Hugging Face's logo

pig4431
/
TextGPT4V-7B-LORA

PEFT
llava
Model card Files Community
1
pig4431 commited on
Commit
c0ea348
•
1 Parent(s): 19200d7

Upload 5 files

Browse files
Files changed (5) hide show
  1. adapter_config.json +26 -0
  2. adapter_model.bin +3 -0
  3. config.json +41 -0
  4. non_lora_trainables.bin +3 -0
  5. trainer_state.json +4255 -0
adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "X2FD/LVIS-Instruct4v-LLaVA-7b",
4
+ "bias": "none",
5
+ "fan_in_fan_out": false,
6
+ "inference_mode": true,
7
+ "init_lora_weights": true,
8
+ "layers_pattern": null,
9
+ "layers_to_transform": null,
10
+ "lora_alpha": 256,
11
+ "lora_dropout": 0.05,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 128,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "q_proj",
18
+ "up_proj",
19
+ "gate_proj",
20
+ "v_proj",
21
+ "down_proj",
22
+ "k_proj",
23
+ "o_proj"
24
+ ],
25
+ "task_type": "CAUSAL_LM"
26
+ }
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:712649f9143e7b8d19eee64cf05d91a9cf99ea15e273ed4957e5eff1db5ec1ff
3
+ size 639786637
config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "X2FD/LVIS-Instruct4v-LLaVA-7b",
3
+ "architectures": [
4
+ "LlavaLlamaForCausalLM"
5
+ ],
6
+ "bos_token_id": 1,
7
+ "eos_token_id": 2,
8
+ "freeze_mm_mlp_adapter": false,
9
+ "hidden_act": "silu",
10
+ "hidden_size": 4096,
11
+ "image_aspect_ratio": "pad",
12
+ "image_grid_pinpoints": null,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 11008,
15
+ "max_position_embeddings": 4096,
16
+ "mm_hidden_size": 1024,
17
+ "mm_projector_lr": 2e-05,
18
+ "mm_projector_type": "mlp2x_gelu",
19
+ "mm_use_im_patch_token": false,
20
+ "mm_use_im_start_end": false,
21
+ "mm_vision_select_feature": "patch",
22
+ "mm_vision_select_layer": -2,
23
+ "mm_vision_tower": "openai/clip-vit-large-patch14-336",
24
+ "model_type": "llava",
25
+ "num_attention_heads": 32,
26
+ "num_hidden_layers": 32,
27
+ "num_key_value_heads": 32,
28
+ "pad_token_id": 0,
29
+ "pretraining_tp": 1,
30
+ "rms_norm_eps": 1e-05,
31
+ "rope_scaling": null,
32
+ "tie_word_embeddings": false,
33
+ "tokenizer_model_max_length": 2048,
34
+ "tokenizer_padding_side": "right",
35
+ "torch_dtype": "bfloat16",
36
+ "transformers_version": "4.31.0",
37
+ "tune_mm_mlp_adapter": false,
38
+ "use_cache": true,
39
+ "use_mm_proj": true,
40
+ "vocab_size": 32000
41
+ }
non_lora_trainables.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03e3ea410c6b079b315eafc2a6f3e33c99f3232f63962aa509fb1323f0c8b71e
3
+ size 41961191
trainer_state.json ADDED
@@ -0,0 +1,4255 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 3.0,
5
+ "global_step": 705,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 9.090909090909091e-06,
13
+ "loss": 1.574,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.01,
18
+ "learning_rate": 1.8181818181818182e-05,
19
+ "loss": 1.618,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.01,
24
+ "learning_rate": 2.7272727272727273e-05,
25
+ "loss": 1.5453,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.02,
30
+ "learning_rate": 3.6363636363636364e-05,
31
+ "loss": 1.4047,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.02,
36
+ "learning_rate": 4.545454545454546e-05,
37
+ "loss": 1.4909,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.03,
42
+ "learning_rate": 5.4545454545454546e-05,
43
+ "loss": 1.4884,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.03,
48
+ "learning_rate": 6.363636363636364e-05,
49
+ "loss": 1.449,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.03,
54
+ "learning_rate": 7.272727272727273e-05,
