Upload folder using huggingface_hub
Browse files- config.json +49 -0
- configuration_aria.py +114 -0
- generation_config.json +7 -0
- model-00001-of-00012.safetensors +3 -0
- model-00002-of-00012.safetensors +3 -0
- model-00003-of-00012.safetensors +3 -0
- model-00004-of-00012.safetensors +3 -0
- model-00005-of-00012.safetensors +3 -0
- model-00006-of-00012.safetensors +3 -0
- model-00007-of-00012.safetensors +3 -0
- model-00008-of-00012.safetensors +3 -0
- model-00009-of-00012.safetensors +3 -0
- model-00010-of-00012.safetensors +3 -0
- model-00011-of-00012.safetensors +3 -0
- model-00012-of-00012.safetensors +3 -0
- model.safetensors.index.json +799 -0
- modeling_aria.py +365 -0
- moe_lm.py +679 -0
- preprocessor_config.json +21 -0
- processing_aria.py +305 -0
- projector.py +189 -0
- special_tokens_map.json +9 -0
- tokenizer_config.json +27 -0
- vision_encoder.py +152 -0
- vision_processor.py +321 -0
config.json
ADDED
|
@@ -0,0 +1,49 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"AriaForConditionalGeneration"
|
| 4 |
+
],
|
| 5 |
+
"auto_map": {
|
| 6 |
+
"AutoConfig": "modeling_aria.AriaConfig",
|
| 7 |
+
"AutoModelForCausalLM": "modeling_aria.AriaForConditionalGeneration"
|
| 8 |
+
},
|
| 9 |
+
"ignore_index": -100,
|
| 10 |
+
"image_token_index": 9,
|
| 11 |
+
"model_type": "aria",
|
| 12 |
+
"projector_patch_to_query_dict": {
|
| 13 |
+
"1225": 128,
|
| 14 |
+
"4900": 256
|
| 15 |
+
},
|
| 16 |
+
"text_config": {
|
| 17 |
+
"hidden_size": 2560,
|
| 18 |
+
"intermediate_size": 13568,
|
| 19 |
+
"max_position_embeddings": 65536,
|
| 20 |
+
"model_type": "aria_moe_lm",
|
| 21 |
+
"moe_intermediate_size": 1664,
|
| 22 |
+
"moe_num_experts": 64,
|
| 23 |
+
"moe_topk": 6,
|
| 24 |
+
"num_attention_heads": 20,
|
| 25 |
+
"num_experts_per_tok": 6,
|
| 26 |
+
"num_hidden_layers": 28,
|
| 27 |
+
"num_key_value_heads": 20,
|
| 28 |
+
"rope_theta": 5000000,
|
| 29 |
+
"vocab_size": 100352
|
| 30 |
+
},
|
| 31 |
+
"torch_dtype": "bfloat16",
|
| 32 |
+
"transformers_version": "4.45.0",
|
| 33 |
+
"_attn_implementation": "flash_attention_2",
|
| 34 |
+
"vision_config": {
|
| 35 |
+
"_flash_attn_2_enabled": true,
|
| 36 |
+
"_attn_implementation": "flash_attention_2",
|
| 37 |
+
"architectures": [
|
| 38 |
+
"AriaVisionModel"
|
| 39 |
+
],
|
| 40 |
+
"hidden_size": 1152,
|
| 41 |
+
"image_size": 980,
|
| 42 |
+
"intermediate_size": 4304,
|
| 43 |
+
"model_type": "aria_vision_model",
|
| 44 |
+
"num_attention_heads": 16,
|
| 45 |
+
"num_hidden_layers": 27,
|
| 46 |
+
"patch_size": 14,
|
| 47 |
+
"torch_dtype": "bfloat16"
|
| 48 |
+
}
|
| 49 |
+
}
|
configuration_aria.py
ADDED
|
@@ -0,0 +1,114 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright 2024 Rhymes AI. All rights reserved.
|
| 2 |
+
#
|
| 3 |
+
# Licensed to the Apache Software Foundation (ASF) under one
|
| 4 |
+
# or more contributor license agreements. See the NOTICE file
|
| 5 |
+
# distributed with this work for additional information
|
| 6 |
+
# regarding copyright ownership. The ASF licenses this file
|
| 7 |
+
# to you under the Apache License, Version 2.0 (the
|
| 8 |
+
# "License"); you may not use this file except in compliance
|
| 9 |
+
# with the License. You may obtain a copy of the License at
|
| 10 |
+
#
|
| 11 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 12 |
+
#
|
| 13 |
+
# Unless required by applicable law or agreed to in writing,
|
| 14 |
+
# software distributed under the License is distributed on an
|
| 15 |
+
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
| 16 |
+
# KIND, either express or implied. See the License for the
|
| 17 |
+
# specific language governing permissions and limitations
|
| 18 |
+
# under the License.
|
| 19 |
+
|
| 20 |
+
import logging
|
| 21 |
+
|
| 22 |
+
from transformers.configuration_utils import PretrainedConfig
|
| 23 |
+
|
| 24 |
+
from .moe_lm import AriaMoELMConfig
|
| 25 |
+
from .vision_encoder import AriaVisionConfig
|
| 26 |
+
|
| 27 |
+
logger = logging.getLogger(__name__)
|
| 28 |
+
|
| 29 |
+
|
| 30 |
+
# adapted from transformers.models.llava.configuration_llava.LlavaConfig
|
| 31 |
+
class AriaConfig(PretrainedConfig):
|
| 32 |
+
"""
|
| 33 |
+
Configuration class for Aria model.
|
| 34 |
+
|
| 35 |
+
This class handles the configuration for both vision and text components of the Aria model,
|
| 36 |
+
as well as additional parameters for image token handling and projector mapping.
|
| 37 |
+
|
| 38 |
+
Args:
|
| 39 |
+
vision_config (AriaVisionConfig or dict): Configuration for the vision component.
|
| 40 |
+
text_config (AriaMoELMConfig or dict): Configuration for the text component.
|
| 41 |
+
projector_patch_to_query_dict (dict): Mapping of patch sizes to query dimensions.
|
| 42 |
+
ignore_index (int): Index to ignore in loss calculation.
|
| 43 |
+
image_token_index (int): Index used to represent image tokens.
|
| 44 |
+
**kwargs: Additional keyword arguments passed to the parent class.
|
| 45 |
+
|
| 46 |
+
Attributes:
|
| 47 |
+
model_type (str): Type of the model, set to "aria".
|
| 48 |
+
is_composition (bool): Whether the model is a composition of multiple components.
|
| 49 |
+
ignore_index (int): Index to ignore in loss calculation.
|
| 50 |
+
image_token_index (int): Index used to represent image tokens.
|
| 51 |
+
projector_patch_to_query_dict (dict): Mapping of patch sizes to query dimensions.
|
| 52 |
+
vision_config (AriaVisionConfig): Configuration for the vision component.
|
| 53 |
+
text_config (AriaMoELMConfig): Configuration for the text component.
|
| 54 |
+
"""
|
| 55 |
+
|
| 56 |
+
model_type = "aria"
|
| 57 |
+
is_composition = False
|
| 58 |
+
|
| 59 |
+
def __init__(
|
| 60 |
+
self,
|
| 61 |
+
vision_config=AriaVisionConfig(),
|
| 62 |
+
text_config=AriaMoELMConfig(),
|
| 63 |
+
projector_patch_to_query_dict={
|
| 64 |
+
1225: 128,
|
| 65 |
+
4900: 256,
|
| 66 |
+
},
|
| 67 |
+
ignore_index=-100,
|
| 68 |
+
image_token_index=32000,
|
| 69 |
+
tie_word_embeddings=False,
|
| 70 |
+
**kwargs,
|
| 71 |
+
):
|
| 72 |
+
super().__init__(**kwargs)
|
| 73 |
+
self.ignore_index = ignore_index
|
| 74 |
+
self.image_token_index = image_token_index
|
| 75 |
+
self.tie_word_embeddings = tie_word_embeddings
|
| 76 |
+
attn_implementation = kwargs.pop("attn_implementation", None)
|
| 77 |
+
|
| 78 |
+
# Set the default attention implementation to flash_attention_2 if not specified
|
| 79 |
+
self._attn_implementation = (
|
| 80 |
+
"flash_attention_2" if attn_implementation is None else attn_implementation
|
| 81 |
+
)
|
| 82 |
+
|
| 83 |
+
# Convert the keys and values of projector_patch_to_query_dict to integers
|
| 84 |
+
# This ensures consistency even if they were provided as strings
|
| 85 |
+
self.projector_patch_to_query_dict = {
|
| 86 |
+
int(k): int(v) for k, v in projector_patch_to_query_dict.items()
|
| 87 |
+
}
|
| 88 |
+
|
| 89 |
+
if isinstance(vision_config, dict) and "model_type" in vision_config:
|
| 90 |
+
vision_config = AriaVisionConfig(**vision_config)
|
| 91 |
+
if attn_implementation is None:
|
| 92 |
+
vision_attn_implementation = "flash_attention_2"
|
| 93 |
+
elif attn_implementation == "sdpa":
|
| 94 |
+
logger.warning(
|
| 95 |
+
"SDPA is not supported for vit, using flash_attention_2 instead"
|
| 96 |
+
)
|
| 97 |
+
vision_attn_implementation = "flash_attention_2"
|
| 98 |
+
else:
|
| 99 |
+
vision_attn_implementation = attn_implementation
|
| 100 |
+
vision_config._attn_implementation = vision_attn_implementation
|
| 101 |
+
|
| 102 |
+
self.vision_config = vision_config
|
| 103 |
+
|
| 104 |
+
if isinstance(text_config, dict) and "model_type" in text_config:
|
| 105 |
+
text_attn_implementation = (
|
| 106 |
+
"sdpa" if attn_implementation is None else attn_implementation
|
| 107 |
+
)
|
| 108 |
+
text_config = AriaMoELMConfig(**text_config)
|
| 109 |
+
text_config._attn_implementation = text_attn_implementation
|
| 110 |
+
|
| 111 |
+
self.text_config = text_config
|
| 112 |
+
|
| 113 |
+
# This is needed for the static kv cache
|
| 114 |
+
self.num_hidden_layers = self.text_config.num_hidden_layers
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"pad_token_id": 2,
|
| 6 |
+
"transformers_version": "4.45.0"
|
| 7 |
+
}
|
model-00001-of-00012.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:60b373aec7648973696bf09f1be8efabee81d70cadd54054cb8d5cfc7b9ef353
|
| 3 |
+
size 4922385968
|
model-00002-of-00012.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e28ea6899bee9fb55c3d684dd3c5138343dbcc3112b67a7bebc4f121a69cb83e
|
| 3 |
+
size 4569849136
|
model-00003-of-00012.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:74d1a04c14400e55c3350de920dba71b762eefa8c0894631e7e79e3dd7041aef
|
| 3 |
+
size 4128475752
|
model-00004-of-00012.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fe416e790b16039c1dc4db6ef84b9e43146da1fd8a21137ed20a1e52fdbae75f
|
| 3 |
+
size 4569849136
|
model-00005-of-00012.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:65f15b84f52d16e686c11a92a0034b83de889a1d82c87bd3873587845a933bb2
|
| 3 |
+
size 4128475792
|
model-00006-of-00012.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:915eff7120204b9366c9404c5a7d5280f02f1b855f7afe90b17f5b0de52c558e
|
| 3 |
+
size 4569849160
|
model-00007-of-00012.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:262115131343776b9dd919bf4955f444900301e0c7fdd4751d6ec2fd713713cd
|
| 3 |
+
size 4128475784
|
model-00008-of-00012.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9da5c8b02a7a0774a35b4b71c17d9b4bcfb1b1b8e74e4b176f958a2d5300a9bb
|
| 3 |
+
size 4569849160
|
model-00009-of-00012.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0a616cb997f240d1fd12b2d5da4f8560f868c0e9e7209202d16496d4c9616c02
|
| 3 |
+
size 4128475784
|
model-00010-of-00012.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:af6548e72103d7b208c639a892249593c7dcbe71280b51dee15e6fce5548ece6
|
| 3 |
+
size 4569849160
|
model-00011-of-00012.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8e06b62376377fa6f5fc479b3d463af45c6053d99c8ba485711a8ef369f375f0
|
| 3 |
+
size 4128475784
|
model-00012-of-00012.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d43c7bbf7a74516d5dcb8aa62b096409c7ade75a88c4d96ab963c512942e52b0
|
| 3 |
+
size 2200715456
|
model.safetensors.index.json
ADDED
|
@@ -0,0 +1,799 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"metadata": {
|
| 3 |
+
"total_size": 50614617824
|
| 4 |
+
},
|
| 5 |
+
"weight_map": {
|
| 6 |
+
"language_model.lm_head.weight": "model-00012-of-00012.safetensors",
|
| 7 |
+
"language_model.model.embed_tokens.weight": "model-00001-of-00012.safetensors",
|
| 8 |
+
"language_model.model.layers.0.input_layernorm.weight": "model-00001-of-00012.safetensors",
|
| 9 |
+
"language_model.model.layers.0.mlp.experts.fc1.weight": "model-00001-of-00012.safetensors",
|
| 10 |
+
"language_model.model.layers.0.mlp.experts.fc2.weight": "model-00001-of-00012.safetensors",
|
| 11 |
+
"language_model.model.layers.0.mlp.router.weight": "model-00001-of-00012.safetensors",
|
| 12 |
+
"language_model.model.layers.0.mlp.shared_experts.down_proj.weight": "model-00001-of-00012.safetensors",
|
| 13 |
+
"language_model.model.layers.0.mlp.shared_experts.gate_proj.weight": "model-00001-of-00012.safetensors",
|
| 14 |
+
"language_model.model.layers.0.mlp.shared_experts.up_proj.weight": "model-00001-of-00012.safetensors",
|
| 15 |
+
"language_model.model.layers.0.post_attention_layernorm.weight": "model-00001-of-00012.safetensors",
|
| 16 |
+
"language_model.model.layers.0.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 17 |
+
"language_model.model.layers.0.self_attn.o_proj.weight": "model-00001-of-00012.safetensors",
|
| 18 |
+
"language_model.model.layers.0.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 19 |
+
"language_model.model.layers.0.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 20 |
+
"language_model.model.layers.1.input_layernorm.weight": "model-00001-of-00012.safetensors",
|
| 21 |
+
"language_model.model.layers.1.mlp.experts.fc1.weight": "model-00001-of-00012.safetensors",
|
| 22 |
+
"language_model.model.layers.1.mlp.experts.fc2.weight": "model-00001-of-00012.safetensors",
|
| 23 |
+
"language_model.model.layers.1.mlp.router.weight": "model-00001-of-00012.safetensors",
|
| 24 |
+
"language_model.model.layers.1.mlp.shared_experts.down_proj.weight": "model-00001-of-00012.safetensors",
|
| 25 |
+
"language_model.model.layers.1.mlp.shared_experts.gate_proj.weight": "model-00001-of-00012.safetensors",
|
| 26 |
+
"language_model.model.layers.1.mlp.shared_experts.up_proj.weight": "model-00001-of-00012.safetensors",
|
| 27 |
+
"language_model.model.layers.1.post_attention_layernorm.weight": "model-00001-of-00012.safetensors",
|
| 28 |
+
"language_model.model.layers.1.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 29 |
+
"language_model.model.layers.1.self_attn.o_proj.weight": "model-00001-of-00012.safetensors",
|
| 30 |
+
"language_model.model.layers.1.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 31 |
+
"language_model.model.layers.1.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 32 |
+
"language_model.model.layers.10.input_layernorm.weight": "model-00005-of-00012.safetensors",
|
| 33 |
+
"language_model.model.layers.10.mlp.experts.fc1.weight": "model-00005-of-00012.safetensors",
|
| 34 |
+
"language_model.model.layers.10.mlp.experts.fc2.weight": "model-00005-of-00012.safetensors",
|
| 35 |
+
"language_model.model.layers.10.mlp.router.weight": "model-00005-of-00012.safetensors",
|
| 36 |
+
"language_model.model.layers.10.mlp.shared_experts.down_proj.weight": "model-00005-of-00012.safetensors",
|
| 37 |
+
"language_model.model.layers.10.mlp.shared_experts.gate_proj.weight": "model-00005-of-00012.safetensors",
|
| 38 |
+
"language_model.model.layers.10.mlp.shared_experts.up_proj.weight": "model-00005-of-00012.safetensors",
|
| 39 |
+
"language_model.model.layers.10.post_attention_layernorm.weight": "model-00005-of-00012.safetensors",
|
| 40 |
+
"language_model.model.layers.10.self_attn.k_proj.weight": "model-00005-of-00012.safetensors",
|
| 41 |
+
"language_model.model.layers.10.self_attn.o_proj.weight": "model-00005-of-00012.safetensors",
|
| 42 |
+
"language_model.model.layers.10.self_attn.q_proj.weight": "model-00005-of-00012.safetensors",
|
| 43 |
+
"language_model.model.layers.10.self_attn.v_proj.weight": "model-00005-of-00012.safetensors",
|
| 44 |
+
"language_model.model.layers.11.input_layernorm.weight": "model-00005-of-00012.safetensors",
|
| 45 |
+
"language_model.model.layers.11.mlp.experts.fc1.weight": "model-00005-of-00012.safetensors",
|
| 46 |
+
"language_model.model.layers.11.mlp.experts.fc2.weight": "model-00005-of-00012.safetensors",
|
| 47 |
+
"language_model.model.layers.11.mlp.router.weight": "model-00005-of-00012.safetensors",
|
| 48 |
+
"language_model.model.layers.11.mlp.shared_experts.down_proj.weight": "model-00005-of-00012.safetensors",
|
| 49 |
+
"language_model.model.layers.11.mlp.shared_experts.gate_proj.weight": "model-00005-of-00012.safetensors",
|
| 50 |
+
"language_model.model.layers.11.mlp.shared_experts.up_proj.weight": "model-00005-of-00012.safetensors",
|
| 51 |
+
"language_model.model.layers.11.post_attention_layernorm.weight": "model-00005-of-00012.safetensors",
|
| 52 |
+
"language_model.model.layers.11.self_attn.k_proj.weight": "model-00005-of-00012.safetensors",
|
| 53 |
+
"language_model.model.layers.11.self_attn.o_proj.weight": "model-00005-of-00012.safetensors",
|
| 54 |
+
"language_model.model.layers.11.self_attn.q_proj.weight": "model-00005-of-00012.safetensors",
|
| 55 |
+
"language_model.model.layers.11.self_attn.v_proj.weight": "model-00005-of-00012.safetensors",
|
| 56 |
+
"language_model.model.layers.12.input_layernorm.weight": "model-00006-of-00012.safetensors",
|
| 57 |
+
"language_model.model.layers.12.mlp.experts.fc1.weight": "model-00006-of-00012.safetensors",
|
| 58 |
+
"language_model.model.layers.12.mlp.experts.fc2.weight": "model-00006-of-00012.safetensors",
|
| 59 |
+
"language_model.model.layers.12.mlp.router.weight": "model-00005-of-00012.safetensors",
|
| 60 |
+
"language_model.model.layers.12.mlp.shared_experts.down_proj.weight": "model-00006-of-00012.safetensors",
|
| 61 |
+
"language_model.model.layers.12.mlp.shared_experts.gate_proj.weight": "model-00006-of-00012.safetensors",
|
| 62 |
+
"language_model.model.layers.12.mlp.shared_experts.up_proj.weight": "model-00006-of-00012.safetensors",
|
| 63 |
+
"language_model.model.layers.12.post_attention_layernorm.weight": "model-00006-of-00012.safetensors",
|
| 64 |
+
"language_model.model.layers.12.self_attn.k_proj.weight": "model-00005-of-00012.safetensors",
|
| 65 |
+
"language_model.model.layers.12.self_attn.o_proj.weight": "model-00005-of-00012.safetensors",
|
| 66 |
+
"language_model.model.layers.12.self_attn.q_proj.weight": "model-00005-of-00012.safetensors",
|
| 67 |
+
"language_model.model.layers.12.self_attn.v_proj.weight": "model-00005-of-00012.safetensors",
|
| 68 |
+
"language_model.model.layers.13.input_layernorm.weight": "model-00006-of-00012.safetensors",
|
| 69 |
+
"language_model.model.layers.13.mlp.experts.fc1.weight": "model-00006-of-00012.safetensors",
|
| 70 |
+
"language_model.model.layers.13.mlp.experts.fc2.weight": "model-00006-of-00012.safetensors",
|
| 71 |
+
"language_model.model.layers.13.mlp.router.weight": "model-00006-of-00012.safetensors",
|
| 72 |
+
"language_model.model.layers.13.mlp.shared_experts.down_proj.weight": "model-00006-of-00012.safetensors",
|
| 73 |
+
"language_model.model.layers.13.mlp.shared_experts.gate_proj.weight": "model-00006-of-00012.safetensors",
|
| 74 |
+
"language_model.model.layers.13.mlp.shared_experts.up_proj.weight": "model-00006-of-00012.safetensors",
|
| 75 |
+
"language_model.model.layers.13.post_attention_layernorm.weight": "model-00006-of-00012.safetensors",
|
| 76 |
+
"language_model.model.layers.13.self_attn.k_proj.weight": "model-00006-of-00012.safetensors",
|
| 77 |
+
"language_model.model.layers.13.self_attn.o_proj.weight": "model-00006-of-00012.safetensors",
|
| 78 |
+
"language_model.model.layers.13.self_attn.q_proj.weight": "model-00006-of-00012.safetensors",
|
| 79 |
+
"language_model.model.layers.13.self_attn.v_proj.weight": "model-00006-of-00012.safetensors",
|
| 80 |
+
"language_model.model.layers.14.input_layernorm.weight": "model-00007-of-00012.safetensors",
|
| 81 |
+
"language_model.model.layers.14.mlp.experts.fc1.weight": "model-00006-of-00012.safetensors",
|
| 82 |
+
"language_model.model.layers.14.mlp.experts.fc2.weight": "model-00007-of-00012.safetensors",
|
| 83 |
+
"language_model.model.layers.14.mlp.router.weight": "model-00006-of-00012.safetensors",
|
| 84 |
+
"language_model.model.layers.14.mlp.shared_experts.down_proj.weight": "model-00007-of-00012.safetensors",
|
| 85 |
+
"language_model.model.layers.14.mlp.shared_experts.gate_proj.weight": "model-00007-of-00012.safetensors",
|
| 86 |
+
"language_model.model.layers.14.mlp.shared_experts.up_proj.weight": "model-00007-of-00012.safetensors",
|
| 87 |
+
"language_model.model.layers.14.post_attention_layernorm.weight": "model-00007-of-00012.safetensors",
|
| 88 |
+
"language_model.model.layers.14.self_attn.k_proj.weight": "model-00006-of-00012.safetensors",
|
| 89 |
+
"language_model.model.layers.14.self_attn.o_proj.weight": "model-00006-of-00012.safetensors",
|
| 90 |
+
"language_model.model.layers.14.self_attn.q_proj.weight": "model-00006-of-00012.safetensors",
|
| 91 |
+
"language_model.model.layers.14.self_attn.v_proj.weight": "model-00006-of-00012.safetensors",
|
| 92 |
+
"language_model.model.layers.15.input_layernorm.weight": "model-00007-of-00012.safetensors",
|
| 93 |
+
"language_model.model.layers.15.mlp.experts.fc1.weight": "model-00007-of-00012.safetensors",
|
| 94 |
+
"language_model.model.layers.15.mlp.experts.fc2.weight": "model-00007-of-00012.safetensors",
|
| 95 |
+
"language_model.model.layers.15.mlp.router.weight": "model-00007-of-00012.safetensors",
|
| 96 |
+
"language_model.model.layers.15.mlp.shared_experts.down_proj.weight": "model-00007-of-00012.safetensors",
|
| 97 |
+
"language_model.model.layers.15.mlp.shared_experts.gate_proj.weight": "model-00007-of-00012.safetensors",
|
| 98 |
+
"language_model.model.layers.15.mlp.shared_experts.up_proj.weight": "model-00007-of-00012.safetensors",
|
| 99 |
+
"language_model.model.layers.15.post_attention_layernorm.weight": "model-00007-of-00012.safetensors",
|
| 100 |
+
"language_model.model.layers.15.self_attn.k_proj.weight": "model-00007-of-00012.safetensors",
|
| 101 |
+
"language_model.model.layers.15.self_attn.o_proj.weight": "model-00007-of-00012.safetensors",
|
| 102 |
+
"language_model.model.layers.15.self_attn.q_proj.weight": "model-00007-of-00012.safetensors",
|
| 103 |
+
"language_model.model.layers.15.self_attn.v_proj.weight": "model-00007-of-00012.safetensors",
|
| 104 |
+
"language_model.model.layers.16.input_layernorm.weight": "model-00007-of-00012.safetensors",
|
| 105 |
+
"language_model.model.layers.16.mlp.experts.fc1.weight": "model-00007-of-00012.safetensors",
|
| 106 |
+
"language_model.model.layers.16.mlp.experts.fc2.weight": "model-00007-of-00012.safetensors",
|
| 107 |
+
"language_model.model.layers.16.mlp.router.weight": "model-00007-of-00012.safetensors",
|
| 108 |
+
"language_model.model.layers.16.mlp.shared_experts.down_proj.weight": "model-00007-of-00012.safetensors",
|
| 109 |
+
"language_model.model.layers.16.mlp.shared_experts.gate_proj.weight": "model-00007-of-00012.safetensors",
|
| 110 |
+
"language_model.model.layers.16.mlp.shared_experts.up_proj.weight": "model-00007-of-00012.safetensors",
|
| 111 |
+
"language_model.model.layers.16.post_attention_layernorm.weight": "model-00007-of-00012.safetensors",
|
| 112 |
+
"language_model.model.layers.16.self_attn.k_proj.weight": "model-00007-of-00012.safetensors",
|
| 113 |
+
"language_model.model.layers.16.self_attn.o_proj.weight": "model-00007-of-00012.safetensors",
|
| 114 |
+
"language_model.model.layers.16.self_attn.q_proj.weight": "model-00007-of-00012.safetensors",
|
| 115 |
+
"language_model.model.layers.16.self_attn.v_proj.weight": "model-00007-of-00012.safetensors",
|
| 116 |
+
"language_model.model.layers.17.input_layernorm.weight": "model-00008-of-00012.safetensors",
|
| 117 |
+
"language_model.model.layers.17.mlp.experts.fc1.weight": "model-00008-of-00012.safetensors",
|
| 118 |
+
"language_model.model.layers.17.mlp.experts.fc2.weight": "model-00008-of-00012.safetensors",
|
| 119 |
+
"language_model.model.layers.17.mlp.router.weight": "model-00007-of-00012.safetensors",
|
| 120 |
+
"language_model.model.layers.17.mlp.shared_experts.down_proj.weight": "model-00008-of-00012.safetensors",
|
| 121 |
+
"language_model.model.layers.17.mlp.shared_experts.gate_proj.weight": "model-00008-of-00012.safetensors",
|
| 122 |
+
"language_model.model.layers.17.mlp.shared_experts.up_proj.weight": "model-00008-of-00012.safetensors",
|
| 123 |
+
"language_model.model.layers.17.post_attention_layernorm.weight": "model-00008-of-00012.safetensors",
|
| 124 |
+
"language_model.model.layers.17.self_attn.k_proj.weight": "model-00007-of-00012.safetensors",
|
| 125 |
+
"language_model.model.layers.17.self_attn.o_proj.weight": "model-00007-of-00012.safetensors",
|
| 126 |
+
"language_model.model.layers.17.self_attn.q_proj.weight": "model-00007-of-00012.safetensors",
|
| 127 |
+
"language_model.model.layers.17.self_attn.v_proj.weight": "model-00007-of-00012.safetensors",
|
| 128 |
+
"language_model.model.layers.18.input_layernorm.weight": "model-00008-of-00012.safetensors",
|
| 129 |
+
"language_model.model.layers.18.mlp.experts.fc1.weight": "model-00008-of-00012.safetensors",
|
| 130 |
+
"language_model.model.layers.18.mlp.experts.fc2.weight": "model-00008-of-00012.safetensors",
|
| 131 |
+
"language_model.model.layers.18.mlp.router.weight": "model-00008-of-00012.safetensors",
|
| 132 |
+
"language_model.model.layers.18.mlp.shared_experts.down_proj.weight": "model-00008-of-00012.safetensors",
|
| 133 |
+
"language_model.model.layers.18.mlp.shared_experts.gate_proj.weight": "model-00008-of-00012.safetensors",
|
| 134 |
+
"language_model.model.layers.18.mlp.shared_experts.up_proj.weight": "model-00008-of-00012.safetensors",
|
| 135 |
+
"language_model.model.layers.18.post_attention_layernorm.weight": "model-00008-of-00012.safetensors",
|
| 136 |
+
"language_model.model.layers.18.self_attn.k_proj.weight": "model-00008-of-00012.safetensors",
|
| 137 |
+
"language_model.model.layers.18.self_attn.o_proj.weight": "model-00008-of-00012.safetensors",
|
| 138 |
+
"language_model.model.layers.18.self_attn.q_proj.weight": "model-00008-of-00012.safetensors",
|
| 139 |
+
"language_model.model.layers.18.self_attn.v_proj.weight": "model-00008-of-00012.safetensors",
|
| 140 |
+
"language_model.model.layers.19.input_layernorm.weight": "model-00009-of-00012.safetensors",
|
| 141 |
+
"language_model.model.layers.19.mlp.experts.fc1.weight": "model-00008-of-00012.safetensors",
|
| 142 |
+
"language_model.model.layers.19.mlp.experts.fc2.weight": "model-00009-of-00012.safetensors",
|
| 143 |
+
"language_model.model.layers.19.mlp.router.weight": "model-00008-of-00012.safetensors",
|
| 144 |
+
"language_model.model.layers.19.mlp.shared_experts.down_proj.weight": "model-00009-of-00012.safetensors",
|
| 145 |
+
"language_model.model.layers.19.mlp.shared_experts.gate_proj.weight": "model-00009-of-00012.safetensors",
|
| 146 |
+
"language_model.model.layers.19.mlp.shared_experts.up_proj.weight": "model-00009-of-00012.safetensors",
|
| 147 |
+
"language_model.model.layers.19.post_attention_layernorm.weight": "model-00009-of-00012.safetensors",
|
| 148 |
+
"language_model.model.layers.19.self_attn.k_proj.weight": "model-00008-of-00012.safetensors",
|
| 149 |
+
"language_model.model.layers.19.self_attn.o_proj.weight": "model-00008-of-00012.safetensors",
|
| 150 |
+
"language_model.model.layers.19.self_attn.q_proj.weight": "model-00008-of-00012.safetensors",
|
| 151 |
+
"language_model.model.layers.19.self_attn.v_proj.weight": "model-00008-of-00012.safetensors",
|
| 152 |
+
"language_model.model.layers.2.input_layernorm.weight": "model-00002-of-00012.safetensors",
|
| 153 |
+
"language_model.model.layers.2.mlp.experts.fc1.weight": "model-00002-of-00012.safetensors",
|
| 154 |
+
"language_model.model.layers.2.mlp.experts.fc2.weight": "model-00002-of-00012.safetensors",
|
| 155 |
