leejuhyoeng
commited on
Commit
•
6251410
1
Parent(s):
744964d
Training in progress, step 353
Browse files- config.json +58 -0
- training_args.bin +3 -0
config.json
ADDED
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "Salesforce/blip2-opt-2.7b",
|
3 |
+
"architectures": [
|
4 |
+
"Blip2ForConditionalGeneration"
|
5 |
+
],
|
6 |
+
"initializer_factor": 1.0,
|
7 |
+
"initializer_range": 0.02,
|
8 |
+
"model_type": "blip-2",
|
9 |
+
"num_query_tokens": 32,
|
10 |
+
"qformer_config": {
|
11 |
+
"classifier_dropout": null,
|
12 |
+
"model_type": "blip_2_qformer"
|
13 |
+
},
|
14 |
+
"text_config": {
|
15 |
+
"_name_or_path": "heegyu/kogpt-j-350m",
|
16 |
+
"activation_function": "gelu_new",
|
17 |
+
"architectures": [
|
18 |
+
"GPTJForCausalLM"
|
19 |
+
],
|
20 |
+
"attn_pdrop": 0.1,
|
21 |
+
"bos_token_id": 1,
|
22 |
+
"embd_pdrop": 0.1,
|
23 |
+
"eos_token_id": 2,
|
24 |
+
"gradient_checkpointing": false,
|
25 |
+
"initializer_range": 0.02,
|
26 |
+
"layer_norm_epsilon": 1e-05,
|
27 |
+
"model_type": "gptj",
|
28 |
+
"n_embd": 1024,
|
29 |
+
"n_head": 16,
|
30 |
+
"n_inner": null,
|
31 |
+
"n_layer": 20,
|
32 |
+
"n_positions": 1024,
|
33 |
+
"pad_token_id": 0,
|
34 |
+
"resid_pdrop": 0.0,
|
35 |
+
"rotary_dim": 32,
|
36 |
+
"scale_attn_weights": true,
|
37 |
+
"summary_activation": null,
|
38 |
+
"summary_first_dropout": 0.1,
|
39 |
+
"summary_proj_to_labels": true,
|
40 |
+
"summary_type": "cls_index",
|
41 |
+
"summary_use_proj": true,
|
42 |
+
"tie_word_embeddings": false,
|
43 |
+
"tokenizer_class": "GPT2Tokenizer",
|
44 |
+
"torch_dtype": "float32",
|
45 |
+
"unk_token_id": 5,
|
46 |
+
"vocab_size": 51200
|
47 |
+
},
|
48 |
+
"torch_dtype": "float32",
|
49 |
+
"transformers_version": "4.36.2",
|
50 |
+
"use_decoder_only_language_model": true,
|
51 |
+
"vision_config": {
|
52 |
+
"dropout": 0.0,
|
53 |
+
"initializer_factor": 1.0,
|
54 |
+
"model_type": "blip_2_vision_model",
|
55 |
+
"num_channels": 3,
|
56 |
+
"projection_dim": 512
|
57 |
+
}
|
58 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5beaa7b666f6d5169746c5764067fcee177eb70ab64543f290fdd4cd6665cc24
|
3 |
+
size 4664
|