55
+ "loss": 1.4063,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.04,
60
+ "learning_rate": 8.181818181818183e-05,
61
+ "loss": 1.4595,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.04,
66
+ "learning_rate": 9.090909090909092e-05,
67
+ "loss": 1.3918,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.05,
72
+ "learning_rate": 0.0001,
73
+ "loss": 1.3847,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.05,
78
+ "learning_rate": 0.00010909090909090909,
79
+ "loss": 1.4178,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.06,
84
+ "learning_rate": 0.0001181818181818182,
85
+ "loss": 1.3751,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.06,
90
+ "learning_rate": 0.00012727272727272728,
91
+ "loss": 1.3377,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.06,
96
+ "learning_rate": 0.00013636363636363637,
97
+ "loss": 1.3761,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.07,
102
+ "learning_rate": 0.00014545454545454546,
103
+ "loss": 1.3358,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.07,
108
+ "learning_rate": 0.00015454545454545454,
109
+ "loss": 1.3477,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.08,
114
+ "learning_rate": 0.00016363636363636366,
115
+ "loss": 1.3174,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.08,
120
+ "learning_rate": 0.00017272727272727275,
121
+ "loss": 1.3381,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.09,
126
+ "learning_rate": 0.00018181818181818183,
127
+ "loss": 1.3524,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.09,
132
+ "learning_rate": 0.00019090909090909092,
133
+ "loss": 1.3306,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.09,
138
+ "learning_rate": 0.0002,
139
+ "loss": 1.3218,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.1,
144
+ "learning_rate": 0.00019999894214150818,
145
+ "loss": 1.3003,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.1,
150
+ "learning_rate": 0.00019999576858841395,
151
+ "loss": 1.3422,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.11,
156
+ "learning_rate": 0.00019999047940786073,
157
+ "loss": 1.3054,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.11,
162
+ "learning_rate": 0.00019998307471175264,
163
+ "loss": 1.3046,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.11,
168
+ "learning_rate": 0.00019997355465675205,
169
+ "loss": 1.3248,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.12,
174
+ "learning_rate": 0.00019996191944427638,
175
+ "loss": 1.3328,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.12,
180
+ "learning_rate": 0.00019994816932049383,
181
+ "loss": 1.3285,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.13,
186
+ "learning_rate": 0.0001999323045763181,
187
+ "loss": 1.3223,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.13,
192
+ "learning_rate": 0.00019991432554740225,
193
+ "loss": 1.3395,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.14,
198
+ "learning_rate": 0.0001998942326141317,
199
+ "loss": 1.3085,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.14,
204
+ "learning_rate": 0.000199872026201616,
205
+ "loss": 1.329,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.14,
210
+ "learning_rate": 0.00019984770677968,
211
+ "loss": 1.2921,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.15,
216
+ "learning_rate": 0.00019982127486285384,
217
+ "loss": 1.3047,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.15,
222
+ "learning_rate": 0.0001997927310103621,
223
+ "loss": 1.2923,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.16,
228
+ "learning_rate": 0.00019976207582611189,