+
"language_model.model.layers.2.mlp.router.weight": "model-00001-of-00012.safetensors",
|
| 156 |
+
"language_model.model.layers.2.mlp.shared_experts.down_proj.weight": "model-00002-of-00012.safetensors",
|
| 157 |
+
"language_model.model.layers.2.mlp.shared_experts.gate_proj.weight": "model-00002-of-00012.safetensors",
|
| 158 |
+
"language_model.model.layers.2.mlp.shared_experts.up_proj.weight": "model-00002-of-00012.safetensors",
|
| 159 |
+
"language_model.model.layers.2.post_attention_layernorm.weight": "model-00002-of-00012.safetensors",
|
| 160 |
+
"language_model.model.layers.2.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 161 |
+
"language_model.model.layers.2.self_attn.o_proj.weight": "model-00001-of-00012.safetensors",
|
| 162 |
+
"language_model.model.layers.2.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 163 |
+
"language_model.model.layers.2.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 164 |
+
"language_model.model.layers.20.input_layernorm.weight": "model-00009-of-00012.safetensors",
|
| 165 |
+
"language_model.model.layers.20.mlp.experts.fc1.weight": "model-00009-of-00012.safetensors",
|
| 166 |
+
"language_model.model.layers.20.mlp.experts.fc2.weight": "model-00009-of-00012.safetensors",
|
| 167 |
+
"language_model.model.layers.20.mlp.router.weight": "model-00009-of-00012.safetensors",
|
| 168 |
+
"language_model.model.layers.20.mlp.shared_experts.down_proj.weight": "model-00009-of-00012.safetensors",
|
| 169 |
+
"language_model.model.layers.20.mlp.shared_experts.gate_proj.weight": "model-00009-of-00012.safetensors",
|
| 170 |
+
"language_model.model.layers.20.mlp.shared_experts.up_proj.weight": "model-00009-of-00012.safetensors",
|
| 171 |
+
"language_model.model.layers.20.post_attention_layernorm.weight": "model-00009-of-00012.safetensors",
|
| 172 |
+
"language_model.model.layers.20.self_attn.k_proj.weight": "model-00009-of-00012.safetensors",
|
| 173 |
+
"language_model.model.layers.20.self_attn.o_proj.weight": "model-00009-of-00012.safetensors",
|
| 174 |
+
"language_model.model.layers.20.self_attn.q_proj.weight": "model-00009-of-00012.safetensors",
|
| 175 |
+
"language_model.model.layers.20.self_attn.v_proj.weight": "model-00009-of-00012.safetensors",
|
| 176 |
+
"language_model.model.layers.21.input_layernorm.weight": "model-00009-of-00012.safetensors",
|
| 177 |
+
"language_model.model.layers.21.mlp.experts.fc1.weight": "model-00009-of-00012.safetensors",
|
| 178 |
+
"language_model.model.layers.21.mlp.experts.fc2.weight": "model-00009-of-00012.safetensors",
|
| 179 |
+
"language_model.model.layers.21.mlp.router.weight": "model-00009-of-00012.safetensors",
|
| 180 |
+
"language_model.model.layers.21.mlp.shared_experts.down_proj.weight": "model-00009-of-00012.safetensors",
|
| 181 |
+
"language_model.model.layers.21.mlp.shared_experts.gate_proj.weight": "model-00009-of-00012.safetensors",
|
| 182 |
+
"language_model.model.layers.21.mlp.shared_experts.up_proj.weight": "model-00009-of-00012.safetensors",
|
| 183 |
+
"language_model.model.layers.21.post_attention_layernorm.weight": "model-00009-of-00012.safetensors",
|
| 184 |
+
"language_model.model.layers.21.self_attn.k_proj.weight": "model-00009-of-00012.safetensors",
|
| 185 |
+
"language_model.model.layers.21.self_attn.o_proj.weight": "model-00009-of-00012.safetensors",
|
| 186 |
+
"language_model.model.layers.21.self_attn.q_proj.weight": "model-00009-of-00012.safetensors",
|
| 187 |
+
"language_model.model.layers.21.self_attn.v_proj.weight": "model-00009-of-00012.safetensors",
|
| 188 |
+
"language_model.model.layers.22.input_layernorm.weight": "model-00010-of-00012.safetensors",
|
| 189 |
+
"language_model.model.layers.22.mlp.experts.fc1.weight": "model-00010-of-00012.safetensors",
|
| 190 |
+
"language_model.model.layers.22.mlp.experts.fc2.weight": "model-00010-of-00012.safetensors",
|
| 191 |
+
"language_model.model.layers.22.mlp.router.weight": "model-00009-of-00012.safetensors",
|
| 192 |
+
"language_model.model.layers.22.mlp.shared_experts.down_proj.weight": "model-00010-of-00012.safetensors",
|
| 193 |
+
"language_model.model.layers.22.mlp.shared_experts.gate_proj.weight": "model-00010-of-00012.safetensors",
|
| 194 |
+
"language_model.model.layers.22.mlp.shared_experts.up_proj.weight": "model-00010-of-00012.safetensors",
|
| 195 |
+
"language_model.model.layers.22.post_attention_layernorm.weight": "model-00010-of-00012.safetensors",
|
| 196 |
+
"language_model.model.layers.22.self_attn.k_proj.weight": "model-00009-of-00012.safetensors",
|
| 197 |
+
"language_model.model.layers.22.self_attn.o_proj.weight": "model-00009-of-00012.safetensors",
|
| 198 |
+
"language_model.model.layers.22.self_attn.q_proj.weight": "model-00009-of-00012.safetensors",
|
| 199 |
+
"language_model.model.layers.22.self_attn.v_proj.weight": "model-00009-of-00012.safetensors",
|
| 200 |
+
"language_model.model.layers.23.input_layernorm.weight": "model-00010-of-00012.safetensors",
|
| 201 |
+
"language_model.model.layers.23.mlp.experts.fc1.weight": "model-00010-of-00012.safetensors",
|
| 202 |
+
"language_model.model.layers.23.mlp.experts.fc2.weight": "model-00010-of-00012.safetensors",
|
| 203 |
+
"language_model.model.layers.23.mlp.router.weight": "model-00010-of-00012.safetensors",
|
| 204 |
+
"language_model.model.layers.23.mlp.shared_experts.down_proj.weight": "model-00010-of-00012.safetensors",
|
| 205 |
+
"language_model.model.layers.23.mlp.shared_experts.gate_proj.weight": "model-00010-of-00012.safetensors",
|
| 206 |
+
"language_model.model.layers.23.mlp.shared_experts.up_proj.weight": "model-00010-of-00012.safetensors",
|
| 207 |
+
"language_model.model.layers.23.post_attention_layernorm.weight": "model-00010-of-00012.safetensors",
|
| 208 |
+
"language_model.model.layers.23.self_attn.k_proj.weight": "model-00010-of-00012.safetensors",
|
| 209 |
+
"language_model.model.layers.23.self_attn.o_proj.weight": "model-00010-of-00012.safetensors",
|
| 210 |
+
"language_model.model.layers.23.self_attn.q_proj.weight": "model-00010-of-00012.safetensors",
|
| 211 |
+
"language_model.model.layers.23.self_attn.v_proj.weight": "model-00010-of-00012.safetensors",
|
| 212 |
+
"language_model.model.layers.24.input_layernorm.weight": "model-00011-of-00012.safetensors",
|
| 213 |
+
"language_model.model.layers.24.mlp.experts.fc1.weight": "model-00010-of-00012.safetensors",
|
| 214 |
+
"language_model.model.layers.24.mlp.experts.fc2.weight": "model-00011-of-00012.safetensors",
|
| 215 |
+
"language_model.model.layers.24.mlp.router.weight": "model-00010-of-00012.safetensors",
|
| 216 |
+
"language_model.model.layers.24.mlp.shared_experts.down_proj.weight": "model-00011-of-00012.safetensors",
|
| 217 |
+
"language_model.model.layers.24.mlp.shared_experts.gate_proj.weight": "model-00011-of-00012.safetensors",
|
| 218 |
+
"language_model.model.layers.24.mlp.shared_experts.up_proj.weight": "model-00011-of-00012.safetensors",
|
| 219 |
+
"language_model.model.layers.24.post_attention_layernorm.weight": "model-00011-of-00012.safetensors",
|
| 220 |
+
"language_model.model.layers.24.self_attn.k_proj.weight": "model-00010-of-00012.safetensors",
|
| 221 |
+
"language_model.model.layers.24.self_attn.o_proj.weight": "model-00010-of-00012.safetensors",
|
| 222 |
+
"language_model.model.layers.24.self_attn.q_proj.weight": "model-00010-of-00012.safetensors",
|
| 223 |
+
"language_model.model.layers.24.self_attn.v_proj.weight": "model-00010-of-00012.safetensors",
|
| 224 |
+
"language_model.model.layers.25.input_layernorm.weight": "model-00011-of-00012.safetensors",
|
| 225 |
+
"language_model.model.layers.25.mlp.experts.fc1.weight": "model-00011-of-00012.safetensors",
|
| 226 |
+
"language_model.model.layers.25.mlp.experts.fc2.weight": "model-00011-of-00012.safetensors",
|
| 227 |
+
"language_model.model.layers.25.mlp.router.weight": "model-00011-of-00012.safetensors",
|
| 228 |
+
"language_model.model.layers.25.mlp.shared_experts.down_proj.weight": "model-00011-of-00012.safetensors",
|
| 229 |
+
"language_model.model.layers.25.mlp.shared_experts.gate_proj.weight": "model-00011-of-00012.safetensors",
|
| 230 |
+
"language_model.model.layers.25.mlp.shared_experts.up_proj.weight": "model-00011-of-00012.safetensors",
|
| 231 |
+
"language_model.model.layers.25.post_attention_layernorm.weight": "model-00011-of-00012.safetensors",
|
| 232 |
+
"language_model.model.layers.25.self_attn.k_proj.weight": "model-00011-of-00012.safetensors",
|
| 233 |
+
"language_model.model.layers.25.self_attn.o_proj.weight": "model-00011-of-00012.safetensors",
|
| 234 |
+
"language_model.model.layers.25.self_attn.q_proj.weight": "model-00011-of-00012.safetensors",
|
| 235 |
+
"language_model.model.layers.25.self_attn.v_proj.weight": "model-00011-of-00012.safetensors",
|
| 236 |
+
"language_model.model.layers.26.input_layernorm.weight": "model-00011-of-00012.safetensors",
|
| 237 |
+
"language_model.model.layers.26.mlp.experts.fc1.weight": "model-00011-of-00012.safetensors",
|
| 238 |
+
"language_model.model.layers.26.mlp.experts.fc2.weight": "model-00011-of-00012.safetensors",
|
| 239 |
+
"language_model.model.layers.26.mlp.router.weight": "model-00011-of-00012.safetensors",
|
| 240 |
+
"language_model.model.layers.26.mlp.shared_experts.down_proj.weight": "model-00011-of-00012.safetensors",
|
| 241 |
+
"language_model.model.layers.26.mlp.shared_experts.gate_proj.weight": "model-00011-of-00012.safetensors",
|
| 242 |
+
"language_model.model.layers.26.mlp.shared_experts.up_proj.weight": "model-00011-of-00012.safetensors",
|
| 243 |
+
"language_model.model.layers.26.post_attention_layernorm.weight": "model-00011-of-00012.safetensors",
|
| 244 |
+
"language_model.model.layers.26.self_attn.k_proj.weight": "model-00011-of-00012.safetensors",
|
| 245 |
+
"language_model.model.layers.26.self_attn.o_proj.weight": "model-00011-of-00012.safetensors",
|
| 246 |
+
"language_model.model.layers.26.self_attn.q_proj.weight": "model-00011-of-00012.safetensors",
|
| 247 |
+
"language_model.model.layers.26.self_attn.v_proj.weight": "model-00011-of-00012.safetensors",
|
| 248 |
+
"language_model.model.layers.27.input_layernorm.weight": "model-00012-of-00012.safetensors",
|
| 249 |
+
"language_model.model.layers.27.mlp.experts.fc1.weight": "model-00012-of-00012.safetensors",
|
| 250 |
+
"language_model.model.layers.27.mlp.experts.fc2.weight": "model-00012-of-00012.safetensors",
|
| 251 |
+
"language_model.model.layers.27.mlp.router.weight": "model-00011-of-00012.safetensors",
|
| 252 |
+
"language_model.model.layers.27.mlp.shared_experts.down_proj.weight": "model-00012-of-00012.safetensors",
|
| 253 |
+
"language_model.model.layers.27.mlp.shared_experts.gate_proj.weight": "model-00012-of-00012.safetensors",
|
| 254 |
+
"language_model.model.layers.27.mlp.shared_experts.up_proj.weight": "model-00012-of-00012.safetensors",
|
| 255 |
+
"language_model.model.layers.27.post_attention_layernorm.weight": "model-00012-of-00012.safetensors",
|
| 256 |
+
"language_model.model.layers.27.self_attn.k_proj.weight": "model-00011-of-00012.safetensors",
|
| 257 |
+
"language_model.model.layers.27.self_attn.o_proj.weight": "model-00011-of-00012.safetensors",
|
| 258 |
+
"language_model.model.layers.27.self_attn.q_proj.weight": "model-00011-of-00012.safetensors",
|
| 259 |
+
"language_model.model.layers.27.self_attn.v_proj.weight": "model-00011-of-00012.safetensors",
|
| 260 |
+
"language_model.model.layers.3.input_layernorm.weight": "model-00002-of-00012.safetensors",
|
| 261 |
+
"language_model.model.layers.3.mlp.experts.fc1.weight": "model-00002-of-00012.safetensors",
|
| 262 |
+
"language_model.model.layers.3.mlp.experts.fc2.weight": "model-00002-of-00012.safetensors",
|
| 263 |
+
"language_model.model.layers.3.mlp.router.weight": "model-00002-of-00012.safetensors",
|
| 264 |
+
"language_model.model.layers.3.mlp.shared_experts.down_proj.weight": "model-00002-of-00012.safetensors",
|
| 265 |
+
"language_model.model.layers.3.mlp.shared_experts.gate_proj.weight": "model-00002-of-00012.safetensors",
|
| 266 |
+
"language_model.model.layers.3.mlp.shared_experts.up_proj.weight": "model-00002-of-00012.safetensors",
|
| 267 |
+
"language_model.model.layers.3.post_attention_layernorm.weight": "model-00002-of-00012.safetensors",
|
| 268 |
+
"language_model.model.layers.3.self_attn.k_proj.weight": "model-00002-of-00012.safetensors",
|
| 269 |
+
"language_model.model.layers.3.self_attn.o_proj.weight": "model-00002-of-00012.safetensors",
|
| 270 |
+
"language_model.model.layers.3.self_attn.q_proj.weight": "model-00002-of-00012.safetensors",
|
| 271 |
+
"language_model.model.layers.3.self_attn.v_proj.weight": "model-00002-of-00012.safetensors",
|
| 272 |
+
"language_model.model.layers.4.input_layernorm.weight": "model-00003-of-00012.safetensors",
|
| 273 |
+
"language_model.model.layers.4.mlp.experts.fc1.weight": "model-00002-of-00012.safetensors",
|
| 274 |
+
"language_model.model.layers.4.mlp.experts.fc2.weight": "model-00003-of-00012.safetensors",
|
| 275 |
+
"language_model.model.layers.4.mlp.router.weight": "model-00002-of-00012.safetensors",
|
| 276 |
+
"language_model.model.layers.4.mlp.shared_experts.down_proj.weight": "model-00003-of-00012.safetensors",
|
| 277 |
+
"language_model.model.layers.4.mlp.shared_experts.gate_proj.weight": "model-00003-of-00012.safetensors",
|
| 278 |
+
"language_model.model.layers.4.mlp.shared_experts.up_proj.weight": "model-00003-of-00012.safetensors",
|
| 279 |
+
"language_model.model.layers.4.post_attention_layernorm.weight": "model-00003-of-00012.safetensors",
|
| 280 |
+
"language_model.model.layers.4.self_attn.k_proj.weight": "model-00002-of-00012.safetensors",
|
| 281 |
+
"language_model.model.layers.4.self_attn.o_proj.weight": "model-00002-of-00012.safetensors",
|
| 282 |
+
"language_model.model.layers.4.self_attn.q_proj.weight": "model-00002-of-00012.safetensors",
|
| 283 |
+
"language_model.model.layers.4.self_attn.v_proj.weight": "model-00002-of-00012.safetensors",
|
| 284 |
+
"language_model.model.layers.5.input_layernorm.weight": "model-00003-of-00012.safetensors",
|
| 285 |
+
"language_model.model.layers.5.mlp.experts.fc1.weight": "model-00003-of-00012.safetensors",
|
| 286 |
+
"language_model.model.layers.5.mlp.experts.fc2.weight": "model-00003-of-00012.safetensors",
|
| 287 |
+
"language_model.model.layers.5.mlp.router.weight": "model-00003-of-00012.safetensors",
|
| 288 |
+
"language_model.model.layers.5.mlp.shared_experts.down_proj.weight": "model-00003-of-00012.safetensors",
|
| 289 |
+
"language_model.model.layers.5.mlp.shared_experts.gate_proj.weight": "model-00003-of-00012.safetensors",
|
| 290 |
+
"language_model.model.layers.5.mlp.shared_experts.up_proj.weight": "model-00003-of-00012.safetensors",
|
| 291 |
+
"language_model.model.layers.5.post_attention_layernorm.weight": "model-00003-of-00012.safetensors",
|
| 292 |
+
"language_model.model.layers.5.self_attn.k_proj.weight": "model-00003-of-00012.safetensors",
|
| 293 |
+
"language_model.model.layers.5.self_attn.o_proj.weight": "model-00003-of-00012.safetensors",
|
| 294 |
+
"language_model.model.layers.5.self_attn.q_proj.weight": "model-00003-of-00012.safetensors",
|
| 295 |
+
"language_model.model.layers.5.self_attn.v_proj.weight": "model-00003-of-00012.safetensors",
|
| 296 |
+
"language_model.model.layers.6.input_layernorm.weight": "model-00003-of-00012.safetensors",
|
| 297 |
+
"language_model.model.layers.6.mlp.experts.fc1.weight": "model-00003-of-00012.safetensors",
|
| 298 |
+
"language_model.model.layers.6.mlp.experts.fc2.weight": "model-00003-of-00012.safetensors",
|
| 299 |
+
"language_model.model.layers.6.mlp.router.weight": "model-00003-of-00012.safetensors",
|
| 300 |
+
"language_model.model.layers.6.mlp.shared_experts.down_proj.weight": "model-00003-of-00012.safetensors",
|
| 301 |
+
"language_model.model.layers.6.mlp.shared_experts.gate_proj.weight": "model-00003-of-00012.safetensors",
|
| 302 |
+
"language_model.model.layers.6.mlp.shared_experts.up_proj.weight": "model-00003-of-00012.safetensors",
|
| 303 |
+
"language_model.model.layers.6.post_attention_layernorm.weight": "model-00003-of-00012.safetensors",
|
| 304 |
+
"language_model.model.layers.6.self_attn.k_proj.weight": "model-00003-of-00012.safetensors",
|
| 305 |
+
"language_model.model.layers.6.self_attn.o_proj.weight": "model-00003-of-00012.safetensors",
|
| 306 |
+
"language_model.model.layers.6.self_attn.q_proj.weight": "model-00003-of-00012.safetensors",
|
| 307 |
+
"language_model.model.layers.6.self_attn.v_proj.weight": "model-00003-of-00012.safetensors",
|
| 308 |
+
"language_model.model.layers.7.input_layernorm.weight": "model-00004-of-00012.safetensors",
|
| 309 |
+
"language_model.model.layers.7.mlp.experts.fc1.weight": "model-00004-of-00012.safetensors",
|
| 310 |
+
"language_model.model.layers.7.mlp.experts.fc2.weight": "model-00004-of-00012.safetensors",
|
| 311 |
+
"language_model.model.layers.7.mlp.router.weight": "model-00003-of-00012.safetensors",
|
| 312 |
+
"language_model.model.layers.7.mlp.shared_experts.down_proj.weight": "model-00004-of-00012.safetensors",
|
| 313 |
+
"language_model.model.layers.7.mlp.shared_experts.gate_proj.weight": "model-00004-of-00012.safetensors",
|
| 314 |
+
"language_model.model.layers.7.mlp.shared_experts.up_proj.weight": "model-00004-of-00012.safetensors",
|
| 315 |
+
"language_model.model.layers.7.post_attention_layernorm.weight": "model-00004-of-00012.safetensors",
|
| 316 |
+
"language_model.model.layers.7.self_attn.k_proj.weight": "model-00003-of-00012.safetensors",
|
| 317 |
+
"language_model.model.layers.7.self_attn.o_proj.weight": "model-00003-of-00012.safetensors",
|
| 318 |
+
"language_model.model.layers.7.self_attn.q_proj.weight": "model-00003-of-00012.safetensors",
|
| 319 |
+
"language_model.model.layers.7.self_attn.v_proj.weight": "model-00003-of-00012.safetensors",
|
| 320 |
+
"language_model.model.layers.8.input_layernorm.weight": "model-00004-of-00012.safetensors",
|
| 321 |
+
"language_model.model.layers.8.mlp.experts.fc1.weight": "model-00004-of-00012.safetensors",
|
| 322 |
+
"language_model.model.layers.8.mlp.experts.fc2.weight": "model-00004-of-00012.safetensors",
|
| 323 |
+
"language_model.model.layers.8.mlp.router.weight": "model-00004-of-00012.safetensors",
|
| 324 |
+
"language_model.model.layers.8.mlp.shared_experts.down_proj.weight": "model-00004-of-00012.safetensors",
|
| 325 |
+
"language_model.model.layers.8.mlp.shared_experts.gate_proj.weight": "model-00004-of-00012.safetensors",
|
| 326 |
+
"language_model.model.layers.8.mlp.shared_experts.up_proj.weight": "model-00004-of-00012.safetensors",
|
| 327 |
+
"language_model.model.layers.8.post_attention_layernorm.weight": "model-00004-of-00012.safetensors",
|
| 328 |
+
"language_model.model.layers.8.self_attn.k_proj.weight": "model-00004-of-00012.safetensors",
|
| 329 |
+
"language_model.model.layers.8.self_attn.o_proj.weight": "model-00004-of-00012.safetensors",
|
| 330 |
+
"language_model.model.layers.8.self_attn.q_proj.weight": "model-00004-of-00012.safetensors",
|
| 331 |
+
"language_model.model.layers.8.self_attn.v_proj.weight": "model-00004-of-00012.safetensors",
|
| 332 |
+
"language_model.model.layers.9.input_layernorm.weight": "model-00005-of-00012.safetensors",
|
| 333 |
+
"language_model.model.layers.9.mlp.experts.fc1.weight": "model-00004-of-00012.safetensors",
|
| 334 |
+
"language_model.model.layers.9.mlp.experts.fc2.weight": "model-00005-of-00012.safetensors",
|
| 335 |
+
"language_model.model.layers.9.mlp.router.weight": "model-00004-of-00012.safetensors",
|
| 336 |
+
"language_model.model.layers.9.mlp.shared_experts.down_proj.weight": "model-00005-of-00012.safetensors",
|
| 337 |
+
"language_model.model.layers.9.mlp.shared_experts.gate_proj.weight": "model-00005-of-00012.safetensors",
|
| 338 |
+
"language_model.model.layers.9.mlp.shared_experts.up_proj.weight": "model-00005-of-00012.safetensors",
|
| 339 |
+
"language_model.model.layers.9.post_attention_layernorm.weight": "model-00005-of-00012.safetensors",
|
| 340 |
+
"language_model.model.layers.9.self_attn.k_proj.weight": "model-00004-of-00012.safetensors",
|
| 341 |
+
"language_model.model.layers.9.self_attn.o_proj.weight": "model-00004-of-00012.safetensors",
|
| 342 |
+
"language_model.model.layers.9.self_attn.q_proj.weight": "model-00004-of-00012.safetensors",
|
| 343 |
+
"language_model.model.layers.9.self_attn.v_proj.weight": "model-00004-of-00012.safetensors",
|
| 344 |
+
"language_model.model.norm.weight": "model-00012-of-00012.safetensors",
|
| 345 |
+
"multi_modal_projector.cross_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 346 |
+
"multi_modal_projector.cross_attn.layer_norm.bias": "model-00001-of-00012.safetensors",
|
| 347 |
+
"multi_modal_projector.cross_attn.layer_norm.weight": "model-00001-of-00012.safetensors",
|
| 348 |
+
"multi_modal_projector.cross_attn.linear.bias": "model-00001-of-00012.safetensors",
|
| 349 |
+
"multi_modal_projector.cross_attn.linear.weight": "model-00001-of-00012.safetensors",
|
| 350 |
+
"multi_modal_projector.cross_attn.ln_kv.bias": "model-00001-of-00012.safetensors",
|
| 351 |
+
"multi_modal_projector.cross_attn.ln_kv.weight": "model-00001-of-00012.safetensors",
|
| 352 |
+
"multi_modal_projector.cross_attn.multihead_attn.in_proj_bias": "model-00001-of-00012.safetensors",
|
| 353 |
+
"multi_modal_projector.cross_attn.multihead_attn.in_proj_weight": "model-00001-of-00012.safetensors",
|
| 354 |
+
"multi_modal_projector.cross_attn.multihead_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 355 |
+
"multi_modal_projector.cross_attn.multihead_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 356 |
+
"multi_modal_projector.cross_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 357 |
+
"multi_modal_projector.cross_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 358 |
+
"multi_modal_projector.ffn.linear_in.weight": "model-00001-of-00012.safetensors",
|
| 359 |
+
"multi_modal_projector.ffn.linear_out.weight": "model-00001-of-00012.safetensors",
|
| 360 |
+
"multi_modal_projector.ln_ffn.bias": "model-00001-of-00012.safetensors",
|
| 361 |
+
"multi_modal_projector.ln_ffn.weight": "model-00001-of-00012.safetensors",
|
| 362 |
+
"multi_modal_projector.query": "model-00001-of-00012.safetensors",
|
| 363 |
+
"vision_tower.vision_model.embeddings.patch_embedding.bias": "model-00001-of-00012.safetensors",
|
| 364 |
+
"vision_tower.vision_model.embeddings.patch_embedding.weight": "model-00001-of-00012.safetensors",
|
| 365 |
+
"vision_tower.vision_model.embeddings.position_embedding.weight": "model-00001-of-00012.safetensors",
|
| 366 |
+
"vision_tower.vision_model.encoder.layers.0.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 367 |
+
"vision_tower.vision_model.encoder.layers.0.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 368 |
+
"vision_tower.vision_model.encoder.layers.0.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 369 |
+
"vision_tower.vision_model.encoder.layers.0.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 370 |
+
"vision_tower.vision_model.encoder.layers.0.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 371 |
+
"vision_tower.vision_model.encoder.layers.0.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 372 |
+
"vision_tower.vision_model.encoder.layers.0.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 373 |
+
"vision_tower.vision_model.encoder.layers.0.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 374 |
+
"vision_tower.vision_model.encoder.layers.0.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 375 |
+
"vision_tower.vision_model.encoder.layers.0.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 376 |
+
"vision_tower.vision_model.encoder.layers.0.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 377 |
+
"vision_tower.vision_model.encoder.layers.0.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 378 |
+
"vision_tower.vision_model.encoder.layers.0.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 379 |
+
"vision_tower.vision_model.encoder.layers.0.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 380 |
+
"vision_tower.vision_model.encoder.layers.0.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 381 |
+
"vision_tower.vision_model.encoder.layers.0.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 382 |
+
"vision_tower.vision_model.encoder.layers.1.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 383 |
+
"vision_tower.vision_model.encoder.layers.1.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 384 |
+
"vision_tower.vision_model.encoder.layers.1.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 385 |
+
"vision_tower.vision_model.encoder.layers.1.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 386 |
+
"vision_tower.vision_model.encoder.layers.1.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 387 |
+
"vision_tower.vision_model.encoder.layers.1.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 388 |
+
"vision_tower.vision_model.encoder.layers.1.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 389 |
+
"vision_tower.vision_model.encoder.layers.1.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 390 |
+
"vision_tower.vision_model.encoder.layers.1.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 391 |
+
"vision_tower.vision_model.encoder.layers.1.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 392 |
+
"vision_tower.vision_model.encoder.layers.1.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 393 |
+
"vision_tower.vision_model.encoder.layers.1.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 394 |
+
"vision_tower.vision_model.encoder.layers.1.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 395 |
+
"vision_tower.vision_model.encoder.layers.1.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 396 |
+
"vision_tower.vision_model.encoder.layers.1.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 397 |
+
"vision_tower.vision_model.encoder.layers.1.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 398 |
+
"vision_tower.vision_model.encoder.layers.10.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 399 |
+
"vision_tower.vision_model.encoder.layers.10.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 400 |
+
"vision_tower.vision_model.encoder.layers.10.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 401 |
+
"vision_tower.vision_model.encoder.layers.10.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 402 |
+
"vision_tower.vision_model.encoder.layers.10.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 403 |
+
"vision_tower.vision_model.encoder.layers.10.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 404 |
+
"vision_tower.vision_model.encoder.layers.10.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 405 |
+
"vision_tower.vision_model.encoder.layers.10.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 406 |
+
"vision_tower.vision_model.encoder.layers.10.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 407 |
+
"vision_tower.vision_model.encoder.layers.10.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 408 |
+
"vision_tower.vision_model.encoder.layers.10.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 409 |
+
"vision_tower.vision_model.encoder.layers.10.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 410 |
+
"vision_tower.vision_model.encoder.layers.10.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 411 |
+
"vision_tower.vision_model.encoder.layers.10.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 412 |
+
"vision_tower.vision_model.encoder.layers.10.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 413 |
+
"vision_tower.vision_model.encoder.layers.10.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 414 |
+
"vision_tower.vision_model.encoder.layers.11.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 415 |
+
"vision_tower.vision_model.encoder.layers.11.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 416 |
+
"vision_tower.vision_model.encoder.layers.11.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 417 |
+
"vision_tower.vision_model.encoder.layers.11.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 418 |
+
"vision_tower.vision_model.encoder.layers.11.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 419 |
+
"vision_tower.vision_model.encoder.layers.11.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 420 |
+
"vision_tower.vision_model.encoder.layers.11.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 421 |
+
"vision_tower.vision_model.encoder.layers.11.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 422 |
+
"vision_tower.vision_model.encoder.layers.11.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 423 |
+
"vision_tower.vision_model.encoder.layers.11.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 424 |
+
"vision_tower.vision_model.encoder.layers.11.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 425 |
+
"vision_tower.vision_model.encoder.layers.11.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 426 |
+
"vision_tower.vision_model.encoder.layers.11.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 427 |
+
"vision_tower.vision_model.encoder.layers.11.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 428 |
+
"vision_tower.vision_model.encoder.layers.11.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 429 |
+
"vision_tower.vision_model.encoder.layers.11.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 430 |
+
"vision_tower.vision_model.encoder.layers.12.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 431 |
+
"vision_tower.vision_model.encoder.layers.12.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 432 |
+
"vision_tower.vision_model.encoder.layers.12.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 433 |
+
"vision_tower.vision_model.encoder.layers.12.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 434 |
+
"vision_tower.vision_model.encoder.layers.12.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 435 |
+
"vision_tower.vision_model.encoder.layers.12.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 436 |
+
"vision_tower.vision_model.encoder.layers.12.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 437 |
+
"vision_tower.vision_model.encoder.layers.12.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 438 |
+
"vision_tower.vision_model.encoder.layers.12.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 439 |
+
"vision_tower.vision_model.encoder.layers.12.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 440 |
+
"vision_tower.vision_model.encoder.layers.12.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 441 |
+
"vision_tower.vision_model.encoder.layers.12.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 442 |
+
"vision_tower.vision_model.encoder.layers.12.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 443 |
+
"vision_tower.vision_model.encoder.layers.12.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 444 |
+
"vision_tower.vision_model.encoder.layers.12.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 445 |
+
"vision_tower.vision_model.encoder.layers.12.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 446 |
+
"vision_tower.vision_model.encoder.layers.13.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 447 |
+
"vision_tower.vision_model.encoder.layers.13.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 448 |
+
"vision_tower.vision_model.encoder.layers.13.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 449 |
+
"vision_tower.vision_model.encoder.layers.13.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 450 |
+
"vision_tower.vision_model.encoder.layers.13.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 451 |
+
"vision_tower.vision_model.encoder.layers.13.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 452 |
+
"vision_tower.vision_model.encoder.layers.13.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 453 |
+
"vision_tower.vision_model.encoder.layers.13.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 454 |
+
"vision_tower.vision_model.encoder.layers.13.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 455 |
+
"vision_tower.vision_model.encoder.layers.13.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 456 |
+
"vision_tower.vision_model.encoder.layers.13.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 457 |
+
"vision_tower.vision_model.encoder.layers.13.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 458 |
+
"vision_tower.vision_model.encoder.layers.13.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 459 |
+
"vision_tower.vision_model.encoder.layers.13.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 460 |
+
"vision_tower.vision_model.encoder.layers.13.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 461 |
+
"vision_tower.vision_model.encoder.layers.13.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 462 |
+
"vision_tower.vision_model.encoder.layers.14.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 463 |
+
"vision_tower.vision_model.encoder.layers.14.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 464 |
+
"vision_tower.vision_model.encoder.layers.14.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 465 |
+
"vision_tower.vision_model.encoder.layers.14.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 466 |
+
"vision_tower.vision_model.encoder.layers.14.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 467 |
+
"vision_tower.vision_model.encoder.layers.14.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 468 |
+
"vision_tower.vision_model.encoder.layers.14.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 469 |
+
"vision_tower.vision_model.encoder.layers.14.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 470 |
+
"vision_tower.vision_model.encoder.layers.14.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 471 |
+
"vision_tower.vision_model.encoder.layers.14.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 472 |
+
"vision_tower.vision_model.encoder.layers.14.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 473 |
+
"vision_tower.vision_model.encoder.layers.14.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 474 |
+
"vision_tower.vision_model.encoder.layers.14.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 475 |
+
"vision_tower.vision_model.encoder.layers.14.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 476 |
+
"vision_tower.vision_model.encoder.layers.14.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 477 |
+
"vision_tower.vision_model.encoder.layers.14.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 478 |
+
"vision_tower.vision_model.encoder.layers.15.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 479 |
+
"vision_tower.vision_model.encoder.layers.15.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 480 |
+
"vision_tower.vision_model.encoder.layers.15.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 481 |
+
"vision_tower.vision_model.encoder.layers.15.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 482 |
+
"vision_tower.vision_model.encoder.layers.15.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 483 |
+
"vision_tower.vision_model.encoder.layers.15.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 484 |
+
"vision_tower.vision_model.encoder.layers.15.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 485 |
+
"vision_tower.vision_model.encoder.layers.15.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 486 |
+
"vision_tower.vision_model.encoder.layers.15.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 487 |
+
"vision_tower.vision_model.encoder.layers.15.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 488 |
+
"vision_tower.vision_model.encoder.layers.15.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 489 |
+
"vision_tower.vision_model.encoder.layers.15.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 490 |
+
"vision_tower.vision_model.encoder.layers.15.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 491 |
+
"vision_tower.vision_model.encoder.layers.15.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 492 |
+
"vision_tower.vision_model.encoder.layers.15.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 493 |
+
"vision_tower.vision_model.encoder.layers.15.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 494 |
+
"vision_tower.vision_model.encoder.layers.16.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 495 |
+
"vision_tower.vision_model.encoder.layers.16.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 496 |
+
"vision_tower.vision_model.encoder.layers.16.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 497 |
+
"vision_tower.vision_model.encoder.layers.16.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 498 |
+
"vision_tower.vision_model.encoder.layers.16.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 499 |
+
"vision_tower.vision_model.encoder.layers.16.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 500 |
+
"vision_tower.vision_model.encoder.layers.16.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 501 |
+
"vision_tower.vision_model.encoder.layers.16.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 502 |
+
"vision_tower.vision_model.encoder.layers.16.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 503 |
+
"vision_tower.vision_model.encoder.layers.16.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 504 |
+
"vision_tower.vision_model.encoder.layers.16.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 505 |
+
"vision_tower.vision_model.encoder.layers.16.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 506 |
+
"vision_tower.vision_model.encoder.layers.16.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 507 |
+
"vision_tower.vision_model.encoder.layers.16.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 508 |
+
"vision_tower.vision_model.encoder.layers.16.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 509 |
+
"vision_tower.vision_model.encoder.layers.16.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 510 |
+
"vision_tower.vision_model.encoder.layers.17.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 511 |
+
"vision_tower.vision_model.encoder.layers.17.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 512 |
+
"vision_tower.vision_model.encoder.layers.17.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 513 |
+
"vision_tower.vision_model.encoder.layers.17.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 514 |
+
"vision_tower.vision_model.encoder.layers.17.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 515 |
+
"vision_tower.vision_model.encoder.layers.17.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 516 |
+
"vision_tower.vision_model.encoder.layers.17.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 517 |
+
"vision_tower.vision_model.encoder.layers.17.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 518 |
+
"vision_tower.vision_model.encoder.layers.17.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 519 |
+
"vision_tower.vision_model.encoder.layers.17.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 520 |
+
"vision_tower.vision_model.encoder.layers.17.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 521 |
+
"vision_tower.vision_model.encoder.layers.17.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 522 |
+
"vision_tower.vision_model.encoder.layers.17.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 523 |
+
"vision_tower.vision_model.encoder.layers.17.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 524 |
+
"vision_tower.vision_model.encoder.layers.17.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 525 |
+
"vision_tower.vision_model.encoder.layers.17.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 526 |
+
"vision_tower.vision_model.encoder.layers.18.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 527 |
+
"vision_tower.vision_model.encoder.layers.18.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 528 |
+
"vision_tower.vision_model.encoder.layers.18.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 529 |
+
"vision_tower.vision_model.encoder.layers.18.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 530 |
+
"vision_tower.vision_model.encoder.layers.18.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 531 |
+
"vision_tower.vision_model.encoder.layers.18.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 532 |
+
"vision_tower.vision_model.encoder.layers.18.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 533 |
+
"vision_tower.vision_model.encoder.layers.18.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 534 |
+
"vision_tower.vision_model.encoder.layers.18.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 535 |
+
"vision_tower.vision_model.encoder.layers.18.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 536 |
+
"vision_tower.vision_model.encoder.layers.18.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 537 |
+
"vision_tower.vision_model.encoder.layers.18.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 538 |
+
"vision_tower.vision_model.encoder.layers.18.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 539 |
+
"vision_tower.vision_model.encoder.layers.18.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 540 |
+
"vision_tower.vision_model.encoder.layers.18.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 541 |
+
"vision_tower.vision_model.encoder.layers.18.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 542 |
+
"vision_tower.vision_model.encoder.layers.19.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 543 |
+
"vision_tower.vision_model.encoder.layers.19.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 544 |
+
"vision_tower.vision_model.encoder.layers.19.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 545 |
+
"vision_tower.vision_model.encoder.layers.19.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 546 |
+
"vision_tower.vision_model.encoder.layers.19.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 547 |
+
"vision_tower.vision_model.encoder.layers.19.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 548 |
+
"vision_tower.vision_model.encoder.layers.19.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 549 |
+
"vision_tower.vision_model.encoder.layers.19.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 550 |
+
"vision_tower.vision_model.encoder.layers.19.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 551 |
+
"vision_tower.vision_model.encoder.layers.19.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 552 |
+
"vision_tower.vision_model.encoder.layers.19.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 553 |
+
"vision_tower.vision_model.encoder.layers.19.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 554 |
+
"vision_tower.vision_model.encoder.layers.19.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 555 |
+
"vision_tower.vision_model.encoder.layers.19.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 556 |
+
"vision_tower.vision_model.encoder.layers.19.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 557 |
+
"vision_tower.vision_model.encoder.layers.19.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 558 |
+
"vision_tower.vision_model.encoder.layers.2.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 559 |
+
"vision_tower.vision_model.encoder.layers.2.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 560 |
+
"vision_tower.vision_model.encoder.layers.2.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 561 |
+
"vision_tower.vision_model.encoder.layers.2.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 562 |
+
"vision_tower.vision_model.encoder.layers.2.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 563 |
+
"vision_tower.vision_model.encoder.layers.2.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 564 |
+
"vision_tower.vision_model.encoder.layers.2.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 565 |
+
"vision_tower.vision_model.encoder.layers.2.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 566 |
+
"vision_tower.vision_model.encoder.layers.2.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 567 |
+
"vision_tower.vision_model.encoder.layers.2.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 568 |
+
"vision_tower.vision_model.encoder.layers.2.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 569 |
+
"vision_tower.vision_model.encoder.layers.2.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 570 |
+
"vision_tower.vision_model.encoder.layers.2.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 571 |
+
"vision_tower.vision_model.encoder.layers.2.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 572 |
+
"vision_tower.vision_model.encoder.layers.2.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 573 |
+
"vision_tower.vision_model.encoder.layers.2.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 574 |
+
"vision_tower.vision_model.encoder.layers.20.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 575 |
+
"vision_tower.vision_model.encoder.layers.20.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 576 |
+
"vision_tower.vision_model.encoder.layers.20.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 577 |
+
"vision_tower.vision_model.encoder.layers.20.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 578 |
+
"vision_tower.vision_model.encoder.layers.20.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 579 |
+
"vision_tower.vision_model.encoder.layers.20.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 580 |
+
"vision_tower.vision_model.encoder.layers.20.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 581 |
+
"vision_tower.vision_model.encoder.layers.20.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 582 |
+
"vision_tower.vision_model.encoder.layers.20.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 583 |
+
"vision_tower.vision_model.encoder.layers.20.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 584 |
+
"vision_tower.vision_model.encoder.layers.20.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 585 |
+
"vision_tower.vision_model.encoder.layers.20.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 586 |
+
"vision_tower.vision_model.encoder.layers.20.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 587 |
+
"vision_tower.vision_model.encoder.layers.20.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 588 |
+
"vision_tower.vision_model.encoder.layers.20.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 589 |
+
"vision_tower.vision_model.encoder.layers.20.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 590 |
+
"vision_tower.vision_model.encoder.layers.21.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 591 |
+
"vision_tower.vision_model.encoder.layers.21.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 592 |
+
"vision_tower.vision_model.encoder.layers.21.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 593 |
+
"vision_tower.vision_model.encoder.layers.21.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 594 |
+
"vision_tower.vision_model.encoder.layers.21.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 595 |
+
"vision_tower.vision_model.encoder.layers.21.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 596 |
+
"vision_tower.vision_model.encoder.layers.21.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 597 |
+
"vision_tower.vision_model.encoder.layers.21.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 598 |
+
"vision_tower.vision_model.encoder.layers.21.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 599 |
+
"vision_tower.vision_model.encoder.layers.21.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 600 |
+
"vision_tower.vision_model.encoder.layers.21.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 601 |
+
"vision_tower.vision_model.encoder.layers.21.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 602 |
+
"vision_tower.vision_model.encoder.layers.21.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 603 |
+
"vision_tower.vision_model.encoder.layers.21.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 604 |
+
"vision_tower.vision_model.encoder.layers.21.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 605 |
+
"vision_tower.vision_model.encoder.layers.21.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 606 |
+
"vision_tower.vision_model.encoder.layers.22.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 607 |
+
"vision_tower.vision_model.encoder.layers.22.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 608 |
+
"vision_tower.vision_model.encoder.layers.22.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 609 |
+
"vision_tower.vision_model.encoder.layers.22.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 610 |
+
"vision_tower.vision_model.encoder.layers.22.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 611 |
+
"vision_tower.vision_model.encoder.layers.22.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 612 |
+
"vision_tower.vision_model.encoder.layers.22.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 613 |
+
"vision_tower.vision_model.encoder.layers.22.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 614 |
+
"vision_tower.vision_model.encoder.layers.22.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 615 |
+
"vision_tower.vision_model.encoder.layers.22.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 616 |
+
"vision_tower.vision_model.encoder.layers.22.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 617 |
+
"vision_tower.vision_model.encoder.layers.22.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 618 |
+
"vision_tower.vision_model.encoder.layers.22.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 619 |
+
"vision_tower.vision_model.encoder.layers.22.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 620 |
+
"vision_tower.vision_model.encoder.layers.22.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 621 |
+
"vision_tower.vision_model.encoder.layers.22.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 622 |
+
"vision_tower.vision_model.encoder.layers.23.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 623 |
+
"vision_tower.vision_model.encoder.layers.23.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 624 |
+
"vision_tower.vision_model.encoder.layers.23.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 625 |
+
"vision_tower.vision_model.encoder.layers.23.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 626 |
+
"vision_tower.vision_model.encoder.layers.23.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 627 |
+
"vision_tower.vision_model.encoder.layers.23.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 628 |
+
"vision_tower.vision_model.encoder.layers.23.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 629 |
+
"vision_tower.vision_model.encoder.layers.23.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 630 |
+
"vision_tower.vision_model.encoder.layers.23.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 631 |
+
"vision_tower.vision_model.encoder.layers.23.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 632 |
+
"vision_tower.vision_model.encoder.layers.23.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 633 |
+
"vision_tower.vision_model.encoder.layers.23.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 634 |
+
"vision_tower.vision_model.encoder.layers.23.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 635 |
+
"vision_tower.vision_model.encoder.layers.23.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 636 |
+
"vision_tower.vision_model.encoder.layers.23.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 637 |
+
"vision_tower.vision_model.encoder.layers.23.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 638 |
+
"vision_tower.vision_model.encoder.layers.24.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 639 |
+
"vision_tower.vision_model.encoder.layers.24.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 640 |
+
"vision_tower.vision_model.encoder.layers.24.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 641 |
+
"vision_tower.vision_model.encoder.layers.24.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 642 |
+
"vision_tower.vision_model.encoder.layers.24.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 643 |
+
"vision_tower.vision_model.encoder.layers.24.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 644 |
+
"vision_tower.vision_model.encoder.layers.24.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 645 |
+
"vision_tower.vision_model.encoder.layers.24.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 646 |
+
"vision_tower.vision_model.encoder.layers.24.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 647 |
+