229
+ "loss": 1.2481,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.16,
234
+ "learning_rate": 0.00019972930995868014,
235
+ "loss": 1.2574,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.17,
240
+ "learning_rate": 0.0001996944341012999,
241
+ "loss": 1.3083,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.17,
246
+ "learning_rate": 0.0001996574489918456,
247
+ "loss": 1.2613,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.17,
252
+ "learning_rate": 0.00019961835541281746,
253
+ "loss": 1.2648,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.18,
258
+ "learning_rate": 0.00019957715419132498,
259
+ "loss": 1.2976,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 0.18,
264
+ "learning_rate": 0.00019953384619906945,
265
+ "loss": 1.3284,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 0.19,
270
+ "learning_rate": 0.00019948843235232535,
271
+ "loss": 1.2759,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 0.19,
276
+ "learning_rate": 0.0001994409136119212,
277
+ "loss": 1.2619,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 0.2,
282
+ "learning_rate": 0.00019939129098321904,
283
+ "loss": 1.2225,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 0.2,
288
+ "learning_rate": 0.00019933956551609322,
289
+ "loss": 1.3009,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 0.2,
294
+ "learning_rate": 0.00019928573830490826,
295
+ "loss": 1.3523,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 0.21,
300
+ "learning_rate": 0.00019922981048849564,
301
+ "loss": 1.2688,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 0.21,
306
+ "learning_rate": 0.00019917178325012963,
307
+ "loss": 1.2939,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 0.22,
312
+ "learning_rate": 0.00019911165781750237,
313
+ "loss": 1.2932,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 0.22,
318
+ "learning_rate": 0.00019904943546269785,
319
+ "loss": 1.2886,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 0.23,
324
+ "learning_rate": 0.00019898511750216505,
325
+ "loss": 1.2656,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 0.23,
330
+ "learning_rate": 0.0001989187052966899,
331
+ "loss": 1.2864,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 0.23,
336
+ "learning_rate": 0.00019885020025136677,
337
+ "loss": 1.287,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 0.24,
342
+ "learning_rate": 0.0001987796038155685,
343
+ "loss": 1.2534,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 0.24,
348
+ "learning_rate": 0.0001987069174829159,
349
+ "loss": 1.3241,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 0.25,
354
+ "learning_rate": 0.00019863214279124608,
355
+ "loss": 1.2915,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 0.25,
360
+ "learning_rate": 0.00019855528132257984,
361
+ "loss": 1.2197,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 0.26,
366
+ "learning_rate": 0.00019847633470308833,
367
+ "loss": 1.2212,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 0.26,
372
+ "learning_rate": 0.00019839530460305862,
373
+ "loss": 1.2421,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 0.26,
378
+ "learning_rate": 0.00019831219273685826,
379
+ "loss": 1.2741,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 0.27,
384
+ "learning_rate": 0.00019822700086289915,
385
+ "loss": 1.2633,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 0.27,
390
+ "learning_rate": 0.00019813973078360025,
391
+ "loss": 1.2529,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 0.28,
396
+ "learning_rate": 0.0001980503843453494,