"vision_tower.vision_model.encoder.layers.24.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 648 |
+
"vision_tower.vision_model.encoder.layers.24.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 649 |
+
"vision_tower.vision_model.encoder.layers.24.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 650 |
+
"vision_tower.vision_model.encoder.layers.24.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 651 |
+
"vision_tower.vision_model.encoder.layers.24.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 652 |
+
"vision_tower.vision_model.encoder.layers.24.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 653 |
+
"vision_tower.vision_model.encoder.layers.24.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 654 |
+
"vision_tower.vision_model.encoder.layers.25.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 655 |
+
"vision_tower.vision_model.encoder.layers.25.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 656 |
+
"vision_tower.vision_model.encoder.layers.25.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 657 |
+
"vision_tower.vision_model.encoder.layers.25.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 658 |
+
"vision_tower.vision_model.encoder.layers.25.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 659 |
+
"vision_tower.vision_model.encoder.layers.25.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 660 |
+
"vision_tower.vision_model.encoder.layers.25.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 661 |
+
"vision_tower.vision_model.encoder.layers.25.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 662 |
+
"vision_tower.vision_model.encoder.layers.25.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 663 |
+
"vision_tower.vision_model.encoder.layers.25.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 664 |
+
"vision_tower.vision_model.encoder.layers.25.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 665 |
+
"vision_tower.vision_model.encoder.layers.25.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 666 |
+
"vision_tower.vision_model.encoder.layers.25.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 667 |
+
"vision_tower.vision_model.encoder.layers.25.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 668 |
+
"vision_tower.vision_model.encoder.layers.25.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 669 |
+
"vision_tower.vision_model.encoder.layers.25.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 670 |
+
"vision_tower.vision_model.encoder.layers.26.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 671 |
+
"vision_tower.vision_model.encoder.layers.26.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 672 |
+
"vision_tower.vision_model.encoder.layers.26.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 673 |
+
"vision_tower.vision_model.encoder.layers.26.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 674 |
+
"vision_tower.vision_model.encoder.layers.26.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 675 |
+
"vision_tower.vision_model.encoder.layers.26.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 676 |
+
"vision_tower.vision_model.encoder.layers.26.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 677 |
+
"vision_tower.vision_model.encoder.layers.26.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 678 |
+
"vision_tower.vision_model.encoder.layers.26.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 679 |
+
"vision_tower.vision_model.encoder.layers.26.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 680 |
+
"vision_tower.vision_model.encoder.layers.26.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 681 |
+
"vision_tower.vision_model.encoder.layers.26.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 682 |
+
"vision_tower.vision_model.encoder.layers.26.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 683 |
+
"vision_tower.vision_model.encoder.layers.26.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 684 |
+
"vision_tower.vision_model.encoder.layers.26.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 685 |
+
"vision_tower.vision_model.encoder.layers.26.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 686 |
+
"vision_tower.vision_model.encoder.layers.3.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 687 |
+
"vision_tower.vision_model.encoder.layers.3.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 688 |
+
"vision_tower.vision_model.encoder.layers.3.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 689 |
+
"vision_tower.vision_model.encoder.layers.3.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 690 |
+
"vision_tower.vision_model.encoder.layers.3.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 691 |
+
"vision_tower.vision_model.encoder.layers.3.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 692 |
+
"vision_tower.vision_model.encoder.layers.3.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 693 |
+
"vision_tower.vision_model.encoder.layers.3.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 694 |
+
"vision_tower.vision_model.encoder.layers.3.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 695 |
+
"vision_tower.vision_model.encoder.layers.3.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 696 |
+
"vision_tower.vision_model.encoder.layers.3.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 697 |
+
"vision_tower.vision_model.encoder.layers.3.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 698 |
+
"vision_tower.vision_model.encoder.layers.3.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 699 |
+
"vision_tower.vision_model.encoder.layers.3.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 700 |
+
"vision_tower.vision_model.encoder.layers.3.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 701 |
+
"vision_tower.vision_model.encoder.layers.3.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 702 |
+
"vision_tower.vision_model.encoder.layers.4.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 703 |
+
"vision_tower.vision_model.encoder.layers.4.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 704 |
+
"vision_tower.vision_model.encoder.layers.4.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 705 |
+
"vision_tower.vision_model.encoder.layers.4.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 706 |
+
"vision_tower.vision_model.encoder.layers.4.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 707 |
+
"vision_tower.vision_model.encoder.layers.4.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 708 |
+
"vision_tower.vision_model.encoder.layers.4.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 709 |
+
"vision_tower.vision_model.encoder.layers.4.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 710 |
+
"vision_tower.vision_model.encoder.layers.4.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 711 |
+
"vision_tower.vision_model.encoder.layers.4.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 712 |
+
"vision_tower.vision_model.encoder.layers.4.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 713 |
+
"vision_tower.vision_model.encoder.layers.4.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 714 |
+
"vision_tower.vision_model.encoder.layers.4.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 715 |
+
"vision_tower.vision_model.encoder.layers.4.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 716 |
+
"vision_tower.vision_model.encoder.layers.4.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 717 |
+
"vision_tower.vision_model.encoder.layers.4.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 718 |
+
"vision_tower.vision_model.encoder.layers.5.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 719 |
+
"vision_tower.vision_model.encoder.layers.5.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 720 |
+
"vision_tower.vision_model.encoder.layers.5.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 721 |
+
"vision_tower.vision_model.encoder.layers.5.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 722 |
+
"vision_tower.vision_model.encoder.layers.5.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 723 |
+
"vision_tower.vision_model.encoder.layers.5.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 724 |
+
"vision_tower.vision_model.encoder.layers.5.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 725 |
+
"vision_tower.vision_model.encoder.layers.5.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 726 |
+
"vision_tower.vision_model.encoder.layers.5.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 727 |
+
"vision_tower.vision_model.encoder.layers.5.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 728 |
+
"vision_tower.vision_model.encoder.layers.5.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 729 |
+
"vision_tower.vision_model.encoder.layers.5.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 730 |
+
"vision_tower.vision_model.encoder.layers.5.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 731 |
+
"vision_tower.vision_model.encoder.layers.5.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 732 |
+
"vision_tower.vision_model.encoder.layers.5.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 733 |
+
"vision_tower.vision_model.encoder.layers.5.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 734 |
+
"vision_tower.vision_model.encoder.layers.6.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 735 |
+
"vision_tower.vision_model.encoder.layers.6.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 736 |
+
"vision_tower.vision_model.encoder.layers.6.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 737 |
+
"vision_tower.vision_model.encoder.layers.6.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 738 |
+
"vision_tower.vision_model.encoder.layers.6.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 739 |
+
"vision_tower.vision_model.encoder.layers.6.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 740 |
+
"vision_tower.vision_model.encoder.layers.6.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 741 |
+
"vision_tower.vision_model.encoder.layers.6.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 742 |
+
"vision_tower.vision_model.encoder.layers.6.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 743 |
+
"vision_tower.vision_model.encoder.layers.6.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 744 |
+
"vision_tower.vision_model.encoder.layers.6.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 745 |
+
"vision_tower.vision_model.encoder.layers.6.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 746 |
+
"vision_tower.vision_model.encoder.layers.6.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 747 |
+
"vision_tower.vision_model.encoder.layers.6.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 748 |
+
"vision_tower.vision_model.encoder.layers.6.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 749 |
+
"vision_tower.vision_model.encoder.layers.6.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 750 |
+
"vision_tower.vision_model.encoder.layers.7.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 751 |
+
"vision_tower.vision_model.encoder.layers.7.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 752 |
+
"vision_tower.vision_model.encoder.layers.7.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 753 |
+
"vision_tower.vision_model.encoder.layers.7.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 754 |
+
"vision_tower.vision_model.encoder.layers.7.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 755 |
+
"vision_tower.vision_model.encoder.layers.7.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 756 |
+
"vision_tower.vision_model.encoder.layers.7.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 757 |
+
"vision_tower.vision_model.encoder.layers.7.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 758 |
+
"vision_tower.vision_model.encoder.layers.7.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 759 |
+
"vision_tower.vision_model.encoder.layers.7.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 760 |
+
"vision_tower.vision_model.encoder.layers.7.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 761 |
+
"vision_tower.vision_model.encoder.layers.7.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 762 |
+
"vision_tower.vision_model.encoder.layers.7.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 763 |
+
"vision_tower.vision_model.encoder.layers.7.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 764 |
+
"vision_tower.vision_model.encoder.layers.7.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 765 |
+
"vision_tower.vision_model.encoder.layers.7.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 766 |
+
"vision_tower.vision_model.encoder.layers.8.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 767 |
+
"vision_tower.vision_model.encoder.layers.8.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 768 |
+
"vision_tower.vision_model.encoder.layers.8.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 769 |
+
"vision_tower.vision_model.encoder.layers.8.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 770 |
+
"vision_tower.vision_model.encoder.layers.8.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 771 |
+
"vision_tower.vision_model.encoder.layers.8.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 772 |
+
"vision_tower.vision_model.encoder.layers.8.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 773 |
+
"vision_tower.vision_model.encoder.layers.8.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 774 |
+
"vision_tower.vision_model.encoder.layers.8.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 775 |
+
"vision_tower.vision_model.encoder.layers.8.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 776 |
+
"vision_tower.vision_model.encoder.layers.8.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 777 |
+
"vision_tower.vision_model.encoder.layers.8.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 778 |
+
"vision_tower.vision_model.encoder.layers.8.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 779 |
+
"vision_tower.vision_model.encoder.layers.8.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 780 |
+
"vision_tower.vision_model.encoder.layers.8.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 781 |
+
"vision_tower.vision_model.encoder.layers.8.self_attn.v_proj.weight": "model-00001-of-00012.safetensors",
|
| 782 |
+
"vision_tower.vision_model.encoder.layers.9.layer_norm1.bias": "model-00001-of-00012.safetensors",
|
| 783 |
+
"vision_tower.vision_model.encoder.layers.9.layer_norm1.weight": "model-00001-of-00012.safetensors",
|
| 784 |
+
"vision_tower.vision_model.encoder.layers.9.layer_norm2.bias": "model-00001-of-00012.safetensors",
|
| 785 |
+
"vision_tower.vision_model.encoder.layers.9.layer_norm2.weight": "model-00001-of-00012.safetensors",
|
| 786 |
+
"vision_tower.vision_model.encoder.layers.9.mlp.fc1.bias": "model-00001-of-00012.safetensors",
|
| 787 |
+
"vision_tower.vision_model.encoder.layers.9.mlp.fc1.weight": "model-00001-of-00012.safetensors",
|
| 788 |
+
"vision_tower.vision_model.encoder.layers.9.mlp.fc2.bias": "model-00001-of-00012.safetensors",
|
| 789 |
+
"vision_tower.vision_model.encoder.layers.9.mlp.fc2.weight": "model-00001-of-00012.safetensors",
|
| 790 |
+
"vision_tower.vision_model.encoder.layers.9.self_attn.k_proj.bias": "model-00001-of-00012.safetensors",
|
| 791 |
+
"vision_tower.vision_model.encoder.layers.9.self_attn.k_proj.weight": "model-00001-of-00012.safetensors",
|
| 792 |
+
"vision_tower.vision_model.encoder.layers.9.self_attn.out_proj.bias": "model-00001-of-00012.safetensors",
|
| 793 |
+
"vision_tower.vision_model.encoder.layers.9.self_attn.out_proj.weight": "model-00001-of-00012.safetensors",
|
| 794 |
+
"vision_tower.vision_model.encoder.layers.9.self_attn.q_proj.bias": "model-00001-of-00012.safetensors",
|
| 795 |
+
"vision_tower.vision_model.encoder.layers.9.self_attn.q_proj.weight": "model-00001-of-00012.safetensors",
|
| 796 |
+
"vision_tower.vision_model.encoder.layers.9.self_attn.v_proj.bias": "model-00001-of-00012.safetensors",
|
| 797 |
+
"vision_tower.vision_model.encoder.layers.9.self_attn.v_proj.weight": "model-00001-of-00012.safetensors"
|
| 798 |
+
}
|
| 799 |
+
}
|
modeling_aria.py
ADDED
|
@@ -0,0 +1,365 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright 2024 Rhymes AI. All rights reserved.
|
| 2 |
+
#
|
| 3 |
+
# Licensed to the Apache Software Foundation (ASF) under one
|
| 4 |
+
# or more contributor license agreements. See the NOTICE file
|
| 5 |
+
# distributed with this work for additional information
|
| 6 |
+
# regarding copyright ownership. The ASF licenses this file
|
| 7 |
+
# to you under the Apache License, Version 2.0 (the
|
| 8 |
+
# "License"); you may not use this file except in compliance
|
| 9 |
+
# with the License. You may obtain a copy of the License at
|
| 10 |
+
#
|
| 11 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 12 |
+
#
|
| 13 |
+
# Unless required by applicable law or agreed to in writing,
|
| 14 |
+
# software distributed under the License is distributed on an
|
| 15 |
+
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
| 16 |
+
# KIND, either express or implied. See the License for the
|
| 17 |
+
# specific language governing permissions and limitations
|
| 18 |
+
# under the License.
|
| 19 |
+
|
| 20 |
+
from dataclasses import dataclass
|
| 21 |
+
from typing import List, Optional, Tuple, Union
|
| 22 |
+
|
| 23 |
+
import torch
|
| 24 |
+
import torch.nn as nn
|
| 25 |
+
from torch import nn
|
| 26 |
+
from transformers import GenerationMixin, PreTrainedModel
|
| 27 |
+
from transformers.modeling_outputs import ModelOutput
|
| 28 |
+
from transformers.utils import logging
|
| 29 |
+
|
| 30 |
+
from .configuration_aria import AriaConfig
|
| 31 |
+
from .moe_lm import AriaMoELMForCausalLM
|
| 32 |
+
from .projector import AriaProjector
|
| 33 |
+
from .vision_encoder import AriaVisionModel
|
| 34 |
+
|
| 35 |
+
logger = logging.get_logger(__name__)
|
| 36 |
+
|
| 37 |
+
|
| 38 |
+
class AriaPretrainedModel(PreTrainedModel):
|
| 39 |
+
"""
|
| 40 |
+
An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained models.
|
| 41 |
+
"""
|
| 42 |
+
|
| 43 |
+
config_class = AriaConfig
|
| 44 |
+
base_model_prefix = "model"
|
| 45 |
+
_no_split_modules = []
|
| 46 |
+
supports_gradient_checkpointing = True
|
| 47 |
+
_skip_keys_device_placement = "past_key_values"
|
| 48 |
+
_supports_flash_attn_2 = True
|
| 49 |
+
_supports_cache_class = True
|
| 50 |
+
_supports_static_cache = True
|
| 51 |
+
|
| 52 |
+
@property
|
| 53 |
+
def _supports_sdpa(self):
|
| 54 |
+
"""
|
| 55 |
+
Retrieve language_model's attribute to check whether the model supports
|
| 56 |
+
SDPA (Scaled Dot Product Attention) or not.
|
| 57 |
+
"""
|
| 58 |
+
return self.language_model._supports_sdpa
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
@dataclass
|
| 62 |
+
# Copied from transformers.models.llava.modeling_llava.LlavaCausalLMOutputWithPast with Llava->Aria
|
| 63 |
+
class AriaCausalLMOutputWithPast(ModelOutput):
|
| 64 |
+
"""
|
| 65 |
+
Base class for Aria causal language model (or autoregressive) outputs.
|
| 66 |
+
|
| 67 |
+
Args:
|
| 68 |
+
loss (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` is provided):
|
| 69 |
+
Language modeling loss (for next-token prediction).
|
| 70 |
+
logits (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.vocab_size)`):
|
| 71 |
+
Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
|
| 72 |
+
past_key_values (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`):
|
| 73 |
+
Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of shape
|
| 74 |
+
`(batch_size, num_heads, sequence_length, embed_size_per_head)`)
|
| 75 |
+
|
| 76 |
+
Contains pre-computed hidden-states (key and values in the self-attention blocks) that can be used (see
|
| 77 |
+
`past_key_values` input) to speed up sequential decoding.
|
| 78 |
+
hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
|
| 79 |
+
Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
|
| 80 |
+
one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.
|
| 81 |
+
|
| 82 |
+
Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
|
| 83 |
+
attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
|
| 84 |
+
Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
|
| 85 |
+
sequence_length)`.
|
| 86 |
+
|
| 87 |
+
Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
|
| 88 |
+
heads.
|
| 89 |
+
image_hidden_states (`tuple(torch.FloatTensor)`, *optional*):
|
| 90 |
+
Tuple of `torch.FloatTensor` (one for the output of the image embeddings, `(batch_size, num_images,
|
| 91 |
+
sequence_length, hidden_size)`.
|
| 92 |
+
|
| 93 |
+
image_hidden_states of the model produced by the vision encoder, and optionally by the perceiver
|
| 94 |
+
"""
|
| 95 |
+
|
| 96 |
+
loss: Optional[torch.FloatTensor] = None
|
| 97 |
+
logits: torch.FloatTensor = None
|
| 98 |
+
past_key_values: Optional[List[torch.FloatTensor]] = None
|
| 99 |
+
hidden_states: Optional[Tuple[torch.FloatTensor]] = None
|
| 100 |
+
attentions: Optional[Tuple[torch.FloatTensor]] = None
|
| 101 |
+
image_hidden_states: Optional[Tuple[torch.FloatTensor]] = None
|
| 102 |
+
|
| 103 |
+
|
| 104 |
+
def build_mm_projector(config: AriaConfig):
|
| 105 |
+
"""
|
| 106 |
+
Builds and returns an AriaProjector instance based on the provided configuration.
|
| 107 |
+
|
| 108 |
+
Args:
|
| 109 |
+
config (AriaConfig): The configuration object containing necessary parameters.
|
| 110 |
+
|
| 111 |
+
Returns:
|
| 112 |
+
AriaProjector: An instance of the AriaProjector class.
|
| 113 |
+
"""
|
| 114 |
+
return AriaProjector(
|
| 115 |
+
patch_to_query_dict=config.projector_patch_to_query_dict,
|
| 116 |
+
embed_dim=config.vision_config.hidden_size,
|
| 117 |
+
num_heads=config.vision_config.num_attention_heads,
|
| 118 |
+
kv_dim=config.vision_config.hidden_size,
|
| 119 |
+
ff_dim=config.text_config.hidden_size,
|
| 120 |
+
output_dim=config.text_config.hidden_size,
|
| 121 |
+
)
|
| 122 |
+
|
| 123 |
+
|
| 124 |
+
# adapted from transformers.models.llava.modeling_llava.LlavaForConditionalGeneration
|
| 125 |
+
class AriaForConditionalGeneration(AriaPretrainedModel, GenerationMixin):
|
| 126 |
+
"""
|
| 127 |
+
Aria model for conditional generation tasks.
|
| 128 |
+
|
| 129 |
+
This model combines a vision tower, a multi-modal projector, and a language model
|
| 130 |
+
to perform tasks that involve both image and text inputs.
|
| 131 |
+
"""
|
| 132 |
+
|
| 133 |
+
def __init__(self, config: AriaConfig):
|
| 134 |
+
super().__init__(config)
|
| 135 |
+
|
| 136 |
+
self.vision_tower = AriaVisionModel(config.vision_config)
|
| 137 |
+
self.multi_modal_projector = build_mm_projector(config)
|
| 138 |
+
self.vocab_size = config.text_config.vocab_size
|
| 139 |
+
self.language_model = AriaMoELMForCausalLM(config.text_config)
|
| 140 |
+
self.pad_token_id = (
|
| 141 |
+
self.config.pad_token_id if self.config.pad_token_id is not None else -1
|
| 142 |
+
)
|
| 143 |
+
self.post_init()
|
| 144 |
+
|
| 145 |
+
def freeze_vit(self):
|
| 146 |
+
"""Freeze the parameters of the vision tower."""
|
| 147 |
+
for param in self.vision_tower.parameters():
|
| 148 |
+
param.requires_grad = False
|
| 149 |
+
|
| 150 |
+
def freeze_projector(self):
|
| 151 |
+
"""Freeze the parameters of the multi-modal projector."""
|
| 152 |
+
for param in self.multi_modal_projector.parameters():
|
| 153 |
+
param.requires_grad = False
|
| 154 |
+
|
| 155 |
+
def freeze_llm(self):
|
| 156 |
+
"""Freeze the parameters of the language model."""
|
| 157 |
+
for param in self.language_model.parameters():
|
| 158 |
+
param.requires_grad = False
|
| 159 |
+
|
| 160 |
+
def get_input_embeddings(self) -> nn.Module:
|
| 161 |
+
"""Retrieve the input embeddings from the language model."""
|
| 162 |
+
return self.language_model.get_input_embeddings()
|
| 163 |
+
|
| 164 |
+
def set_input_embeddings(self, value):
|
| 165 |
+
"""Set the input embeddings for the language model."""
|
| 166 |
+
self.language_model.set_input_embeddings(value)
|
| 167 |
+
|
| 168 |
+
def get_output_embeddings(self):
|
| 169 |
+
"""Retrieve the output embeddings from the language model."""
|
| 170 |
+
return self.language_model.get_output_embeddings()
|
| 171 |
+
|
| 172 |
+
def set_output_embeddings(self, value):
|
| 173 |
+
"""Set the output embeddings for the language model."""
|
| 174 |
+
self.language_model.set_output_embeddings(value)
|
| 175 |
+
|
| 176 |
+
def set_moe_z_loss_coeff(self, value):
|
| 177 |
+
"""
|
| 178 |
+
Set the z-loss coefficient for Mixture of Experts (MoE) models.
|
| 179 |
+
|
| 180 |
+
Args:
|
| 181 |
+
value: The z-loss coefficient value to set.
|
| 182 |
+
"""
|
| 183 |
+
self.language_model.set_z_loss_coeff(value)
|
| 184 |
+
|
| 185 |
+
def set_moe_aux_loss_coeff(self, value):
|
| 186 |
+
"""
|
| 187 |
+
Set the auxiliary loss coefficient for Mixture of Experts (MoE) models.
|
| 188 |
+
|
| 189 |
+
Args:
|
| 190 |
+
value: The auxiliary loss coefficient value to set.
|
| 191 |
+
"""
|
| 192 |
+
self.language_model.set_aux_loss_coeff(value)
|
| 193 |
+
|
| 194 |
+
def forward(
|
| 195 |
+
self,
|
| 196 |
+
input_ids: torch.LongTensor = None,
|
| 197 |
+
pixel_values: torch.FloatTensor = None,
|
| 198 |
+
pixel_mask: torch.LongTensor = None,
|
| 199 |
+
attention_mask: Optional[torch.Tensor] = None,
|
| 200 |
+
position_ids: Optional[torch.LongTensor] = None,
|
| 201 |
+
past_key_values: Optional[List[torch.FloatTensor]] = None,
|
| 202 |
+
inputs_embeds: Optional[torch.FloatTensor] = None,
|
| 203 |
+
labels: Optional[torch.LongTensor] = None,
|
| 204 |
+
use_cache: Optional[bool] = None,
|
| 205 |
+
output_attentions: Optional[bool] = None,
|
| 206 |
+
output_hidden_states: Optional[bool] = None,
|
| 207 |
+
return_dict: Optional[bool] = None,
|
| 208 |
+
cache_position: Optional[torch.LongTensor] = None,
|
| 209 |
+
num_logits_to_keep: int = 0,
|
| 210 |
+
) -> Union[Tuple, AriaCausalLMOutputWithPast]:
|
| 211 |
+
"""
|
| 212 |
+
Forward pass of the AriaForConditionalGeneration model.
|
| 213 |
+
|
| 214 |
+
This method processes both text and image inputs, merges them if necessary,
|
| 215 |
+
and generates output using the language model.
|
| 216 |
+
|
| 217 |
+
Args:
|
| 218 |
+
input_ids (torch.LongTensor, optional): Input token ids.
|
| 219 |
+
pixel_values (torch.FloatTensor, optional): Pixel values of the images.
|
| 220 |
+
pixel_mask (torch.LongTensor, optional): Mask for the pixel values.
|
| 221 |
+
attention_mask (torch.Tensor, optional): Attention mask.
|
| 222 |
+
position_ids (torch.LongTensor, optional): Position ids.
|
| 223 |
+
past_key_values (List[torch.FloatTensor], optional): Past key values for efficient processing.
|
| 224 |
+
inputs_embeds (torch.FloatTensor, optional): Input embeddings.
|
| 225 |
+
labels (torch.LongTensor, optional): Labels for computing the language modeling loss.
|
| 226 |
+
use_cache (bool, optional): Whether to use the model's cache mechanism.
|
| 227 |
+
output_attentions (bool, optional): Whether to output attention weights.
|
| 228 |
+
output_hidden_states (bool, optional): Whether to output hidden states.
|
| 229 |
+
return_dict (bool, optional): Whether to return a ModelOutput object.
|
| 230 |
+
|
| 231 |
+
Returns:
|
| 232 |
+
Union[Tuple, AriaCausalLMOutputWithPast]: Model outputs.
|
| 233 |
+
"""
|
| 234 |
+
output_attentions = (
|
| 235 |
+
output_attentions
|
| 236 |
+
if output_attentions is not None
|
| 237 |
+
else self.config.output_attentions
|
| 238 |
+
)
|
| 239 |
+
output_hidden_states = (
|
| 240 |
+
output_hidden_states
|
| 241 |
+
if output_hidden_states is not None
|
| 242 |
+
else self.config.output_hidden_states
|
| 243 |
+
)
|
| 244 |
+
return_dict = (
|
| 245 |
+
return_dict if return_dict is not None else self.config.use_return_dict
|
| 246 |
+
)
|
| 247 |
+
|
| 248 |
+
if inputs_embeds is None:
|
| 249 |
+
# 1. Extra the input embeddings
|
| 250 |
+
inputs_embeds = self.get_input_embeddings()(input_ids)
|
| 251 |
+
|
| 252 |
+
image_features = None
|
| 253 |
+
if pixel_values is not None:
|
| 254 |
+
image_outputs, image_attn_mask = self.vision_tower(
|
| 255 |
+
pixel_values,
|
| 256 |
+
pixel_mask=pixel_mask,
|
| 257 |
+
)
|
| 258 |
+
|
| 259 |
+
selected_image_feature = image_outputs.last_hidden_state
|
| 260 |
+
image_features = self.multi_modal_projector(
|
| 261 |
+
selected_image_feature, attn_mask=image_attn_mask
|
| 262 |
+
)
|
| 263 |
+
|
| 264 |
+
if image_features is not None:
|
| 265 |
+
n_image_tokens = (input_ids == self.config.image_token_index).sum().item()
|
| 266 |
+
n_image_features = image_features.shape[0] * image_features.shape[1]
|
| 267 |
+
|
| 268 |
+
if n_image_tokens != n_image_features:
|
| 269 |
+
raise ValueError(
|
| 270 |
+
f"Image features and image tokens do not match: tokens: {n_image_tokens}, features {n_image_features}"
|
| 271 |
+
)
|
| 272 |
+
special_image_mask = (
|
| 273 |
+
(input_ids == self.config.image_token_index)
|
| 274 |
+
.unsqueeze(-1)
|
| 275 |
+
.expand_as(inputs_embeds)
|
| 276 |
+
.to(inputs_embeds.device)
|
| 277 |
+
)
|
| 278 |
+
image_features = image_features.to(
|
| 279 |
+
inputs_embeds.device, inputs_embeds.dtype
|
| 280 |
+
)
|
| 281 |
+
inputs_embeds = inputs_embeds.masked_scatter(
|
| 282 |
+
special_image_mask, image_features
|
| 283 |
+
)
|
| 284 |
+
|
| 285 |
+
outputs = self.language_model(
|
| 286 |
+
attention_mask=attention_mask,
|
| 287 |
+
position_ids=position_ids,
|
| 288 |
+
past_key_values=past_key_values,
|
| 289 |
+
inputs_embeds=inputs_embeds,
|
| 290 |
+
use_cache=use_cache,
|
| 291 |
+
output_attentions=output_attentions,
|
| 292 |
+
output_hidden_states=output_hidden_states,
|
| 293 |
+
return_dict=return_dict,
|
| 294 |
+
cache_position=cache_position,
|
| 295 |
+
num_logits_to_keep=num_logits_to_keep,
|
| 296 |
+
)
|
| 297 |
+
|
| 298 |
+
logits = outputs[0]
|
| 299 |
+
|
| 300 |
+
loss = None
|
| 301 |
+
if labels is not None:
|
| 302 |
+
# Shift so that tokens < n predict n
|
| 303 |
+
if attention_mask is not None:
|
| 304 |
+
# we use the input attention mask to shift the logits and labels, because it is 2D.
|
| 305 |
+
# we also crop attn mask in case it is longer, which happens in PrefixTuning with peft
|
| 306 |
+
shift_attention_mask = attention_mask[:, -(logits.shape[1] - 1) :].to(
|
| 307 |
+
logits.device
|
| 308 |
+
)
|
| 309 |
+
shift_logits = logits[..., :-1, :][
|
| 310 |
+
shift_attention_mask.to(logits.device) != 0
|
| 311 |
+
].contiguous()
|
| 312 |
+
shift_labels = labels[..., 1:][
|
| 313 |
+
shift_attention_mask.to(labels.device) != 0
|
| 314 |
+
].contiguous()
|
| 315 |
+
else:
|
| 316 |
+
shift_logits = logits[..., :-1, :].contiguous()
|
| 317 |
+
shift_labels = labels[..., 1:].contiguous()
|
| 318 |
+
# Flatten the tokens
|
| 319 |
+
loss_fct = nn.CrossEntropyLoss()
|
| 320 |
+
loss = loss_fct(
|
| 321 |
+
shift_logits.view(-1, shift_logits.size(-1)),
|
| 322 |
+
shift_labels.view(-1).to(shift_logits.device),
|
| 323 |
+
)
|
| 324 |
+
|
| 325 |
+
if not return_dict:
|
| 326 |
+
output = (logits,) + outputs[1:]
|
| 327 |
+
return (loss,) + output if loss is not None else output
|
| 328 |
+
|
| 329 |
+
return AriaCausalLMOutputWithPast(
|
| 330 |
+
loss=loss,
|
| 331 |
+
logits=logits,
|
| 332 |
+
past_key_values=outputs.past_key_values,
|
| 333 |
+
hidden_states=outputs.hidden_states,
|
| 334 |
+
attentions=outputs.attentions,
|
| 335 |
+
)
|
| 336 |
+
|
| 337 |
+
def prepare_inputs_for_generation(
|
| 338 |
+
self,
|
| 339 |
+
input_ids,
|
| 340 |
+
past_key_values=None,
|
| 341 |
+
inputs_embeds=None,
|
| 342 |
+
pixel_values=None,
|
| 343 |
+
pixel_mask=None,
|
| 344 |
+
attention_mask=None,
|
| 345 |
+
cache_position=None,
|
| 346 |
+
num_logits_to_keep=None,
|
| 347 |
+
**kwargs,
|
| 348 |
+
):
|
| 349 |
+
model_inputs = self.language_model.prepare_inputs_for_generation(
|
| 350 |
+
input_ids,
|
| 351 |
+
past_key_values=past_key_values,
|
| 352 |
+
inputs_embeds=inputs_embeds,
|
| 353 |
+
attention_mask=attention_mask,
|
| 354 |
+
cache_position=cache_position,
|
| 355 |
+
num_logits_to_keep=num_logits_to_keep,
|
| 356 |
+
**kwargs,
|
| 357 |
+
)
|
| 358 |
+
|
| 359 |
+
if cache_position[0] == 0:
|
| 360 |
+
# If we're in cached decoding stage, pixel values should be None because input ids do not contain special image token anymore
|
| 361 |
+
# Otherwise we need pixel values to be passed to model
|
| 362 |
+
model_inputs["pixel_values"] = pixel_values
|
| 363 |
+
model_inputs["pixel_mask"] = pixel_mask
|
| 364 |
+
|
| 365 |
+
return model_inputs
|
moe_lm.py
ADDED
|
@@ -0,0 +1,679 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright 2024 Rhymes AI. All rights reserved.
|
| 2 |
+
#
|
| 3 |
+
# Licensed to the Apache Software Foundation (ASF) under one
|
| 4 |
+
# or more contributor license agreements. See the NOTICE file
|
| 5 |
+
# distributed with this work for additional information
|
| 6 |
+
# regarding copyright ownership. The ASF licenses this file
|
| 7 |
+
# to you under the Apache License, Version 2.0 (the
|
| 8 |
+
# "License"); you may not use this file except in compliance
|
| 9 |
+
# with the License. You may obtain a copy of the License at
|
| 10 |
+
#
|
| 11 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 12 |
+
#
|
| 13 |
+
# Unless required by applicable law or agreed to in writing,
|
| 14 |
+
# software distributed under the License is distributed on an
|
| 15 |
+
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
| 16 |
+
# KIND, either express or implied. See the License for the
|
| 17 |
+
# specific language governing permissions and limitations
|
| 18 |
+
# under the License.
|
| 19 |
+
|
| 20 |
+
import logging
|
| 21 |
+
import os
|
| 22 |
+
from typing import Tuple
|
| 23 |
+
|
| 24 |
+
import torch
|
| 25 |
+
import torch.nn as nn
|
| 26 |
+
import torch.nn.functional as F
|
| 27 |
+
from torch import nn
|
| 28 |
+
from transformers import GenerationMixin, LlamaConfig
|
| 29 |
+
from transformers.models.llama.modeling_llama import (
|
| 30 |
+
ACT2FN,
|
| 31 |
+
LLAMA_ATTENTION_CLASSES,
|
| 32 |
+
LlamaDecoderLayer,
|
| 33 |
+
LlamaForCausalLM,
|
| 34 |
+
LlamaMLP,
|
| 35 |
+
LlamaModel,
|
| 36 |
+
LlamaRMSNorm,
|
| 37 |
+
LlamaRotaryEmbedding,
|
| 38 |
+
)
|
| 39 |
+
|
| 40 |
+
logger = logging.getLogger(__name__)
|
| 41 |
+
|
| 42 |
+
|
| 43 |
+
class AriaMoELMConfig(LlamaConfig):
|
| 44 |
+
"""
|
| 45 |
+
Configuration class for AriaMoE language model.
|
| 46 |
+
|
| 47 |
+
This class extends the LlamaConfig to include additional parameters specific to the Mixture of Experts (MoE) architecture.
|
| 48 |
+
"""
|
| 49 |
+
|
| 50 |
+
model_type = "aria_moe_lm"
|
| 51 |
+
|
| 52 |
+
def __init__(
|
| 53 |
+
self,
|
| 54 |
+
moe_intermediate_size: int = 4096,
|
| 55 |
+
moe_num_experts: int = 8,
|
| 56 |
+
moe_topk: int = 2,
|
| 57 |
+
moe_z_loss_coeff: float = 1e-5,
|
| 58 |
+
moe_aux_loss_coeff: float = 1e-3,
|
| 59 |
+
moe_num_shared_experts: int = 2,
|
| 60 |
+
**kwargs,
|
| 61 |
+
):
|
| 62 |
+
"""
|
| 63 |
+
Initialize the AriaMoELMConfig.
|
| 64 |
+
|
| 65 |
+
Args:
|
| 66 |
+
moe_intermediate_size (int): The intermediate size for MoE layers. Default is 4096.
|
| 67 |
+
moe_num_experts (int): The number of experts in the MoE layer. Default is 8.
|
| 68 |
+
moe_topk (int): The number of top experts to route to for each token. Default is 2.
|
| 69 |
+
moe_z_loss_coeff (float): The coefficient for the auxiliary z-loss. Default is 1e-5.
|
| 70 |
+
moe_aux_loss_coeff (float): The coefficient for the auxiliary load balancing loss. Default is 1e-3.
|
| 71 |
+
moe_num_shared_experts (int): The number of shared experts. Default is 2.
|
| 72 |
+
**kwargs: Additional keyword arguments to be passed to the parent LlamaConfig.
|
| 73 |
+
"""
|
| 74 |
+
super().__init__(**kwargs)
|
| 75 |
+
self.moe_intermediate_size = moe_intermediate_size
|
| 76 |
+
self.moe_num_experts = moe_num_experts
|
| 77 |
+
self.moe_topk = moe_topk
|
| 78 |
+
self.moe_z_loss_coeff = moe_z_loss_coeff
|
| 79 |
+
self.moe_aux_loss_coeff = moe_aux_loss_coeff
|
| 80 |
+
self.moe_num_shared_experts = moe_num_shared_experts
|
| 81 |
+
|
| 82 |
+
|
| 83 |
+
# copied from https://github.com/NVIDIA/Megatron-LM/blob/54f1f78529cbc2b9cddad313e7f9d96ac0420a27/megatron/core/transformer/moe/moe_utils.py#L101-L142
|
| 84 |
+
class MoEAuxLossAutoScaler(torch.autograd.Function):
|
| 85 |
+
"""An AutoScaler that compute and scales the grad for auxiliary loss."""
|
| 86 |
+
|
| 87 |
+
main_loss_backward_scale: torch.Tensor = torch.tensor(1.0)
|
| 88 |
+
|
| 89 |
+
@staticmethod
|
| 90 |
+
def forward(ctx, output: torch.Tensor, aux_loss: torch.Tensor):
|
| 91 |
+
"""Preserve the aux_loss by storing it in the context to avoid garbage collection.
|
| 92 |
+
|
| 93 |
+
Args:
|
| 94 |
+
output (torch.Tensor): The output tensor.
|
| 95 |
+
aux_loss (torch.Tensor): The auxiliary loss tensor.
|
| 96 |
+
|
| 97 |
+
Returns:
|
| 98 |
+
torch.Tensor: The output tensor.
|
| 99 |
+
"""
|
| 100 |
+
ctx.save_for_backward(aux_loss)
|
| 101 |
+
return output
|
| 102 |
+
|
| 103 |
+
@staticmethod
|
| 104 |
+
def backward(ctx, grad_output: torch.Tensor):
|
| 105 |
+
"""Compute and scale the gradient for auxiliary loss..
|
| 106 |
+
|
| 107 |
+
Args:
|
| 108 |
+
grad_output (torch.Tensor): The gradient of the output.
|
| 109 |
+
|
| 110 |
+
Returns:
|
| 111 |
+
Tuple[torch.Tensor, torch.Tensor]: The gradient of the output, scaled auxiliary loss gradient.
|
| 112 |
+
"""
|
| 113 |
+
(aux_loss,) = ctx.saved_tensors
|
| 114 |
+
aux_loss_backward_scale = MoEAuxLossAutoScaler.main_loss_backward_scale
|
| 115 |
+
scaled_aux_loss_grad = torch.ones_like(aux_loss) * aux_loss_backward_scale
|
| 116 |
+
return grad_output, scaled_aux_loss_grad
|
| 117 |
+
|
| 118 |
+
@staticmethod
|
| 119 |
+
def set_loss_scale(scale: torch.Tensor):
|
| 120 |
+
"""set the scale of the aux loss.
|
| 121 |
+
|
| 122 |
+
Args:
|
| 123 |
+
scale (torch.Tensor): The scale value to set. Please ensure that the scale passed in matches the scale of the main_loss.
|
| 124 |
+
"""
|
| 125 |
+
MoEAuxLossAutoScaler.main_loss_backward_scale = scale
|
| 126 |
+
|
| 127 |
+
|
| 128 |
+
def z_loss_func(logits, z_loss_coeff):
|
| 129 |
+
"""Encourages the router's logits to remain small to enhance stability.
|
| 130 |
+
Please refer to the ST-MoE paper (https://arxiv.org/pdf/2202.08906.pdf) for details.
|
| 131 |
+
|
| 132 |
+
Args:
|
| 133 |
+
logits (torch.Tensor): The logits of the router.
|
| 134 |
+
|
| 135 |
+
Returns:
|
| 136 |
+
torch.Tensor: The logits after applying the z-loss.
|
| 137 |
+
"""
|
| 138 |
+
|
| 139 |
+
z_loss = torch.mean(torch.square(torch.logsumexp(logits, dim=-1))) * z_loss_coeff
|
| 140 |
+
return z_loss
|
| 141 |
+
|
| 142 |
+
|
| 143 |
+
def switch_load_balancing_loss_func(
|
| 144 |
+
probs: torch.Tensor,
|
| 145 |
+
tokens_per_expert: torch.Tensor,
|
| 146 |
+
topk: int,
|
| 147 |
+
moe_aux_loss_coeff: float,
|
| 148 |
+
):
|
| 149 |
+
"""Calculate the auxiliary loss for better load balancing.
|
| 150 |
+
Please refer to the Switch Transformer paper (https://arxiv.org/abs/2101.03961) for details.
|
| 151 |
+
|
| 152 |
+
Args:
|
| 153 |
+
probs (torch.Tensor): The softmax probs output by the router for each token. [num_tokens, num_experts]
|
| 154 |
+
tokens_per_expert (torch.Tensor): The number of assigned tokens for each expert. [num_experts]
|
| 155 |
+
|
| 156 |
+
Returns:
|
| 157 |
+
torch.Tensor: The auxiliary loss for load balancing.
|
| 158 |
+
"""
|
| 159 |
+
num_tokens = probs.shape[0] * topk
|
| 160 |
+
num_experts = probs.shape[1]
|
| 161 |
+
|
| 162 |
+
probs_mean_per_expert = probs.mean(dim=0)
|
| 163 |
+
aux_loss = torch.sum(probs_mean_per_expert * tokens_per_expert) * (
|
| 164 |
+
num_experts / num_tokens * moe_aux_loss_coeff
|
| 165 |
+
)
|
| 166 |
+
return aux_loss
|
| 167 |
+
|
| 168 |
+
|
| 169 |
+
# adapted from https://github.com/NVIDIA/Megatron-LM/blob/54f1f78529cbc2b9cddad313e7f9d96ac0420a27/megatron/core/transformer/moe/router.py#L96-L304
|
| 170 |
+
class TopKRouter(nn.Module):
|
| 171 |
+
"""
|
| 172 |
+
Top-K Router for Mixture of Experts (MoE) models.
|
| 173 |
+
|
| 174 |
+
This router determines which experts should process each token based on the top-k scoring experts.
|
| 175 |
+
It also applies auxiliary losses to encourage load balancing among experts.
|
| 176 |
+
|
| 177 |
+
Args:
|
| 178 |
+
config (AriaMoELMConfig): Configuration object containing MoE-related parameters.
|
| 179 |
+
"""
|
| 180 |
+
|
| 181 |
+
def __init__(self, config: AriaMoELMConfig):
|
| 182 |
+
super().__init__()
|
| 183 |
+
self.config = config
|
| 184 |
+
|
| 185 |
+
self.weight = nn.Parameter(
|
| 186 |
+
torch.empty((self.config.moe_num_experts, self.config.hidden_size))
|
| 187 |
+
)
|
| 188 |
+
# FIXME: initialize the weight
|
| 189 |
+
|
| 190 |
+
def gating(self, input: torch.Tensor) -> torch.Tensor:
|
| 191 |
+
"""
|
| 192 |
+
Compute the gating logits for each token-expert pair.
|
| 193 |
+
|
| 194 |
+
Args:
|
| 195 |
+
input (torch.Tensor): Input tensor of shape [batch_size * seq_len, hidden_size].
|
| 196 |
+
|
| 197 |
+
Returns:
|
| 198 |
+
torch.Tensor: Logits tensor of shape [batch_size * seq_len, num_experts].
|
| 199 |
+
"""
|
| 200 |
+
logits = torch.nn.functional.linear(input, self.weight)
|
| 201 |
+
return logits
|
| 202 |
+
|
| 203 |
+
def apply_z_loss(self, logits: torch.Tensor) -> torch.Tensor:
|
| 204 |
+
"""
|
| 205 |
+
Apply z-loss to encourage router logits to remain small for enhanced stability.
|
| 206 |
+
|
| 207 |
+
Args:
|
| 208 |
+
logits (torch.Tensor): Router logits.
|
| 209 |
+
|
| 210 |
+
Returns:
|
| 211 |
+
torch.Tensor: Logits with z-loss applied.
|
| 212 |
+
"""
|
| 213 |
+
z_loss = z_loss_func(logits, self.config.moe_z_loss_coeff)
|
| 214 |
+
logits = MoEAuxLossAutoScaler.apply(logits, z_loss)
|
| 215 |
+
return logits
|
| 216 |
+
|
| 217 |
+
def apply_aux_loss(
|
| 218 |
+
self,
|
| 219 |
+
logits: torch.Tensor,
|
| 220 |
+
tokens_per_expert: torch.Tensor,
|
| 221 |
+
activation: torch.Tensor,
|
| 222 |
+
) -> torch.Tensor:
|
| 223 |
+
"""
|
| 224 |
+
Apply auxiliary loss for load balancing among experts.
|
| 225 |
+
|
| 226 |
+
Args:
|
| 227 |
+
logits (torch.Tensor): Router logits.
|
| 228 |
+
tokens_per_expert (torch.Tensor): Number of tokens assigned to each expert.
|
| 229 |
+
activation (torch.Tensor): Activation values.
|
| 230 |
+
|
| 231 |
+
Returns:
|
| 232 |
+
torch.Tensor: Activation with auxiliary loss applied.
|
| 233 |
+
"""
|
| 234 |
+
probs = torch.softmax(logits, dim=-1, dtype=torch.float32)
|
| 235 |
+
aux_loss = switch_load_balancing_loss_func(
|
| 236 |
+
probs,
|
| 237 |
+
tokens_per_expert,
|
| 238 |
+
self.config.moe_topk,
|
| 239 |
+
self.config.moe_aux_loss_coeff,
|
| 240 |
+
)
|
| 241 |
+
return MoEAuxLossAutoScaler.apply(activation, aux_loss)
|
| 242 |
+
|
| 243 |
+
def routing(
|
| 244 |
+
self, logits: torch.Tensor
|
| 245 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 246 |
+
"""
|
| 247 |
+
Perform the routing operation to determine expert assignments.
|
| 248 |
+
|
| 249 |
+
Args:
|
| 250 |
+
logits (torch.Tensor): Router logits.
|
| 251 |
+
|
| 252 |
+
Returns:
|
| 253 |
+
Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 254 |
+
- scores: Softmax probabilities for top-k experts.
|
| 255 |
+
- top_indices: Indices of top-k experts for each token.
|
| 256 |
+
- tokens_per_expert: Number of tokens assigned to each expert.
|
| 257 |
+
"""
|
| 258 |
+
if self.training:
|
| 259 |
+
logits = self.apply_z_loss(logits)
|
| 260 |
+
|
| 261 |
+
top_logits, top_indices = torch.topk(logits, k=self.config.moe_topk, dim=1)
|
| 262 |
+
scores = torch.softmax(top_logits, dim=-1, dtype=torch.float32).type_as(logits)
|
| 263 |
+
|
| 264 |
+
tokens_per_expert = torch.histc(
|
| 265 |
+
top_indices.flatten(),
|
| 266 |
+
bins=self.config.moe_num_experts,
|
| 267 |
+
min=0,
|
| 268 |
+
max=self.config.moe_num_experts - 1,
|
| 269 |
+
)
|
| 270 |
+
|
| 271 |
+
if self.training:
|
| 272 |
+
scores = self.apply_aux_loss(logits, tokens_per_expert, scores)
|
| 273 |
+
return scores, top_indices, tokens_per_expert
|
| 274 |
+
|
| 275 |
+
def forward(
|
| 276 |
+
self, input: torch.Tensor
|
| 277 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 278 |
+
"""
|
| 279 |
+
Forward pass of the TopKRouter.
|
| 280 |
+
|
| 281 |
+
Args:
|
| 282 |
+
input (torch.Tensor): Input tensor of shape [batch_size * seq_len, hidden_size].
|
| 283 |
+
|
| 284 |
+
Returns:
|
| 285 |
+
Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 286 |
+
- scores: Softmax probabilities for top-k experts.
|
| 287 |
+
- top_indices: Indices of top-k experts for each token.
|
| 288 |
+
- tokens_per_expert: Number of tokens assigned to each expert.
|
| 289 |
+
"""
|
| 290 |
+
logits = self.gating(input)
|
| 291 |
+
logits = logits.view(-1, self.config.moe_num_experts)
|
| 292 |
+
scores, top_indices, tokens_per_expert = self.routing(logits)
|
| 293 |
+
return scores, top_indices, tokens_per_expert
|
| 294 |
+
|
| 295 |
+
|
| 296 |
+
# adapted from https://github.com/NVIDIA/Megatron-LM/blob/54f1f78529cbc2b9cddad313e7f9d96ac0420a27/megatron/core/transformer/moe/token_dispatcher.py#L291-L587
|
| 297 |
+
class TokenDispatcher:
|
| 298 |
+
"""
|
| 299 |
+
Handles the dispatching and gathering of tokens to and from experts.
|
| 300 |
+
|
| 301 |
+
This class is responsible for permuting tokens based on expert assignments and
|
| 302 |
+
unpermuting them after expert processing.
|
| 303 |
+
|
| 304 |
+
Args:
|
| 305 |
+
config (AriaMoELMConfig): Configuration object containing MoE-related parameters.