397
+ "loss": 1.2905,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 0.28,
402
+ "learning_rate": 0.00019795896343846437,
403
+ "loss": 1.2651,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 0.29,
408
+ "learning_rate": 0.00019786546999715285,
409
+ "loss": 1.2507,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 0.29,
414
+ "learning_rate": 0.00019776990599947147,
415
+ "loss": 1.2113,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 0.29,
420
+ "learning_rate": 0.00019767227346728392,
421
+ "loss": 1.262,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 0.3,
426
+ "learning_rate": 0.00019757257446621827,
427
+ "loss": 1.2706,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 0.3,
432
+ "learning_rate": 0.00019747081110562322,
433
+ "loss": 1.2491,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 0.31,
438
+ "learning_rate": 0.0001973669855385235,
439
+ "loss": 1.3059,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 0.31,
444
+ "learning_rate": 0.00019726109996157424,
445
+ "loss": 1.2589,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 0.31,
450
+ "learning_rate": 0.0001971531566150145,
451
+ "loss": 1.2999,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 0.32,
456
+ "learning_rate": 0.00019704315778262016,
457
+ "loss": 1.2421,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 0.32,
462
+ "learning_rate": 0.00019693110579165513,
463
+ "loss": 1.3066,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 0.33,
468
+ "learning_rate": 0.00019681700301282234,
469
+ "loss": 1.2259,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 0.33,
474
+ "learning_rate": 0.00019670085186021375,
475
+ "loss": 1.2563,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 0.34,
480
+ "learning_rate": 0.000196582654791259,
481
+ "loss": 1.2243,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 0.34,
486
+ "learning_rate": 0.00019646241430667353,
487
+ "loss": 1.2172,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 0.34,
492
+ "learning_rate": 0.0001963401329504057,
493
+ "loss": 1.2653,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 0.35,
498
+ "learning_rate": 0.00019621581330958295,
499
+ "loss": 1.3283,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 0.35,
504
+ "learning_rate": 0.000196089458014457,
505
+ "loss": 1.2118,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 0.36,
510
+ "learning_rate": 0.00019596106973834835,
511
+ "loss": 1.2633,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 0.36,
516
+ "learning_rate": 0.0001958306511975895,
517
+ "loss": 1.2357,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 0.37,
522
+ "learning_rate": 0.00019569820515146768,
523
+ "loss": 1.2624,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 0.37,
528
+ "learning_rate": 0.0001955637344021664,
529
+ "loss": 1.2221,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 0.37,
534
+ "learning_rate": 0.00019542724179470616,
535
+ "loss": 1.2304,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 0.38,
540
+ "learning_rate": 0.0001952887302168842,
541
+ "loss": 1.235,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 0.38,
546
+ "learning_rate": 0.00019514820259921352,
547
+ "loss": 1.2106,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 0.39,
552
+ "learning_rate": 0.00019500566191486075,
553
+ "loss": 1.2013,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 0.39,
558
+ "learning_rate": 0.00019486111117958342,
559
+ "loss": 1.2796,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 0.4,
564
+ "learning_rate": 0.00019471455345166595,