|
| 306 |
+
"""
|
| 307 |
+
|
| 308 |
+
def __init__(self, config: AriaMoELMConfig):
|
| 309 |
+
self.config = config
|
| 310 |
+
self.hidden_states_shape = None
|
| 311 |
+
self.reversed_input_permutation_mapping = None
|
| 312 |
+
|
| 313 |
+
def token_permutation(
|
| 314 |
+
self, hidden_states: torch.Tensor, indices: torch.Tensor
|
| 315 |
+
) -> torch.Tensor:
|
| 316 |
+
"""
|
| 317 |
+
Permute tokens based on expert assignments.
|
| 318 |
+
|
| 319 |
+
Args:
|
| 320 |
+
hidden_states (torch.Tensor): Input hidden states.
|
| 321 |
+
indices (torch.Tensor): Expert assignment indices.
|
| 322 |
+
|
| 323 |
+
Returns:
|
| 324 |
+
torch.Tensor: Permuted tokens.
|
| 325 |
+
"""
|
| 326 |
+
self.hidden_states_shape = hidden_states.shape
|
| 327 |
+
hidden_states = hidden_states.view(-1, hidden_states.size(-1))
|
| 328 |
+
flatten_indices = indices.flatten()
|
| 329 |
+
sorted_indices = torch.argsort(flatten_indices, stable=True)
|
| 330 |
+
permuted_tokens = hidden_states.index_select(
|
| 331 |
+
0, sorted_indices // self.config.moe_topk
|
| 332 |
+
)
|
| 333 |
+
self.reversed_input_permutation_mapping = sorted_indices
|
| 334 |
+
return permuted_tokens
|
| 335 |
+
|
| 336 |
+
def token_unpermutation(
|
| 337 |
+
self, permuted_tokens: torch.Tensor, scores: torch.Tensor
|
| 338 |
+
) -> torch.Tensor:
|
| 339 |
+
"""
|
| 340 |
+
Unpermute tokens and combine expert outputs.
|
| 341 |
+
|
| 342 |
+
Args:
|
| 343 |
+
permuted_tokens (torch.Tensor): Tokens after expert processing.
|
| 344 |
+
scores (torch.Tensor): Expert assignment scores.
|
| 345 |
+
|
| 346 |
+
Returns:
|
| 347 |
+
torch.Tensor: Unpermuted and combined output.
|
| 348 |
+
"""
|
| 349 |
+
num_unpermuted_tokens = scores.numel()
|
| 350 |
+
unpermuted_tokens = torch.zeros(
|
| 351 |
+
(num_unpermuted_tokens, permuted_tokens.size(1)),
|
| 352 |
+
dtype=permuted_tokens.dtype,
|
| 353 |
+
device=permuted_tokens.device,
|
| 354 |
+
)
|
| 355 |
+
unpermuted_tokens.index_copy_(
|
| 356 |
+
0, self.reversed_input_permutation_mapping, permuted_tokens
|
| 357 |
+
)
|
| 358 |
+
unpermuted_tokens = unpermuted_tokens.reshape(
|
| 359 |
+
-1, self.config.moe_topk, permuted_tokens.size(1)
|
| 360 |
+
)
|
| 361 |
+
|
| 362 |
+
unpermuted_tokens = unpermuted_tokens * scores.unsqueeze(-1)
|
| 363 |
+
unpermuted_tokens = unpermuted_tokens.sum(dim=1).type_as(permuted_tokens)
|
| 364 |
+
output = unpermuted_tokens.view(self.hidden_states_shape)
|
| 365 |
+
return output
|
| 366 |
+
|
| 367 |
+
|
| 368 |
+
class SharedExpertMLP(LlamaMLP):
|
| 369 |
+
"""
|
| 370 |
+
Shared Expert MLP for shared experts.
|
| 371 |
+
|
| 372 |
+
Unlike routed experts, shared experts process all tokens without routing.
|
| 373 |
+
This class reconfigures the intermediate size in comparison to the LlamaMLP.
|
| 374 |
+
|
| 375 |
+
Args:
|
| 376 |
+
config (AriaMoELMConfig): Configuration object for the AriaMoE language model.
|
| 377 |
+
"""
|
| 378 |
+
|
| 379 |
+
def __init__(self, config: AriaMoELMConfig):
|
| 380 |
+
nn.Module.__init__(self)
|
| 381 |
+
self.config = config
|
| 382 |
+
self.hidden_size = config.hidden_size
|
| 383 |
+
self.intermediate_size = (
|
| 384 |
+
config.moe_intermediate_size * config.moe_num_shared_experts
|
| 385 |
+
)
|
| 386 |
+
self.gate_proj = nn.Linear(
|
| 387 |
+
self.hidden_size, self.intermediate_size, bias=config.mlp_bias
|
| 388 |
+
)
|
| 389 |
+
self.up_proj = nn.Linear(
|
| 390 |
+
self.hidden_size, self.intermediate_size, bias=config.mlp_bias
|
| 391 |
+
)
|
| 392 |
+
self.down_proj = nn.Linear(
|
| 393 |
+
self.intermediate_size, self.hidden_size, bias=config.mlp_bias
|
| 394 |
+
)
|
| 395 |
+
self.act_fn = ACT2FN[config.hidden_act]
|
| 396 |
+
|
| 397 |
+
|
| 398 |
+
def sequential_gemm(input, weight, tokens_per_expert):
|
| 399 |
+
"""
|
| 400 |
+
Compute the matrix multiplication (GEMM) for each expert sequentially. This approach is computationally inefficient, especially when dealing with a large number of experts.
|
| 401 |
+
|
| 402 |
+
Args:
|
| 403 |
+
input (torch.Tensor): Input tensor of shape (num_tokens, in_features).
|
| 404 |
+
weight (torch.Tensor): Weight tensor of shape (num_experts, in_features, out_features).
|
| 405 |
+
tokens_per_expert (torch.Tensor): Number of tokens assigned to each expert.
|
| 406 |
+
|
| 407 |
+
Returns:
|
| 408 |
+
torch.Tensor: Output tensor of shape (num_tokens, out_features).
|
| 409 |
+
"""
|
| 410 |
+
num_tokens = input.shape[0]
|
| 411 |
+
out_features = weight.shape[-1]
|
| 412 |
+
output = torch.zeros(
|
| 413 |
+
num_tokens, out_features, dtype=input.dtype, device=input.device
|
| 414 |
+
)
|
| 415 |
+
|
| 416 |
+
cumsum_num_tokens = torch.cumsum(tokens_per_expert, dim=0)
|
| 417 |
+
# Insert zero at the begining for offset index's convenience
|
| 418 |
+
zero_tensor = torch.zeros(1, dtype=torch.long, device=cumsum_num_tokens.device)
|
| 419 |
+
cumsum_num_tokens = torch.cat((zero_tensor, cumsum_num_tokens))
|
| 420 |
+
|
| 421 |
+
for expert_num in range(weight.shape[0]):
|
| 422 |
+
start = cumsum_num_tokens[expert_num]
|
| 423 |
+
end = cumsum_num_tokens[expert_num + 1]
|
| 424 |
+
tokens = input[start:end]
|
| 425 |
+
|
| 426 |
+
out = torch.matmul(tokens, weight[expert_num])
|
| 427 |
+
output[start:end] = out
|
| 428 |
+
return output
|
| 429 |
+
|
| 430 |
+
|
| 431 |
+
try:
|
| 432 |
+
from grouped_gemm.ops import gmm as experts_gemm
|
| 433 |
+
|
| 434 |
+
if os.environ.get("USE_GROUPED_GEMM", "1") == "0":
|
| 435 |
+
logger.warning(
|
| 436 |
+
"environment variable USE_GROUPED_GEMM is set to 0, using sequential GEMM instead."
|
| 437 |
+
)
|
| 438 |
+
experts_gemm = sequential_gemm
|
| 439 |
+
except ImportError:
|
| 440 |
+
logger.warning(
|
| 441 |
+
"`grouped_gemm` is not installed, using sequential GEMM, which is slower."
|
| 442 |
+
)
|
| 443 |
+
experts_gemm = sequential_gemm
|
| 444 |
+
|
| 445 |
+
|
| 446 |
+
class GroupedGEMM(nn.Module):
|
| 447 |
+
"""
|
| 448 |
+
Grouped GEMM (General Matrix Multiplication) module for efficient expert computation.
|
| 449 |
+
This module utilizes the grouped_gemm library (https://github.com/fanshiqing/grouped_gemm)
|
| 450 |
+
for optimized performance. If the grouped_gemm library is not installed, it gracefully
|
| 451 |
+
falls back to a sequential GEMM implementation, which may be slower but ensures
|
| 452 |
+
functionality.
|
| 453 |
+
|
| 454 |
+
Args:
|
| 455 |
+
in_features (int): Number of input features.
|
| 456 |
+
out_features (int): Number of output features.
|
| 457 |
+
groups (int): Number of expert groups.
|
| 458 |
+
"""
|
| 459 |
+
|
| 460 |
+
def __init__(self, in_features, out_features, groups):
|
| 461 |
+
super().__init__()
|
| 462 |
+
self.in_features = in_features
|
| 463 |
+
self.out_features = out_features
|
| 464 |
+
self.groups = groups
|
| 465 |
+
self.weight = nn.Parameter(torch.empty(groups, in_features, out_features))
|
| 466 |
+
|
| 467 |
+
def forward(self, input, tokens_per_expert):
|
| 468 |
+
"""
|
| 469 |
+
Perform grouped matrix multiplication.
|
| 470 |
+
|
| 471 |
+
Args:
|
| 472 |
+
input (torch.Tensor): Input tensor of shape (num_tokens, in_features).
|
| 473 |
+
tokens_per_expert (torch.Tensor): Number of tokens assigned to each expert.
|
| 474 |
+
|
| 475 |
+
Returns:
|
| 476 |
+
torch.Tensor: Output tensor of shape (num_tokens, out_features).
|
| 477 |
+
"""
|
| 478 |
+
tokens_per_expert = tokens_per_expert.cpu()
|
| 479 |
+
|
| 480 |
+
# Ensure the CUDA device matches the input tensor's device.
|
| 481 |
+
# This mismatch can occur when using `transformers.AutoModel.from_pretrained`
|
| 482 |
+
# with `device_map="auto"` on a multi-GPU setup.
|
| 483 |
+
torch.cuda.set_device(input.device)
|
| 484 |
+
return experts_gemm(input, self.weight, tokens_per_expert)
|
| 485 |
+
|
| 486 |
+
|
| 487 |
+
class GroupedMLP(nn.Module):
|
| 488 |
+
"""
|
| 489 |
+
Grouped MLP module for Mixture of Experts.
|
| 490 |
+
|
| 491 |
+
Args:
|
| 492 |
+
config (AriaMoELMConfig): Configuration object for the model.
|
| 493 |
+
"""
|
| 494 |
+
|
| 495 |
+
def __init__(self, config: AriaMoELMConfig) -> None:
|
| 496 |
+
super().__init__()
|
| 497 |
+
self.config = config
|
| 498 |
+
self.fc1 = GroupedGEMM(
|
| 499 |
+
config.hidden_size, config.moe_intermediate_size * 2, config.moe_num_experts
|
| 500 |
+
)
|
| 501 |
+
self.fc2 = GroupedGEMM(
|
| 502 |
+
config.moe_intermediate_size, config.hidden_size, config.moe_num_experts
|
| 503 |
+
)
|
| 504 |
+
|
| 505 |
+
def glu(x):
|
| 506 |
+
x = torch.chunk(x, 2, dim=-1)
|
| 507 |
+
return F.silu(x[0]) * x[1]
|
| 508 |
+
|
| 509 |
+
self.activation_func = glu
|
| 510 |
+
|
| 511 |
+
def forward(self, permuted_tokens, tokens_per_expert):
|
| 512 |
+
"""
|
| 513 |
+
Forward pass of the Grouped MLP.
|
| 514 |
+
|
| 515 |
+
Args:
|
| 516 |
+
permuted_tokens (torch.Tensor): Permuted input tokens.
|
| 517 |
+
tokens_per_expert (torch.Tensor): Number of tokens assigned to each expert.
|
| 518 |
+
|
| 519 |
+
Returns:
|
| 520 |
+
torch.Tensor: Output tensor after passing through the MLP.
|
| 521 |
+
"""
|
| 522 |
+
fc1_output = self.fc1(permuted_tokens, tokens_per_expert)
|
| 523 |
+
fc1_output = self.activation_func(fc1_output)
|
| 524 |
+
fc2_output = self.fc2(fc1_output, tokens_per_expert)
|
| 525 |
+
return fc2_output
|
| 526 |
+
|
| 527 |
+
|
| 528 |
+
class MoELayer(nn.Module):
|
| 529 |
+
"""
|
| 530 |
+
Mixture of Experts (MoE) Layer for the AriaMoE model.
|
| 531 |
+
|
| 532 |
+
This layer implements the MoE mechanism, which routes input tokens to different experts
|
| 533 |
+
based on a routing algorithm, processes them through the experts, and then combines
|
| 534 |
+
the outputs.
|
| 535 |
+
|
| 536 |
+
Args:
|
| 537 |
+
config (AriaMoELMConfig): Configuration object for the MoE layer.
|
| 538 |
+
"""
|
| 539 |
+
|
| 540 |
+
def __init__(self, config: AriaMoELMConfig):
|
| 541 |
+
super().__init__()
|
| 542 |
+
|
| 543 |
+
self.router = TopKRouter(config)
|
| 544 |
+
self.token_dispatcher = TokenDispatcher(config)
|
| 545 |
+
self.experts = GroupedMLP(config)
|
| 546 |
+
self.shared_experts = SharedExpertMLP(config)
|
| 547 |
+
|
| 548 |
+
def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
|
| 549 |
+
"""
|
| 550 |
+
Forward pass of the MoE Layer.
|
| 551 |
+
|
| 552 |
+
Args:
|
| 553 |
+
hidden_states (torch.Tensor): Input tensor of shape (batch_size, sequence_length, hidden_size).
|
| 554 |
+
|
| 555 |
+
Returns:
|
| 556 |
+
torch.Tensor: Output tensor after passing through the MoE layer.
|
| 557 |
+
|
| 558 |
+
Process:
|
| 559 |
+
1. Route tokens to experts using the router.
|
| 560 |
+
2. Permute tokens based on routing decisions.
|
| 561 |
+
3. Process tokens through experts.
|
| 562 |
+
4. Unpermute and combine expert outputs.
|
| 563 |
+
5. Add shared expert output to the final result.
|
| 564 |
+
"""
|
| 565 |
+
scores, indices, tokens_per_expert = self.router(hidden_states)
|
| 566 |
+
|
| 567 |
+
permuted_tokens = self.token_dispatcher.token_permutation(
|
| 568 |
+
hidden_states, indices
|
| 569 |
+
)
|
| 570 |
+
|
| 571 |
+
expert_output = self.experts(permuted_tokens, tokens_per_expert)
|
| 572 |
+
|
| 573 |
+
output = self.token_dispatcher.token_unpermutation(expert_output, scores)
|
| 574 |
+
|
| 575 |
+
shared_expert_output = self.shared_experts(hidden_states)
|
| 576 |
+
output += shared_expert_output
|
| 577 |
+
return output
|
| 578 |
+
|
| 579 |
+
|
| 580 |
+
class MoEDecoderLayer(LlamaDecoderLayer):
|
| 581 |
+
"""
|
| 582 |
+
Custom Decoder Layer for the AriaMoE model which modifies the standard `LlamaDecoderLayer` by
|
| 583 |
+
replacing the traditional MLP with a Mixture of Experts (MoE) Layer.
|
| 584 |
+
|
| 585 |
+
Args:
|
| 586 |
+
config (LlamaConfig): Configuration object for the layer.
|
| 587 |
+
layer_idx (int): Index of the current layer in the model.
|
| 588 |
+
"""
|
| 589 |
+
|
| 590 |
+
def __init__(self, config: LlamaConfig, layer_idx: int):
|
| 591 |
+
nn.Module.__init__(self)
|
| 592 |
+
self.hidden_size = config.hidden_size
|
| 593 |
+
|
| 594 |
+
self.self_attn = LLAMA_ATTENTION_CLASSES[config._attn_implementation](
|
| 595 |
+
config=config, layer_idx=layer_idx
|
| 596 |
+
)
|
| 597 |
+
|
| 598 |
+
self.mlp = MoELayer(config)
|
| 599 |
+
self.input_layernorm = LlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
|
| 600 |
+
self.post_attention_layernorm = LlamaRMSNorm(
|
| 601 |
+
config.hidden_size, eps=config.rms_norm_eps
|
| 602 |
+
)
|
| 603 |
+
|
| 604 |
+
|
| 605 |
+
class AriaMoELMModel(LlamaModel):
|
| 606 |
+
"""
|
| 607 |
+
Custom LlamaModel for the AriaMoE model which modifies the standard LlamaModel by
|
| 608 |
+
replacing the `LlamaDecoderLayer` with `MoEDecoderLayer`.
|
| 609 |
+
|
| 610 |
+
This model implements a Mixture of Experts (MoE) approach, where each layer contains
|
| 611 |
+
multiple expert networks that specialize in different aspects of the input.
|
| 612 |
+
|
| 613 |
+
Args:
|
| 614 |
+
config (LlamaConfig): Configuration object for the model.
|
| 615 |
+
"""
|
| 616 |
+
|
| 617 |
+
def __init__(self, config: LlamaConfig):
|
| 618 |
+
super().__init__(config)
|
| 619 |
+
self.padding_idx = config.pad_token_id
|
| 620 |
+
self.vocab_size = config.vocab_size
|
| 621 |
+
|
| 622 |
+
self.embed_tokens = nn.Embedding(
|
| 623 |
+
config.vocab_size, config.hidden_size, self.padding_idx
|
| 624 |
+
)
|
| 625 |
+
self.layers = nn.ModuleList(
|
| 626 |
+
[
|
| 627 |
+
MoEDecoderLayer(config, layer_idx)
|
| 628 |
+
for layer_idx in range(config.num_hidden_layers)
|
| 629 |
+
]
|
| 630 |
+
)
|
| 631 |
+
self.norm = LlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
|
| 632 |
+
self.rotary_emb = LlamaRotaryEmbedding(config=config)
|
| 633 |
+
self.gradient_checkpointing = False
|
| 634 |
+
|
| 635 |
+
# Initialize weights and apply final processing
|
| 636 |
+
self.post_init()
|
| 637 |
+
|
| 638 |
+
|
| 639 |
+
class AriaMoELMForCausalLM(LlamaForCausalLM, GenerationMixin):
|
| 640 |
+
"""
|
| 641 |
+
AriaMoE model for causal language modeling tasks.
|
| 642 |
+
|
| 643 |
+
This class extends LlamaForCausalLM to incorporate the Mixture of Experts (MoE) approach,
|
| 644 |
+
allowing for more efficient and scalable language modeling.
|
| 645 |
+
|
| 646 |
+
Args:
|
| 647 |
+
config (AriaMoELMConfig): Configuration object for the model.
|
| 648 |
+
"""
|
| 649 |
+
|
| 650 |
+
_tied_weights_keys = ["lm_head.weight"]
|
| 651 |
+
config_class = AriaMoELMConfig
|
| 652 |
+
_no_split_modules = ["MoEDecoderLayer"]
|
| 653 |
+
|
| 654 |
+
def __init__(self, config):
|
| 655 |
+
super().__init__(config)
|
| 656 |
+
self.model = AriaMoELMModel(config)
|
| 657 |
+
self.vocab_size = config.vocab_size
|
| 658 |
+
self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
|
| 659 |
+
|
| 660 |
+
# Initialize weights and apply final processing
|
| 661 |
+
self.post_init()
|
| 662 |
+
|
| 663 |
+
def set_z_loss_coeff(self, z_loss_coeff: float):
|
| 664 |
+
"""
|
| 665 |
+
Set the coefficient for the z-loss in the MoE routing.
|
| 666 |
+
|
| 667 |
+
Args:
|
| 668 |
+
z_loss_coeff (float): The coefficient for the z-loss.
|
| 669 |
+
"""
|
| 670 |
+
self.config.moe_z_loss_coeff = z_loss_coeff
|
| 671 |
+
|
| 672 |
+
def set_aux_loss_coeff(self, aux_loss_coeff: float):
|
| 673 |
+
"""
|
| 674 |
+
Set the coefficient for the auxiliary loss in the MoE routing.
|
| 675 |
+
|
| 676 |
+
Args:
|
| 677 |
+
aux_loss_coeff (float): The coefficient for the auxiliary loss.
|
| 678 |
+
"""
|
| 679 |
+
self.config.moe_aux_loss_coeff = aux_loss_coeff
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_transform": null,
|
| 3 |
+
"auto_map": {
|
| 4 |
+
"AutoImageProcessor": "vision_processor.AriaVisionProcessor",
|
| 5 |
+
"AutoProcessor": "processing_aria.AriaProcessor"
|
| 6 |
+
},
|
| 7 |
+
"image_mean": [
|
| 8 |
+
0.5,
|
| 9 |
+
0.5,
|
| 10 |
+
0.5
|
| 11 |
+
],
|
| 12 |
+
"image_processor_type": "AriaVisionProcessor",
|
| 13 |
+
"image_std": [
|
| 14 |
+
0.5,
|
| 15 |
+
0.5,
|
| 16 |
+
0.5
|
| 17 |
+
],
|
| 18 |
+
"max_image_size": 980,
|
| 19 |
+
"min_image_size": 336,
|
| 20 |
+
"processor_class": "AriaProcessor"
|
| 21 |
+
}
|
processing_aria.py
ADDED
|
@@ -0,0 +1,305 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright 2024 Rhymes AI. All rights reserved.
|
| 2 |
+
#
|
| 3 |
+
# Licensed to the Apache Software Foundation (ASF) under one
|
| 4 |
+
# or more contributor license agreements. See the NOTICE file
|
| 5 |
+
# distributed with this work for additional information
|
| 6 |
+
# regarding copyright ownership. The ASF licenses this file
|
| 7 |
+
# to you under the Apache License, Version 2.0 (the
|
| 8 |
+
# "License"); you may not use this file except in compliance
|
| 9 |
+
# with the License. You may obtain a copy of the License at
|
| 10 |
+
#
|
| 11 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 12 |
+
#
|
| 13 |
+
# Unless required by applicable law or agreed to in writing,
|
| 14 |
+
# software distributed under the License is distributed on an
|
| 15 |
+
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
| 16 |
+
# KIND, either express or implied. See the License for the
|
| 17 |
+
# specific language governing permissions and limitations
|
| 18 |
+
# under the License.
|
| 19 |
+
|
| 20 |
+
import inspect
|
| 21 |
+
import logging
|
| 22 |
+
import re
|
| 23 |
+
from typing import List, Optional, Union
|
| 24 |
+
|
| 25 |
+
from transformers import AutoTokenizer, BatchFeature
|
| 26 |
+
from transformers.image_utils import ImageInput
|
| 27 |
+
from transformers.processing_utils import ProcessorMixin
|
| 28 |
+
from transformers.tokenization_utils import (
|
| 29 |
+
PaddingStrategy,
|
| 30 |
+
PreTokenizedInput,
|
| 31 |
+
TensorType,
|
| 32 |
+
TextInput,
|
| 33 |
+
TruncationStrategy,
|
| 34 |
+
)
|
| 35 |
+
|
| 36 |
+
from .vision_processor import AriaVisionProcessor
|
| 37 |
+
|
| 38 |
+
logger = logging.getLogger(__name__)
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
class AriaProcessor(ProcessorMixin):
|
| 42 |
+
"""
|
| 43 |
+
AriaProcessor is a processor for the Aria model which wraps the Aria image preprocessor and the LLama slow tokenizer.
|
| 44 |
+
Args:
|
| 45 |
+
image_processor(AriaVisionProcessor): The AriaVisionProcessor to use for image preprocessing.
|
| 46 |
+
tokenizer(AutoTokenizer): The AutoTokenizer to use for tokenizing the text.
|
| 47 |
+
patch_size(int): The patch size to use for the image processor.
|
| 48 |
+
chat_template(str): The chat template to use for the tokenizer.
|
| 49 |
+
image_token(str): The image token to use for the tokenizer.