565
+ "loss": 1.195,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 0.4,
570
+ "learning_rate": 0.00019456599183185507,
571
+ "loss": 1.2813,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 0.4,
576
+ "learning_rate": 0.00019441542946329422,
577
+ "loss": 1.2356,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 0.41,
582
+ "learning_rate": 0.00019426286953145704,
583
+ "loss": 1.2815,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 0.41,
588
+ "learning_rate": 0.00019410831526407984,
589
+ "loss": 1.2286,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 0.42,
594
+ "learning_rate": 0.00019395176993109356,
595
+ "loss": 1.2427,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 0.42,
600
+ "learning_rate": 0.0001937932368445544,
601
+ "loss": 1.2076,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 0.43,
606
+ "learning_rate": 0.00019363271935857372,
607
+ "loss": 1.2544,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 0.43,
612
+ "learning_rate": 0.00019347022086924732,
613
+ "loss": 1.2632,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 0.43,
618
+ "learning_rate": 0.00019330574481458333,
619
+ "loss": 1.2088,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 0.44,
624
+ "learning_rate": 0.00019313929467442952,
625
+ "loss": 1.2348,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 0.44,
630
+ "learning_rate": 0.00019297087397039984,
631
+ "loss": 1.2186,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 0.45,
636
+ "learning_rate": 0.00019280048626579962,
637
+ "loss": 1.2039,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 0.45,
642
+ "learning_rate": 0.0001926281351655506,
643
+ "loss": 1.2665,
644
+ "step": 106
645
+ },
646
+ {
647
+ "epoch": 0.46,
648
+ "learning_rate": 0.0001924538243161142,
649
+ "loss": 1.2293,
650
+ "step": 107
651
+ },
652
+ {
653
+ "epoch": 0.46,
654
+ "learning_rate": 0.0001922775574054147,
655
+ "loss": 1.2712,
656
+ "step": 108
657
+ },
658
+ {
659
+ "epoch": 0.46,
660
+ "learning_rate": 0.00019209933816276102,
661
+ "loss": 1.2019,
662
+ "step": 109
663
+ },
664
+ {
665
+ "epoch": 0.47,
666
+ "learning_rate": 0.00019191917035876798,
667
+ "loss": 1.2248,
668
+ "step": 110
669
+ },
670
+ {
671
+ "epoch": 0.47,
672
+ "learning_rate": 0.00019173705780527642,
673
+ "loss": 1.2718,
674
+ "step": 111
675
+ },
676
+ {
677
+ "epoch": 0.48,
678
+ "learning_rate": 0.00019155300435527256,
679
+ "loss": 1.1894,
680
+ "step": 112
681
+ },
682
+ {
683
+ "epoch": 0.48,
684
+ "learning_rate": 0.00019136701390280644,
685
+ "loss": 1.2603,
686
+ "step": 113
687
+ },
688
+ {
689
+ "epoch": 0.49,
690
+ "learning_rate": 0.00019117909038290974,
691
+ "loss": 1.2292,
692
+ "step": 114
693
+ },
694
+ {
695
+ "epoch": 0.49,
696
+ "learning_rate": 0.00019098923777151222,
697
+ "loss": 1.2264,
698
+ "step": 115
699
+ },
700
+ {
701
+ "epoch": 0.49,
702
+ "learning_rate": 0.00019079746008535784,
703
+ "loss": 1.2553,
704
+ "step": 116
705
+ },
706
+ {
707
+ "epoch": 0.5,
708
+ "learning_rate": 0.0001906037613819197,
709
+ "loss": 1.287,
710
+ "step": 117
711
+ },
712
+ {
713
+ "epoch": 0.5,
714
+ "learning_rate": 0.00019040814575931413,
715
+ "loss": 1.1868,
716
+ "step": 118
717
+ },
718
+ {
719
+ "epoch": 0.51,
720
+ "learning_rate": 0.00019021061735621412,
721
+ "loss": 1.2465,
722
+ "step": 119
723
+ },
724
+ {
725
+ "epoch": 0.51,
726
+ "learning_rate": 0.00019001118035176162,
727
+ "loss": 1.2324,
728
+ "step": 120
729
+ },
730
+ {
731
+ "epoch": 0.51,
732
+ "learning_rate": 0.0001898098389654792,
733
+ "loss": 1.242,
734
+ "step": 121
735
+ },
736
+ {
737
+ "epoch": 0.52,
738
+ "learning_rate": 0.0001896065974571808,
739
+ "loss": 1.2202,
740
+ "step": 122
741
+ },
742
+ {
743
+ "epoch": 0.52,
744
+ "learning_rate": 0.00018940146012688146,
745
+ "loss": 1.2269,
746
+ "step": 123
747
+ },
748
+ {
749
+ "epoch": 0.53,
750
+ "learning_rate": 0.00018919443131470658,
751
+ "loss": 1.2456,
752
+ "step": 124
753
+ },
754
+ {
755
+ "epoch": 0.53,
756
+ "learning_rate": 0.00018898551540079989,
757
+ "loss": 1.1917,
758
+ "step": 125
759
+ },
760
+ {
761
+ "epoch": 0.54,
762
+ "learning_rate": 0.00018877471680523082,
763
+ "loss": 1.2336,
764
+ "step": 126
765
+ },
766
+ {
767
+ "epoch": 0.54,
768
+ "learning_rate": 0.00018856203998790112,
769
+ "loss": 1.2126,
770
+ "step": 127
771
+ },
772
+ {
773
+ "epoch": 0.54,
774
+ "learning_rate": 0.00018834748944845028,
775
+ "loss": 1.2487,
776
+ "step": 128
777
+ },
778
+ {
779
+ "epoch": 0.55,
780
+ "learning_rate": 0.00018813106972616055,
781
+ "loss": 1.2732,
782
+ "step": 129
783
+ },
784
+ {
785
+ "epoch": 0.55,
786
+ "learning_rate": 0.0001879127853998607,
787
+ "loss": 1.2385,
788
+ "step": 130
789
+ },
790
+ {
791
+ "epoch": 0.56,
792
+ "learning_rate": 0.00018769264108782933,
793
+ "loss": 1.2294,
794
+ "step": 131
795
+ },
796
+ {
797
+ "epoch": 0.56,
798
+ "learning_rate": 0.00018747064144769703,
799
+ "loss": 1.2875,
800
+ "step": 132
801
+ },
802
+ {
803
+ "epoch": 0.57,
804
+ "learning_rate": 0.0001872467911763479,
805
+ "loss": 1.266,
806
+ "step": 133
807
+ },
808
+ {
809
+ "epoch": 0.57,
810
+ "learning_rate": 0.0001870210950098201,
811
+ "loss": 1.1919,
812
+ "step": 134
813
+ },
814
+ {
815
+ "epoch": 0.57,
816
+ "learning_rate": 0.00018679355772320585,
817
+ "loss": 1.2646,
818
+ "step": 135
819
+ },
820
+ {
821
+ "epoch": 0.58,
822
+ "learning_rate": 0.00018656418413055007,
823
+ "loss": 1.2151,
824
+ "step": 136
825
+ },
826
+ {
827
+ "epoch": 0.58,
828
+ "learning_rate": 0.0001863329790847488,
829
+ "loss": 1.224,
830
+ "step": 137
831
+ },
832
+ {
833
+ "epoch": 0.59,
834
+ "learning_rate": 0.0001860999474774466,
835
+ "loss": 1.2197,
836
+ "step": 138
837
+ },
838
+ {
839
+ "epoch": 0.59,
840
+ "learning_rate": 0.00018586509423893267,
841
+ "loss": 1.2487,
842
+ "step": 139
843
+ },
844
+ {
845
+ "epoch": 0.6,
846
+ "learning_rate": 0.00018562842433803687,
847
+ "loss": 1.2164,
848
+ "step": 140
849
+ },
850
+ {
851
+ "epoch": 0.6,
852
+ "learning_rate": 0.00018538994278202448,
853
+ "loss": 1.2395,
854
+ "step": 141
855
+ },
856
+ {
857
+ "epoch": 0.6,
858
+ "learning_rate": 0.0001851496546164903,
859
+ "loss": 1.2454,
860
+ "step": 142
861
+ },
862
+ {
863
+ "epoch": 0.61,
864
+ "learning_rate": 0.00018490756492525187,
865
+ "loss": 1.2096,
866
+ "step": 143
867
+ },
868
+ {
869
+ "epoch": 0.61,
870
+ "learning_rate": 0.00018466367883024186,
871
+ "loss": 1.1975,
872
+ "step": 144
873
+ },
874
+ {
875
+ "epoch": 0.62,
876
+ "learning_rate": 0.00018441800149139988,
877
+ "loss": 1.19,
878
+ "step": 145
879
+ },
880
+ {
881
+ "epoch": 0.62,
882
+ "learning_rate": 0.00018417053810656302,
883
+ "loss": 1.2228,
884
+ "step": 146
885
+ },
886
+ {
887
+ "epoch": 0.63,
888
+ "learning_rate": 0.0001839212939113562,
889
+ "loss": 1.2044,
890
+ "step": 147
891
+ },
892
+ {
893
+ "epoch": 0.63,
894
+ "learning_rate": 0.00018367027417908117,
895
+ "loss": 1.2381,
896
+ "step": 148
897
+ },
898
+ {
899
+ "epoch": 0.63,
900
+ "learning_rate": 0.00018341748422060503,
901
+ "loss": 1.225,
902
+ "step": 149
903
+ },
904
+ {
905
+ "epoch": 0.64,
906
+ "learning_rate": 0.00018316292938424787,
907