|
| 50 |
+
"""
|
| 51 |
+
|
| 52 |
+
attributes = []
|
| 53 |
+
valid_kwargs = ["chat_template", "patch_size", "image_token"]
|
| 54 |
+
image_processor_class = None
|
| 55 |
+
tokenizer_class = "AutoTokenizer"
|
| 56 |
+
|
| 57 |
+
def __init__(
|
| 58 |
+
self,
|
| 59 |
+
image_processor: AriaVisionProcessor = None,
|
| 60 |
+
tokenizer: Union[AutoTokenizer, str] = None,
|
| 61 |
+
patch_size: int = 490,
|
| 62 |
+
chat_template: str = None,
|
| 63 |
+
image_token: str = "<|img|>",
|
| 64 |
+
):
|
| 65 |
+
super().__init__(chat_template=chat_template)
|
| 66 |
+
|
| 67 |
+
if image_processor is None:
|
| 68 |
+
self.image_processor = AriaVisionProcessor(max_image_size=patch_size)
|
| 69 |
+
else:
|
| 70 |
+
self.image_processor = image_processor
|
| 71 |
+
|
| 72 |
+
if isinstance(tokenizer, str):
|
| 73 |
+
self.tokenizer = AutoTokenizer.from_pretrained(
|
| 74 |
+
tokenizer, trust_remote_code=True, use_fast=False
|
| 75 |
+
)
|
| 76 |
+
else:
|
| 77 |
+
self.tokenizer = tokenizer
|
| 78 |
+
|
| 79 |
+
if self.tokenizer is not None and self.tokenizer.pad_token is None:
|
| 80 |
+
self.tokenizer.pad_token = self.tokenizer.unk_token
|
| 81 |
+
|
| 82 |
+
self.image_token = image_token
|
| 83 |
+
|
| 84 |
+
# Copied from transformers.models.llava_next.processing_llave_next.LlavaNextProcessor.__call__
|
| 85 |
+
def __call__(
|
| 86 |
+
self,
|
| 87 |
+
text: Union[
|
| 88 |
+
TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]
|
| 89 |
+
],
|
| 90 |
+
images: ImageInput = None,
|
| 91 |
+
padding: Union[bool, str, PaddingStrategy] = False,
|
| 92 |
+
truncation: Union[bool, str, TruncationStrategy] = None,
|
| 93 |
+
max_length: Optional[int] = None,
|
| 94 |
+
max_image_size: Optional[int] = 980,
|
| 95 |
+
split_image: Optional[bool] = False,
|
| 96 |
+
return_tensors: Optional[Union[str, TensorType]] = TensorType.PYTORCH,
|
| 97 |
+
return_final_prompts: Optional[bool] = False,
|
| 98 |
+
) -> BatchFeature:
|
| 99 |
+
"""
|
| 100 |
+
Main method to prepare for the model one or several sequences(s) and image(s). Please refer to the doctsring
|
| 101 |
+
of the above two methods for more information.
|
| 102 |
+
|
| 103 |
+
Args:
|
| 104 |
+
text (`str`, `List[str]`, `List[List[str]]`):
|
| 105 |
+
The sequence or batch of sequences to be encoded. Each sequence can be a string or a list of strings
|
| 106 |
+
(pretokenized string). If the sequences are provided as list of strings (pretokenized), you must set
|
| 107 |
+
`is_split_into_words=True` (to lift the ambiguity with a batch of sequences).
|
| 108 |
+
images (`PIL.Image.Image`, `np.ndarray`, `torch.Tensor`, `List[PIL.Image.Image]`, `List[np.ndarray]`, `List[torch.Tensor]`):
|
| 109 |
+
The image or batch of images to be prepared. Each image can be a PIL image, NumPy array or PyTorch
|
| 110 |
+
tensor. Both channels-first and channels-last formats are supported.
|
| 111 |
+
padding (`bool`, `str` or [`~utils.PaddingStrategy`], *optional*, defaults to `False`):
|
| 112 |
+
Select a strategy to pad the returned sequences (according to the model's padding side and padding
|
| 113 |
+
index) among:
|
| 114 |
+
- `True` or `'longest'`: Pad to the longest sequence in the batch (or no padding if only a single
|
| 115 |
+
sequence if provided).
|
| 116 |
+
- `'max_length'`: Pad to a maximum length specified with the argument `max_length` or to the maximum
|
| 117 |
+
acceptable input length for the model if that argument is not provided.
|
| 118 |
+
- `False` or `'do_not_pad'` (default): No padding (i.e., can output a batch with sequences of different
|
| 119 |
+
lengths).
|
| 120 |
+
max_length (`int`, *optional*):
|
| 121 |
+
Maximum length of the returned list and optionally padding length (see above).
|
| 122 |
+
max_image_size (`int`, *optional*):
|
| 123 |
+
Maximum size of the image to be processed.
|
| 124 |
+
split_image (`bool`, *optional*):
|
| 125 |
+
Whether to split the image into patches before processing.
|
| 126 |
+
truncation (`bool`, *optional*):
|
| 127 |
+
Activates truncation to cut input sequences longer than `max_length` to `max_length`.
|
| 128 |
+
return_tensors (`str` or [`~utils.TensorType`], *optional*):
|
| 129 |
+
If set, will return tensors of a particular framework. Acceptable values are:
|
| 130 |
+
|
| 131 |
+
- `'tf'`: Return TensorFlow `tf.constant` objects.
|
| 132 |
+
- `'pt'`: Return PyTorch `torch.Tensor` objects.
|
| 133 |
+
- `'np'`: Return NumPy `np.ndarray` objects.
|
| 134 |
+
- `'jax'`: Return JAX `jnp.ndarray` objects.
|
| 135 |
+
|
| 136 |
+
Returns:
|
| 137 |
+
[`BatchFeature`]: A [`BatchFeature`] with the following fields:
|
| 138 |
+
|
| 139 |
+
- **input_ids** -- List of token ids to be fed to a model. Returned when `text` is not `None`.
|
| 140 |
+
- **attention_mask** -- List of indices specifying which tokens should be attended to by the model (when
|
| 141 |
+
`return_attention_mask=True` or if *"attention_mask"* is in `self.model_input_names` and if `text` is not
|
| 142 |
+
`None`).
|
| 143 |
+
- **pixel_values** -- Pixel values to be fed to a model. Returned when `images` is not `None`.
|
| 144 |
+
- **pixel_mask** -- Pixel mask to be fed to a model. Returned when `images` is not `None`.
|
| 145 |
+
"""
|
| 146 |
+
if isinstance(text, str):
|
| 147 |
+
text = [text]
|
| 148 |
+
elif not isinstance(text, list) and not isinstance(text[0], str):
|
| 149 |
+
raise ValueError(
|
| 150 |
+
"Invalid input text. Please provide a string, or a list of strings"
|
| 151 |
+
)
|
| 152 |
+
|
| 153 |
+
if images is not None:
|
| 154 |
+
image_inputs = self.image_processor(
|
| 155 |
+
images,
|
| 156 |
+
return_tensors=return_tensors,
|
| 157 |
+
max_image_size=max_image_size,
|
| 158 |
+
split_image=split_image,
|
| 159 |
+
)
|
| 160 |
+
# expand the image_token according to the num_crops of image
|
| 161 |
+
prompt_strings = []
|
| 162 |
+
crop_iter = iter(image_inputs.pop("num_crops"))
|
| 163 |
+
for prompt in text:
|
| 164 |
+
prompt_strings.append(
|
| 165 |
+
re.sub(
|
| 166 |
+
re.escape(self.image_token),
|
| 167 |
+
lambda _: next(crop_iter) * self.image_token,
|
| 168 |
+
prompt,
|
| 169 |
+
)
|
| 170 |
+
)
|
| 171 |
+
|
| 172 |
+
max_image_size = (
|
| 173 |
+
max_image_size
|
| 174 |
+
if max_image_size is not None
|
| 175 |
+
else self.image_processor.max_image_size
|
| 176 |
+
)
|
| 177 |
+
if max_image_size == 490:
|
| 178 |
+
num_image_tokens = 128
|
| 179 |
+
elif max_image_size == 980:
|
| 180 |
+
num_image_tokens = 256
|
| 181 |
+
else:
|
| 182 |
+
raise ValueError(
|
| 183 |
+
f"max_image_size must be either 490 or 980, got {max_image_size}"
|
| 184 |
+
)
|
| 185 |
+
prompt_strings = [
|
| 186 |
+
sample.replace(self.image_token, self.image_token * num_image_tokens)
|
| 187 |
+
for sample in prompt_strings
|
| 188 |
+
]
|
| 189 |
+
|
| 190 |
+
else:
|
| 191 |
+
image_inputs = {}
|
| 192 |
+
prompt_strings = text
|
| 193 |
+
|
| 194 |
+
text_inputs = self.tokenizer(
|
| 195 |
+
prompt_strings,
|
| 196 |
+
return_tensors=return_tensors,
|
| 197 |
+
padding=padding,
|
| 198 |
+
truncation=truncation,
|
| 199 |
+
max_length=max_length,
|
| 200 |
+
)
|
| 201 |
+
|
| 202 |
+
if return_final_prompts:
|
| 203 |
+
return BatchFeature(data={**text_inputs, **image_inputs}), prompt_strings
|
| 204 |
+
else:
|
| 205 |
+
return BatchFeature(data={**text_inputs, **image_inputs})
|
| 206 |
+
|
| 207 |
+
@staticmethod
|
| 208 |
+
def _extract_kwargs(func: callable, **kwargs) -> dict:
|
| 209 |
+
"""
|
| 210 |
+
Extract the kwargs that are valid for the given function.
|
| 211 |
+
"""
|
| 212 |
+
return {
|
| 213 |
+
k: v for k, v in kwargs.items() if k in inspect.signature(func).parameters
|
| 214 |
+
}
|
| 215 |
+
|
| 216 |
+
def save_pretrained(self, save_directory, **kwargs):
|
| 217 |
+
"""
|
| 218 |
+
Save both the image processor and tokenizer.
|
| 219 |
+
"""
|
| 220 |
+
if self.image_processor is not None:
|
| 221 |
+
self.image_processor.save_pretrained(
|
| 222 |
+
save_directory,
|
| 223 |
+
**self._extract_kwargs(self.image_processor.save_pretrained, **kwargs),
|
| 224 |
+
)
|
| 225 |
+
if self.tokenizer is not None:
|
| 226 |
+
self.tokenizer.save_pretrained(
|
| 227 |
+
save_directory,
|
| 228 |
+
**self._extract_kwargs(self.tokenizer.save_pretrained, **kwargs),
|
| 229 |
+
)
|
| 230 |
+
|
| 231 |
+
@classmethod
|
| 232 |
+
def from_pretrained(
|
| 233 |
+
cls,
|
| 234 |
+
pretrained_model_name_or_path,
|
| 235 |
+
tokenizer_path=None,
|
| 236 |
+
image_processor_path=None,
|
| 237 |
+
**kwargs,
|
| 238 |
+
):
|
| 239 |
+
"""
|
| 240 |
+
Load both the image processor and tokenizer from a pretrained model path.
|
| 241 |
+
"""
|
| 242 |
+
tokenizer_path = (
|
| 243 |
+
tokenizer_path
|
| 244 |
+
if tokenizer_path is not None
|
| 245 |
+
else pretrained_model_name_or_path
|
| 246 |
+
)
|
| 247 |
+
image_processor_path = (
|
| 248 |
+
image_processor_path
|
| 249 |
+
if image_processor_path is not None
|
| 250 |
+
else pretrained_model_name_or_path
|
| 251 |
+
)
|
| 252 |
+
image_processor = AriaVisionProcessor.from_pretrained(
|
| 253 |
+
image_processor_path,
|
| 254 |
+
**cls._extract_kwargs(AriaVisionProcessor.from_pretrained, **kwargs),
|
| 255 |
+
)
|
| 256 |
+
if "use_fast" in kwargs:
|
| 257 |
+
logger.warning("use_fast is not supported for AriaProcessor. Ignoring...")
|
| 258 |
+
kwargs.pop("use_fast")
|
| 259 |
+
try:
|
| 260 |
+
tokenizer = AutoTokenizer.from_pretrained(
|
| 261 |
+
tokenizer_path,
|
| 262 |
+
use_fast=False,
|
| 263 |
+
**cls._extract_kwargs(AutoTokenizer.from_pretrained, **kwargs),
|
| 264 |
+
)
|
| 265 |
+
chat_template = tokenizer.chat_template
|
| 266 |
+
except Exception as e:
|
| 267 |
+
logger.warning(f"Failed to load tokenizer from {tokenizer_path}: {e}")
|
| 268 |
+
tokenizer = None
|
| 269 |
+
chat_template = None
|
| 270 |
+
return cls(
|
| 271 |
+
image_processor=image_processor,
|
| 272 |
+
tokenizer=tokenizer,
|
| 273 |
+
chat_template=chat_template,
|
| 274 |
+
)
|
| 275 |
+
|
| 276 |
+
# Copied from transformers.models.clip.processing_clip.CLIPProcessor.batch_decode with CLIP->Llama
|
| 277 |
+
def batch_decode(self, *args, **kwargs):
|
| 278 |
+
"""
|
| 279 |
+
This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.batch_decode`]. Please
|
| 280 |
+
refer to the docstring of this method for more information.
|
| 281 |
+
"""
|
| 282 |
+
if self.tokenizer is None:
|
| 283 |
+
raise ValueError(
|
| 284 |
+
"Tokenizer is not initialized. Please provide a valid tokenizer."
|
| 285 |
+
)
|
| 286 |
+
return self.tokenizer.batch_decode(*args, **kwargs)
|
| 287 |
+
|
| 288 |
+
# Copied from transformers.models.clip.processing_clip.CLIPProcessor.decode with CLIP->Llama
|
| 289 |
+
def decode(self, *args, **kwargs):
|
| 290 |
+
"""
|
| 291 |
+
This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.decode`]. Please refer to
|
| 292 |
+
the docstring of this method for more information.
|
| 293 |
+
"""
|
| 294 |
+
if self.tokenizer is None:
|
| 295 |
+
raise ValueError(
|
| 296 |
+
"Tokenizer is not initialized. Please provide a valid tokenizer."
|
| 297 |
+
)
|
| 298 |
+
return self.tokenizer.decode(*args, **kwargs)
|
| 299 |
+
|
| 300 |
+
@property
|
| 301 |
+
# Copied from transformers.models.clip.processing_clip.CLIPProcessor.model_input_names
|
| 302 |
+
def model_input_names(self):
|
| 303 |
+
tokenizer_input_names = self.tokenizer.model_input_names
|
| 304 |
+
image_processor_input_names = self.image_processor.model_input_names
|
| 305 |
+
return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names))
|
projector.py
ADDED
|
@@ -0,0 +1,189 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright 2024 Rhymes AI. All rights reserved.
|
| 2 |
+
#
|
| 3 |
+
# Licensed to the Apache Software Foundation (ASF) under one
|
| 4 |
+
# or more contributor license agreements. See the NOTICE file
|
| 5 |
+
# distributed with this work for additional information
|
| 6 |
+
# regarding copyright ownership. The ASF licenses this file
|
| 7 |
+
# to you under the Apache License, Version 2.0 (the
|
| 8 |
+
# "License"); you may not use this file except in compliance
|
| 9 |
+
# with the License. You may obtain a copy of the License at
|
| 10 |
+
#
|
| 11 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 12 |
+
#
|
| 13 |
+
# Unless required by applicable law or agreed to in writing,
|
| 14 |
+
# software distributed under the License is distributed on an
|
| 15 |
+
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
| 16 |
+
# KIND, either express or implied. See the License for the
|
| 17 |
+
# specific language governing permissions and limitations
|
| 18 |
+
# under the License.
|
| 19 |
+
|
| 20 |
+
import torch
|
| 21 |
+
import torch.nn as nn
|
| 22 |
+
from torch.nn.init import trunc_normal_
|
| 23 |
+
from transformers.activations import ACT2FN
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
class FFN(nn.Module):
|
| 27 |
+
"""
|
| 28 |
+
Feed-Forward Network module.
|
| 29 |
+
|
| 30 |
+
Args:
|
| 31 |
+
embed_dim (int): Input embedding dimension.
|
| 32 |
+
ff_dim (int): Hidden dimension of the feed-forward network.
|
| 33 |
+
output_dim (int): Output dimension.
|
| 34 |
+
"""
|
| 35 |
+
|
| 36 |
+
def __init__(self, embed_dim, ff_dim, output_dim):
|
| 37 |
+
super().__init__()
|
| 38 |
+
self.linear_in = nn.Linear(embed_dim, ff_dim, bias=False)
|
| 39 |
+
self.linear_out = nn.Linear(ff_dim, output_dim, bias=False)
|
| 40 |
+
self.act = ACT2FN["gelu_new"]
|
| 41 |
+
|
| 42 |
+
def forward(self, hidden_states):
|
| 43 |
+
hidden_states = self.act(self.linear_in(hidden_states))
|
| 44 |
+
hidden_states = self.linear_out(hidden_states)
|
| 45 |
+
return hidden_states
|
| 46 |
+
|
| 47 |
+
|
| 48 |
+
class CrossAttention(nn.Module):
|
| 49 |
+
"""
|
| 50 |
+
Cross-Attention module.
|
| 51 |
+
|
| 52 |
+
Args:
|
| 53 |
+
kv_dim (int): Dimension of key and value.
|
| 54 |
+
embed_dim (int): Embedding dimension.
|
| 55 |
+
num_heads (int): Number of attention heads.
|
| 56 |
+
drop_out_rate (float): Dropout rate. Default is 0.
|
| 57 |
+
"""
|
| 58 |
+
|
| 59 |
+
def __init__(self, kv_dim, embed_dim, num_heads, drop_out_rate=0):
|
| 60 |
+
super().__init__()
|
| 61 |
+
self.num_heads = num_heads
|
| 62 |
+
self.q_proj = nn.Linear(embed_dim, embed_dim, bias=False)
|
| 63 |
+
self.k_proj = nn.Linear(kv_dim, embed_dim, bias=False)
|
| 64 |
+
self.v_proj = nn.Linear(kv_dim, embed_dim, bias=False)
|
| 65 |
+
|
| 66 |
+
self.multihead_attn = nn.MultiheadAttention(embed_dim, num_heads)
|
| 67 |
+
self.linear = nn.Linear(embed_dim, embed_dim)
|
| 68 |
+
self.dropout = nn.Dropout(drop_out_rate)
|
| 69 |
+
|
| 70 |
+
self.layer_norm = nn.LayerNorm(embed_dim)
|
| 71 |
+
self.ln_kv = nn.LayerNorm(kv_dim)
|
| 72 |
+
|
| 73 |
+
def forward(self, x, hidden_states, attn_mask=None, add_residual=False):
|
| 74 |
+
"""
|
| 75 |
+
Forward pass of the CrossAttention module.
|
| 76 |
+
|
| 77 |
+
Args:
|
| 78 |
+
x (torch.Tensor): Input tensor for key and value.
|
| 79 |
+
hidden_states (torch.Tensor): Input tensor for query.
|
| 80 |
+
attn_mask (torch.Tensor, optional): Attention mask. Default is None.
|
| 81 |
+
add_residual (bool): Whether to add residual connection. Default is False.
|
| 82 |
+
|
| 83 |
+
Returns:
|
| 84 |
+
torch.Tensor: Output tensor after cross-attention.
|
| 85 |
+
"""
|
| 86 |
+
normed_hidden_states = self.layer_norm(hidden_states)
|
| 87 |
+
query = self.q_proj(normed_hidden_states).permute(1, 0, 2)
|
| 88 |
+
|
| 89 |
+
x = self.ln_kv(x)
|
| 90 |
+
key = self.k_proj(x).permute(1, 0, 2)
|
| 91 |
+
value = self.v_proj(x).permute(1, 0, 2)
|
| 92 |
+
|
| 93 |
+
attn_output, _ = self.multihead_attn(query, key, value, attn_mask=attn_mask)
|
| 94 |
+
|
| 95 |
+
attn_output = attn_output.permute(1, 0, 2)
|
| 96 |
+
|
| 97 |
+
if add_residual:
|
| 98 |
+
attn_output = hidden_states + self.dropout(self.linear(attn_output))
|
| 99 |
+
else:
|
| 100 |
+
attn_output = self.dropout(self.linear(attn_output))
|
| 101 |
+
|
| 102 |
+
return attn_output
|
| 103 |
+
|
| 104 |
+
|
| 105 |
+
class AriaProjector(nn.Module):
|
| 106 |
+
"""
|
| 107 |
+
A projection module with one cross attention layer and one FFN layer, which projects ViT's outputs into MoE's inputs.
|
| 108 |
+
|
| 109 |
+
Args:
|
| 110 |
+
patch_to_query_dict (dict): Maps patch numbers to their corresponding query numbers,
|
| 111 |
+
e.g., {1225: 128, 4900: 256}. This allows for different query sizes based on image resolution.
|
| 112 |
+
embed_dim (int): Embedding dimension.
|
| 113 |
+
num_heads (int): Number of attention heads.
|
| 114 |
+
kv_dim (int): Dimension of key and value.
|
| 115 |
+
ff_dim (int): Hidden dimension of the feed-forward network.
|
| 116 |
+
output_dim (int): Output dimension.
|
| 117 |
+
norm_layer (nn.Module): Normalization layer. Default is nn.LayerNorm.
|
| 118 |
+
|
| 119 |
+
Outputs:
|
| 120 |
+
A tensor with the shape of (batch_size, query_number, output_dim)
|
| 121 |
+
"""
|
| 122 |
+
|
| 123 |
+
def __init__(
|
| 124 |
+
self,
|
| 125 |
+
patch_to_query_dict,
|
| 126 |
+
embed_dim,
|
| 127 |
+
num_heads,
|
| 128 |
+
kv_dim,
|
| 129 |
+
ff_dim,
|
| 130 |
+
output_dim,
|
| 131 |
+
norm_layer=nn.LayerNorm,
|
| 132 |
+
):
|
| 133 |
+
super().__init__()
|
| 134 |
+
self.patch_to_query_dict = patch_to_query_dict
|
| 135 |
+
self.embed_dim = embed_dim
|
| 136 |
+
self.num_heads = num_heads
|
| 137 |
+
|
| 138 |
+
self.query = nn.Parameter(
|
| 139 |
+
torch.zeros(max(patch_to_query_dict.values()), self.embed_dim)
|
| 140 |
+
)
|
| 141 |
+
|
| 142 |
+
trunc_normal_(self.query, std=0.02)
|
| 143 |
+
|
| 144 |
+
self.cross_attn = CrossAttention(kv_dim, embed_dim, num_heads)
|
| 145 |
+
|
| 146 |
+
self.ln_ffn = norm_layer(embed_dim)
|
| 147 |
+
self.ffn = FFN(embed_dim, ff_dim, output_dim)
|
| 148 |
+
|
| 149 |
+
self.apply(self._init_weights)
|
| 150 |
+
|
| 151 |
+
def _init_weights(self, m):
|
| 152 |
+
if isinstance(m, nn.Linear):
|
| 153 |
+
trunc_normal_(m.weight, std=0.02)
|
| 154 |
+
if isinstance(m, nn.Linear) and m.bias is not None:
|
| 155 |
+
nn.init.constant_(m.bias, 0)
|
| 156 |
+
elif isinstance(m, nn.LayerNorm):
|
| 157 |
+
nn.init.constant_(m.bias, 0)
|
| 158 |
+
nn.init.constant_(m.weight, 1.0)
|
| 159 |
+
|
| 160 |
+
def forward(self, x, attn_mask=None):
|
| 161 |
+
"""
|
| 162 |
+
Forward pass of the Projector module.
|
| 163 |
+
|
| 164 |
+
Args:
|
| 165 |
+
x (torch.Tensor): Input tensor of shape (batch_size, num_patches, kv_dim).
|
| 166 |
+
attn_mask (torch.Tensor, optional): Attention mask. Default is None.
|
| 167 |
+
|
| 168 |
+
Returns:
|
| 169 |
+
torch.Tensor: Output tensor of shape (batch_size, query_number, output_dim).
|
| 170 |
+
"""
|
| 171 |
+
bs = x.shape[0]
|
| 172 |
+
queries = self.query.unsqueeze(0).repeat(bs, 1, 1)
|
| 173 |
+
|
| 174 |
+
query_num = self.patch_to_query_dict.get(x.shape[1], None)
|
| 175 |
+
assert (
|
| 176 |
+
query_num is not None
|
| 177 |
+
), f"Query number for {x.shape[1]} patches is not provided"
|
| 178 |
+
|
| 179 |
+
queries = queries[:, :query_num, :]
|
| 180 |
+
|
| 181 |
+
if attn_mask is not None:
|
| 182 |
+
attn_mask = attn_mask.repeat_interleave(self.num_heads, 0)
|
| 183 |
+
attn_mask = attn_mask.unsqueeze(1).expand(-1, queries.size(1), -1)
|
| 184 |
+
|
| 185 |
+
attention_out = self.cross_attn(x, queries, attn_mask=attn_mask)
|
| 186 |
+
|
| 187 |
+
out = self.ffn(self.ln_ffn(attention_out))
|
| 188 |
+
|
| 189 |
+
return out
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"unk_token": {
|
| 3 |
+
"content": "<unk>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
}
|
| 9 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": false,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"add_prefix_space": true,
|
| 5 |
+
"added_tokens_decoder": {
|
| 6 |
+
"0": {
|
| 7 |
+
"content": "<unk>",
|
| 8 |
+
"lstrip": false,
|
| 9 |
+
"normalized": false,
|
| 10 |
+
"rstrip": false,
|
| 11 |
+
"single_word": false,
|
| 12 |
+
"special": true
|
| 13 |
+
}
|
| 14 |
+
},
|
| 15 |
+
"bos_token": null,
|
| 16 |
+
"chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% for message in messages %}<|im_start|>{{ message['role'] }}\n{% if message['content'] is string %}{{ message['content'] }}{% elif message['content'] is iterable %}{% for item in message['content'] %}{% if item['type'] == 'text' %}{{ item['text'] }}{% elif item['type'] == 'image' %}<fim_prefix><|img|><fim_suffix>{% endif %}{% endfor %}{% endif %}<|im_end|>\n{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant\n{% endif %}",
|
| 17 |
+
"clean_up_tokenization_spaces": false,
|
| 18 |
+
"eos_token": null,
|
| 19 |
+
"legacy": true,
|
| 20 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 21 |
+
"pad_token": null,
|
| 22 |
+
"sp_model_kwargs": {},
|
| 23 |
+
"spaces_between_special_tokens": false,
|
| 24 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 25 |
+
"unk_token": "<unk>",
|
| 26 |
+
"use_default_system_prompt": false
|
| 27 |
+
}
|
vision_encoder.py
ADDED
|
@@ -0,0 +1,152 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright 2024 Rhymes AI. All rights reserved.
|
| 2 |
+
#
|
| 3 |
+
# Licensed to the Apache Software Foundation (ASF) under one
|
| 4 |
+
# or more contributor license agreements. See the NOTICE file
|
| 5 |
+
# distributed with this work for additional information
|
| 6 |
+
# regarding copyright ownership. The ASF licenses this file
|
| 7 |
+
# to you under the Apache License, Version 2.0 (the
|
| 8 |
+
# "License"); you may not use this file except in compliance
|
| 9 |
+
# with the License. You may obtain a copy of the License at
|
| 10 |
+
#
|
| 11 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 12 |
+
#
|
| 13 |
+
# Unless required by applicable law or agreed to in writing,
|
| 14 |
+
# software distributed under the License is distributed on an
|
| 15 |
+
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
| 16 |
+
# KIND, either express or implied. See the License for the
|
| 17 |
+
# specific language governing permissions and limitations
|
| 18 |
+
# under the License.
|
| 19 |
+
|
| 20 |
+
"""PyTorch Aria vision transformer."""
|
| 21 |
+
|
| 22 |
+
from typing import Optional, Tuple, Union
|
| 23 |
+
|
| 24 |
+
import torch
|
| 25 |
+
import torch.utils.checkpoint
|
| 26 |
+
from transformers import SiglipVisionConfig, SiglipVisionModel
|
| 27 |
+
from transformers.modeling_outputs import BaseModelOutputWithPooling
|
| 28 |
+
from transformers.models.idefics2.modeling_idefics2 import Idefics2VisionTransformer
|
| 29 |
+
|
| 30 |
+
|
| 31 |
+
class AriaVisionConfig(SiglipVisionConfig):
|
| 32 |
+
"""Configuration class for AriaVisionModel."""
|
| 33 |
+
|
| 34 |
+
model_type = "aria_vision_model"
|
| 35 |
+
|
| 36 |
+
def __init__(
|
| 37 |
+
self,
|
| 38 |
+
**kwargs,
|
| 39 |
+
):
|
| 40 |
+
super().__init__(**kwargs)
|
| 41 |
+
|
| 42 |
+
|
| 43 |
+
class IdentityOp(torch.nn.Module):
|
| 44 |
+
"""
|
| 45 |
+
An identity operation that returns the input unchanged.
|
| 46 |
+
|
| 47 |
+
This can be used as a placeholder or to maintain architectural consistency
|
| 48 |
+
when a specific operation is not needed.
|
| 49 |
+
"""
|
| 50 |
+
|
| 51 |
+
def __init__(self, *args, **kwargs):
|
| 52 |
+
super().__init__()
|
| 53 |
+
|
| 54 |
+
def forward(self, x, *args, **kwargs):
|
| 55 |
+
return x
|
| 56 |
+
|
| 57 |
+
|
| 58 |
+
class AriaVisionTransformer(Idefics2VisionTransformer):
|
| 59 |
+
"""
|
| 60 |
+
Aria Vision Transformer model based on Idefics2VisionTransformer.
|
| 61 |
+
|
| 62 |
+
This class extends the original Idefics2VisionTransformer by removing the post-layernorm operation.
|
| 63 |
+
"""
|
| 64 |
+
|
| 65 |
+
def __init__(self, config: AriaVisionConfig):
|
| 66 |
+
super().__init__(config)
|
| 67 |
+
self.post_layernorm = IdentityOp()
|
| 68 |
+
|
| 69 |
+
|
| 70 |
+
class AriaVisionModel(SiglipVisionModel):
|
| 71 |
+
"""
|
| 72 |
+
Aria Vision Model extends SiglipVisionModel to support pixel_mask.
|
| 73 |
+
|
| 74 |
+
The pixel_mask is a 2D boolean tensor that indicates which pixels in the input
|
| 75 |
+
image are actual content and which are padding. It has the same height and width
|
| 76 |
+
as the input image, where:
|
| 77 |
+
- True (1) values represent pixels from the original image
|
| 78 |
+
- False (0) values represent padding pixels
|
| 79 |
+
|
| 80 |
+
This mask helps the model focus on the relevant parts of the image during processing.
|
| 81 |
+
"""
|
| 82 |
+
|
| 83 |
+
config_class = AriaVisionConfig
|
| 84 |
+
main_input_name = "pixel_values"
|
| 85 |
+
_supports_sdpa = False
|
| 86 |
+
|
| 87 |
+
def __init__(self, config: AriaVisionConfig):
|
| 88 |
+
super().__init__(config)
|
| 89 |
+
self.vision_model = AriaVisionTransformer(config)
|
| 90 |
+
|
| 91 |
+
# Initialize weights and apply final processing
|
| 92 |
+
self.post_init()
|
| 93 |
+
|
| 94 |
+
def forward(
|
| 95 |
+
self,
|
| 96 |
+
pixel_values: torch.Tensor,
|
| 97 |
+
pixel_mask: Optional[torch.BoolTensor] = None,
|
| 98 |
+
output_attentions: Optional[bool] = None,
|
| 99 |
+
output_hidden_states: Optional[bool] = None,
|
| 100 |
+
return_dict: Optional[bool] = None,
|
| 101 |
+
) -> Union[Tuple, BaseModelOutputWithPooling]:
|
| 102 |
+
"""
|
| 103 |
+
Forward pass of the AriaVisionModel.
|
| 104 |
+
|
| 105 |
+
Args:
|
| 106 |
+
pixel_values (torch.Tensor): The pixel values of the input images.
|
| 107 |
+
pixel_mask (Optional[torch.BoolTensor]): Mask for the pixel values.
|
| 108 |
+
output_attentions (Optional[bool]): Whether to output attentions.
|
| 109 |
+
output_hidden_states (Optional[bool]): Whether to output hidden states.
|
| 110 |
+
return_dict (Optional[bool]): Whether to return a ModelOutput object.
|
| 111 |
+
|
| 112 |
+
Returns:
|
| 113 |
+
Union[Tuple, BaseModelOutputWithPooling]: The model's output.
|
| 114 |
+
"""
|
| 115 |
+
return_dict = (
|
| 116 |
+
return_dict if return_dict is not None else self.config.use_return_dict
|
| 117 |
+
)
|
| 118 |
+
patch_attention_mask = self._create_patch_attention_mask(pixel_mask)
|
| 119 |
+
|
| 120 |
+
vit_oup = self.vision_model(
|
| 121 |
+
pixel_values=pixel_values,
|
| 122 |
+
patch_attention_mask=patch_attention_mask,
|
| 123 |
+
output_attentions=output_attentions,
|
| 124 |
+
output_hidden_states=output_hidden_states,
|
| 125 |
+
return_dict=return_dict,
|
| 126 |
+
)
|
| 127 |
+
|
| 128 |
+
image_atts = self._create_image_attention_mask(patch_attention_mask)
|
| 129 |
+
|
| 130 |
+
return vit_oup, image_atts
|
| 131 |
+
|
| 132 |
+
def _create_patch_attention_mask(self, pixel_mask):
|
| 133 |
+
if pixel_mask is None:
|
| 134 |
+
return None
|
| 135 |
+
|
| 136 |
+
patches_subgrid = pixel_mask.unfold(
|
| 137 |
+
dimension=1,
|
| 138 |
+
size=self.vision_model.config.patch_size,
|
| 139 |
+
step=self.vision_model.config.patch_size,
|
| 140 |
+
).unfold(
|
| 141 |
+
dimension=2,
|
| 142 |
+
size=self.vision_model.config.patch_size,
|
| 143 |
+
step=self.vision_model.config.patch_size,
|
| 144 |
+
)
|
| 145 |
+
return (patches_subgrid.sum(dim=(-1, -2)) > 0).bool()
|
| 146 |
+
|
| 147 |
+
def _create_image_attention_mask(self, patch_attention_mask):
|
| 148 |
+
if patch_attention_mask is None:
|
| 149 |
+
return None
|
| 150 |
+
|
| 151 |
+
flattened_mask = patch_attention_mask.flatten(1)
|
| 152 |
+
return torch.logical_not(flattened_mask)
|
vision_processor.py
ADDED
|
@@ -0,0 +1,321 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright 2024 Rhymes AI. All rights reserved.
|
| 2 |
+
#
|
| 3 |
+
# Licensed to the Apache Software Foundation (ASF) under one
|
| 4 |
+
# or more contributor license agreements. See the NOTICE file
|
| 5 |
+
# distributed with this work for additional information
|
| 6 |
+
# regarding copyright ownership. The ASF licenses this file
|
| 7 |
+
# to you under the Apache License, Version 2.0 (the
|
| 8 |
+
# "License"); you may not use this file except in compliance
|
| 9 |
+
# with the License. You may obtain a copy of the License at
|
| 10 |
+
#
|
| 11 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 12 |
+
#
|
| 13 |
+
# Unless required by applicable law or agreed to in writing,
|
| 14 |
+
# software distributed under the License is distributed on an
|
| 15 |
+
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
| 16 |
+
# KIND, either express or implied. See the License for the
|
| 17 |
+
# specific language governing permissions and limitations
|
| 18 |
+
# under the License.
|
| 19 |
+
|
| 20 |
+
from typing import List, Optional, Union
|
| 21 |
+
|
| 22 |
+
import numpy as np
|
| 23 |
+
import torch
|
| 24 |
+
from PIL import Image, ImageOps
|
| 25 |
+
from torchvision import transforms
|
| 26 |
+
from transformers import BaseImageProcessor, BatchFeature, TensorType
|
| 27 |
+
|
| 28 |
+
|
| 29 |
+
def _select_best_resolution(
|
| 30 |
+
img_width: int, img_height: int, target_ratios: List[List[int]], patch_size: int
|
| 31 |
+
):
|
| 32 |
+
"""
|
| 33 |
+
Selects the best resolution from a list of possible resolutions based on the original size.
|
| 34 |
+
|
| 35 |
+
Args:
|
| 36 |
+
img_width: the original widths of images.
|
| 37 |
+
img_height: the original heights of images.
|
| 38 |
+
target_ratios (2d numpy array): dimension size (M,2)
|
| 39 |
+
patch_size (int): image patch size
|
| 40 |
+
|
| 41 |
+
Returns:
|
| 42 |
+
tuple: The best fit resolution in the format (width, height).
|
| 43 |
+
"""
|
| 44 |
+
|
| 45 |
+
aspect_ratio = img_width / img_height
|
| 46 |
+
best_ratio_diff = float("inf")
|
| 47 |
+
best_ratio_w, best_ratio_h = 1, 1
|
| 48 |
+
area = np.int32(img_width) * np.int32(img_height)
|
| 49 |
+
for ratio in target_ratios:
|
| 50 |
+
target_aspect_ratio = ratio[0] / ratio[1]
|
| 51 |
+
ratio_diff = abs(aspect_ratio - target_aspect_ratio)
|
| 52 |
+
if ratio_diff < best_ratio_diff:
|
| 53 |
+
best_ratio_diff = ratio_diff
|
| 54 |
+
best_ratio_w, best_ratio_h = ratio[0], ratio[1]
|
| 55 |
+
elif (
|
| 56 |
+
ratio_diff == best_ratio_diff
|
| 57 |
+
and area > 0.5 * patch_size * patch_size * ratio[0] * ratio[1]
|
| 58 |
+
):
|
| 59 |
+
best_ratio_w, best_ratio_h = ratio[0], ratio[1]
|
| 60 |
+
|
| 61 |
+
return best_ratio_w, best_ratio_h
|
| 62 |
+
|
| 63 |
+
|
| 64 |
+
def _split_image(
|
| 65 |
+
image: Image.Image,
|
| 66 |
+
split_image: bool,
|
| 67 |
+
split_ratio: List[List[int]],
|
| 68 |
+
patch_size: int,
|
| 69 |
+
) -> List[Image.Image]:
|
| 70 |
+
"""
|
| 71 |
+
Split image into multiple patches
|
| 72 |
+
|
| 73 |
+
Args:
|
| 74 |
+
image (PIL.Image): Input image.
|
| 75 |
+
split_image (bool): Whether to split the image into patches.
|
| 76 |
+
split_ratio (2d numpy array): dimension size (M,2)
|
| 77 |
+
patch_size (int): image patch size
|
| 78 |
+
|
| 79 |
+
Returns:
|
| 80 |
+
List[PIL.Image]: List of splitted images.
|
| 81 |
+
"""
|
| 82 |
+
if split_image:
|
| 83 |
+
ratio_width, ratio_height = _select_best_resolution(
|
| 84 |
+
image.width, image.height, split_ratio, patch_size
|
| 85 |
+
)
|
| 86 |
+
resize_width = patch_size * ratio_width
|
| 87 |
+
resize_height = patch_size * ratio_height
|
| 88 |
+
blocks = ratio_width * ratio_height
|
| 89 |
+
resized_img = image.resize((resize_width, resize_height))
|
| 90 |
+
processed_images = []
|
| 91 |
+
for i in range(blocks):
|
| 92 |
+
box = (
|
| 93 |
+
(i % (resize_width // patch_size)) * patch_size,
|
| 94 |
+
(i // (resize_width // patch_size)) * patch_size,
|
| 95 |
+
((i % (resize_width // patch_size)) + 1) * patch_size,
|
| 96 |
+
((i // (resize_width // patch_size)) + 1) * patch_size,
|
| 97 |
+
)
|
| 98 |
+
# split the image
|
| 99 |
+
split_img = resized_img.crop(box)
|
| 100 |
+
processed_images.append(split_img)
|
| 101 |
+
assert len(processed_images) == blocks
|
| 102 |
+
if len(processed_images) != 1:
|
| 103 |
+
processed_images.insert(0, image)
|
| 104 |
+
return processed_images
|
| 105 |
+
else:
|
| 106 |
+
return [image]
|
| 107 |
+
|
| 108 |
+
|
| 109 |
+
def keep_ratio_resize_and_pixel_mask(
|
| 110 |
+
img: Image.Image, max_size, min_size=336, padding_value=0
|
| 111 |
+
):
|
| 112 |
+
"""
|
| 113 |
+
Resize an image while maintaining aspect ratio and create a pixel mask.
|
| 114 |
+
|
| 115 |
+
Args:
|
| 116 |
+
img (PIL.Image): Input image.
|
| 117 |
+
max_size (int): Maximum size for the larger dimension of the image.
|
| 118 |
+
min_size (int, optional): Minimum size for the smaller dimension. Defaults to 336.
|
| 119 |
+
padding_value (int, optional): Value used for padding. Defaults to 0.
|
| 120 |
+
|
| 121 |
+
Returns:
|
| 122 |
+
tuple: A tuple containing:
|
| 123 |
+
- PIL.Image: Resized and padded image.
|
| 124 |
+
- torch.Tensor: Boolean pixel mask. This mask is a 2D tensor of shape (max_size, max_size) where:
|
| 125 |
+
- True (1) values indicate pixels that belong to the original resized image.
|
| 126 |
+
- False (0) values indicate pixels that are part of the padding.
|
| 127 |
+
The mask helps distinguish between actual image content and padded areas in subsequent processing steps.
|
| 128 |
+
"""
|
| 129 |
+
img = img.convert("RGB")
|
| 130 |
+
# rescale the given image, keep the aspect ratio
|
| 131 |
+
scale = max_size / max(img.size)
|
| 132 |
+
|
| 133 |
+
w, h = img.size
|
| 134 |
+
if w >= h:
|
| 135 |
+
new_size = (max_size, max(int(h * scale), min_size)) # w, h
|
| 136 |
+
else:
|
| 137 |
+
new_size = (max(int(w * scale), min_size), max_size) # w, h
|
| 138 |
+
|
| 139 |
+
img_resized = img.resize(new_size, resample=Image.Resampling.BICUBIC)
|
| 140 |
+
|
| 141 |
+
# padding the right/bottom
|
| 142 |
+
padding_right, padding_bottom = max_size - new_size[0], max_size - new_size[1]
|
| 143 |
+
img_padded = ImageOps.expand(
|
| 144 |
+
img_resized, (0, 0, padding_right, padding_bottom), fill=padding_value
|
| 145 |
+
)
|
| 146 |
+
|
| 147 |
+
# Create a pixel mask
|
| 148 |
+
pixel_mask = torch.zeros(max_size, max_size)
|
| 149 |
+
pixel_mask[: new_size[1], : new_size[0]] = 1
|
| 150 |
+
pixel_mask = pixel_mask.bool()
|
| 151 |
+
return img_padded, pixel_mask
|
| 152 |
+
|
| 153 |
+
|
| 154 |
+
class AriaVisionProcessor(BaseImageProcessor):
|
| 155 |
+
"""
|
| 156 |
+
A vision processor for the Aria model that handles image preprocessing.
|
| 157 |
+
"""
|
| 158 |
+
|
| 159 |
+
def __init__(
|
| 160 |
+
self,
|
| 161 |
+
max_image_size=980,
|
| 162 |
+
min_image_size=336,
|
| 163 |
+
image_mean=[0.5, 0.5, 0.5],
|
| 164 |
+
image_std=[0.5, 0.5, 0.5],
|
| 165 |
+
**kwargs,
|
| 166 |
+
):
|
| 167 |
+
"""
|
| 168 |
+
Initialize the AriaVisionProcessor.
|
| 169 |
+
|
| 170 |
+
Args:
|
| 171 |
+
max_image_size (int, optional): Maximum image size. Defaults to 980.
|
| 172 |
+
min_image_size (int, optional): Minimum image size. Defaults to 336.
|
| 173 |
+
mean (list, optional): Mean values for normalization. Defaults to [0.5, 0.5, 0.5].
|
| 174 |
+
std (list, optional): Standard deviation values for normalization. Defaults to [0.5, 0.5, 0.5].
|
| 175 |
+
"""
|
| 176 |
+
super().__init__(**kwargs)
|
| 177 |
+
|
| 178 |
+
self.max_image_size = max_image_size
|
| 179 |
+
self.min_image_size = min_image_size
|
| 180 |
+
self.image_mean = image_mean
|
| 181 |
+
self.image_std = image_std
|
| 182 |
+
self.auto_map = {
|
| 183 |
+
"AutoProcessor": "processing_aria.AriaProcessor",
|
| 184 |
+
"AutoImageProcessor": "vision_processor.AriaVisionProcessor",
|
| 185 |
+
}
|
| 186 |
+
|
| 187 |
+
# we make the transform a property so that it is lazily initialized,
|
| 188 |
+
# this could avoid the error "TypeError: Object of type Normalize is not JSON serializable"
|
| 189 |
+
# when we used save_pretrained or from_pretrained.
|
| 190 |
+
self._transform = None
|
| 191 |
+
self._set_processor_class("AriaProcessor")
|
| 192 |
+
|
| 193 |
+
@property
|
| 194 |
+
def transform(self):
|
| 195 |
+
if self._transform is None:
|
| 196 |
+
# Recreate the transform when accessed
|
| 197 |
+
self._transform = transforms.Compose(
|
| 198 |
+
[
|
| 199 |
+
transforms.ToTensor(),
|
| 200 |
+
transforms.Normalize(self.image_mean, self.image_std),
|
| 201 |
+
]
|
| 202 |
+
)
|
| 203 |
+
return self._transform
|
| 204 |
+
|
| 205 |
+
def __call__(
|
| 206 |
+
self,
|
| 207 |
+
images: Union[Image.Image, List[Image.Image]],
|
| 208 |
+
max_image_size: Optional[int] = 980,
|
| 209 |
+
min_image_size: Optional[int] = 336,
|
| 210 |
+
return_tensors: Optional[Union[str, TensorType]] = "pt",
|
| 211 |
+
split_image: Optional[bool] = False,
|
| 212 |
+
split_ratio: Optional[List[List[int]]] = [
|
| 213 |
+
[1, 2],
|
| 214 |
+
[1, 3],
|
| 215 |
+
[1, 4],
|
| 216 |
+
[1, 5],
|
| 217 |
+
[1, 6],
|
| 218 |
+
[1, 7],
|
| 219 |
+
[1, 8],
|
| 220 |
+
[2, 4],
|
| 221 |
+
[2, 3],
|
| 222 |
+
[2, 2],
|
| 223 |
+
[2, 1],
|
| 224 |
+
[3, 1],
|
| 225 |
+
[3, 2],
|
| 226 |
+
[4, 1],
|
| 227 |
+
[4, 2],
|
| 228 |
+
[5, 1],
|
| 229 |
+
[6, 1],
|
| 230 |
+
[7, 1],
|
| 231 |
+
[8, 1],
|
| 232 |
+
],
|
| 233 |
+
):
|
| 234 |
+
"""
|
| 235 |
+
Process a list of images.
|
| 236 |
+
|
| 237 |
+
Args:
|
| 238 |
+
images (list): List of PIL.Image objects.
|
| 239 |
+
max_image_size (int, optional): Override the default max image size. Defaults to None.
|
| 240 |
+
return_tensors (str or TensorType, optional): The type of tensor to return. Defaults to "pt".
|
| 241 |
+
split_image (bool, optional): Whether to split the image. Defaults to False.
|
| 242 |
+
split_ratio (list, optional): The ratio for splitting the image. Defaults to a list of common split ratios.
|
| 243 |
+
Returns:
|
| 244 |
+
BatchFeature: A BatchFeature object containing:
|
| 245 |
+
- 'pixel_values': Tensor of processed image pixel values.
|
| 246 |
+
- 'pixel_mask': Boolean pixel mask. This mask is a 2D tensor of shape (max_size, max_size) where:
|
| 247 |
+
- True (1) values indicate pixels that belong to the original resized image.
|
| 248 |
+
- False (0) values indicate pixels that are part of the padding.
|
| 249 |
+
The mask helps distinguish between actual image content and padded areas in subsequent processing steps.
|
| 250 |
+
- 'num_crops': Tensor of the number of crops for each image.
|
| 251 |
+
"""
|
| 252 |
+
max_size = self.max_image_size if max_image_size is None else max_image_size
|
| 253 |
+
min_size = self.min_image_size if min_image_size is None else min_image_size
|
| 254 |
+
|
| 255 |
+
if max_size not in [490, 980]:
|
| 256 |
+
raise ValueError("max_image_size must be either 490 or 980")
|
| 257 |
+
|
| 258 |
+
if isinstance(images, Image.Image):
|
| 259 |
+
images = [images]
|
| 260 |
+
|
| 261 |
+
pixel_values = []
|
| 262 |
+
pixel_masks = []
|
| 263 |
+
num_crops = []
|
| 264 |
+
|
| 265 |
+
for image in images:
|
| 266 |
+
crop_images = _split_image(image, split_image, split_ratio, max_size)
|
| 267 |
+
num_crops.append(torch.tensor(len(crop_images)))
|
| 268 |
+
for crop_image in crop_images:
|
| 269 |
+
img_padded, pixel_mask = keep_ratio_resize_and_pixel_mask(
|
| 270 |
+
crop_image, max_size, min_size
|
| 271 |
+
)
|
| 272 |
+
img_padded = self.transform(img_padded)
|
| 273 |
+
pixel_values.append(img_padded)
|
| 274 |
+
pixel_masks.append(pixel_mask)
|
| 275 |
+
|
| 276 |
+
return BatchFeature(
|
| 277 |
+
data={
|
| 278 |
+
"pixel_values": torch.stack(pixel_values),
|
| 279 |
+
"pixel_mask": torch.stack(pixel_masks),
|
| 280 |
+
"num_crops": torch.stack(num_crops),
|
| 281 |
+
},
|
| 282 |
+
tensor_type=return_tensors,
|
| 283 |
+
)
|
| 284 |
+
|
| 285 |
+
def preprocess(
|
| 286 |
+
self,
|
| 287 |
+
images,
|
| 288 |
+
max_image_size=None,
|
| 289 |
+
min_image_size=None,
|
| 290 |
+
return_tensors: Optional[Union[str, TensorType]] = None,
|
| 291 |
+
split_image: Optional[bool] = False,
|
| 292 |
+
split_ratio: Optional[List[List[int]]] = [
|
| 293 |
+
[1, 2],
|
| 294 |
+
[1, 3],
|
| 295 |
+
[1, 4],
|
| 296 |
+
[1, 5],
|
| 297 |
+
[1, 6],
|
| 298 |
+
[1, 7],
|
| 299 |
+
[1, 8],
|
| 300 |
+
[2, 4],
|
| 301 |
+
[2, 3],
|
| 302 |
+
[2, 2],
|
| 303 |
+
[2, 1],
|
| 304 |
+
[3, 1],
|
| 305 |
+
[3, 2],
|
| 306 |
+
[4, 1],
|
| 307 |
+
[4, 2],
|
| 308 |
+
[5, 1],
|
| 309 |
+
[6, 1],
|
| 310 |
+
[7, 1],
|
| 311 |
+
[8, 1],
|
| 312 |
+
],
|
| 313 |
+
):
|
| 314 |
+
return self.__call__(
|
| 315 |
+
images,
|
| 316 |
+
max_image_size=max_image_size,
|
| 317 |
+
min_image_size=min_image_size,
|
| 318 |
+
return_tensors=return_tensors,
|
| 319 |
+
split_image=split_image,
|
| 320 |
+
split_ratio=split_ratio,
|
| 321 |
+
)
|