+ "loss": 1.2104,
908
+ "step": 150
909
+ },
910
+ {
911
+ "epoch": 0.64,
912
+ "learning_rate": 0.00018290661505566963,
913
+ "loss": 1.229,
914
+ "step": 151
915
+ },
916
+ {
917
+ "epoch": 0.65,
918
+ "learning_rate": 0.00018264854665775605,
919
+ "loss": 1.2396,
920
+ "step": 152
921
+ },
922
+ {
923
+ "epoch": 0.65,
924
+ "learning_rate": 0.0001823887296505041,
925
+ "loss": 1.2132,
926
+ "step": 153
927
+ },
928
+ {
929
+ "epoch": 0.66,
930
+ "learning_rate": 0.00018212716953090624,
931
+ "loss": 1.2255,
932
+ "step": 154
933
+ },
934
+ {
935
+ "epoch": 0.66,
936
+ "learning_rate": 0.00018186387183283443,
937
+ "loss": 1.2393,
938
+ "step": 155
939
+ },
940
+ {
941
+ "epoch": 0.66,
942
+ "learning_rate": 0.00018159884212692274,
943
+ "loss": 1.273,
944
+ "step": 156
945
+ },
946
+ {
947
+ "epoch": 0.67,
948
+ "learning_rate": 0.00018133208602044972,
949
+ "loss": 1.2314,
950
+ "step": 157
951
+ },
952
+ {
953
+ "epoch": 0.67,
954
+ "learning_rate": 0.00018106360915721956,
955
+ "loss": 1.1983,
956
+ "step": 158
957
+ },
958
+ {
959
+ "epoch": 0.68,
960
+ "learning_rate": 0.0001807934172174429,
961
+ "loss": 1.1759,
962
+ "step": 159
963
+ },
964
+ {
965
+ "epoch": 0.68,
966
+ "learning_rate": 0.00018052151591761644,
967
+ "loss": 1.2081,
968
+ "step": 160
969
+ },
970
+ {
971
+ "epoch": 0.69,
972
+ "learning_rate": 0.0001802479110104022,
973
+ "loss": 1.1594,
974
+ "step": 161
975
+ },
976
+ {
977
+ "epoch": 0.69,
978
+ "learning_rate": 0.0001799726082845057,
979
+ "loss": 1.2388,
980
+ "step": 162
981
+ },
982
+ {
983
+ "epoch": 0.69,
984
+ "learning_rate": 0.00017969561356455336,
985
+ "loss": 1.2052,
986
+ "step": 163
987
+ },
988
+ {
989
+ "epoch": 0.7,
990
+ "learning_rate": 0.00017941693271096966,
991
+ "loss": 1.2385,
992
+ "step": 164
993
+ },
994
+ {
995
+ "epoch": 0.7,
996
+ "learning_rate": 0.00017913657161985268,
997
+ "loss": 1.225,
998
+ "step": 165
999
+ },
1000
+ {
1001
+ "epoch": 0.71,
1002
+ "learning_rate": 0.0001788545362228496,
1003
+ "loss": 1.2173,
1004
+ "step": 166
1005
+ },
1006
+ {
1007
+ "epoch": 0.71,
1008
+ "learning_rate": 0.00017857083248703126,
1009
+ "loss": 1.1866,
1010
+ "step": 167
1011
+ },
1012
+ {
1013
+ "epoch": 0.71,
1014
+ "learning_rate": 0.00017828546641476578,
1015
+ "loss": 1.2196,
1016
+ "step": 168
1017
+ },
1018
+ {
1019
+ "epoch": 0.72,
1020
+ "learning_rate": 0.0001779984440435916,
1021
+ "loss": 1.2535,
1022
+ "step": 169
1023
+ },
1024
+ {
1025
+ "epoch": 0.72,
1026
+ "learning_rate": 0.00017770977144608978,
1027
+ "loss": 1.2124,
1028
+ "step": 170
1029
+ },
1030
+ {
1031
+ "epoch": 0.73,
1032
+ "learning_rate": 0.0001774194547297555,
1033
+ "loss": 1.2213,
1034
+ "step": 171
1035
+ },
1036
+ {
1037
+ "epoch": 0.73,
1038
+ "learning_rate": 0.00017712750003686883,
1039
+ "loss": 1.165,
1040
+ "step": 172
1041
+ },
1042
+ {
1043
+ "epoch": 0.74,
1044
+ "learning_rate": 0.0001768339135443648,
1045
+ "loss": 1.2336,
1046
+ "step": 173
1047
+ },
1048
+ {
1049
+ "epoch": 0.74,
1050
+ "learning_rate": 0.00017653870146370267,
1051
+ "loss": 1.2574,
1052
+ "step": 174
1053
+ },
1054
+ {
1055
+ "epoch": 0.74,
1056
+ "learning_rate": 0.00017624187004073463,
1057
+ "loss": 1.201,
1058
+ "step": 175
1059
+ },
1060
+ {
1061
+ "epoch": 0.75,
1062
+ "learning_rate": 0.0001759434255555734,
1063
+ "loss": 1.2127,
1064
+ "step": 176
1065
+ },
1066
+ {
1067
+ "epoch": 0.75,
1068
+ "learning_rate": 0.00017564337432245976,
1069
+ "loss": 1.1924,
1070
+ "step": 177
1071
+ },
1072
+ {
1073
+ "epoch": 0.76,
1074
+ "learning_rate": 0.00017534172268962852,
1075
+ "loss": 1.293,
1076
+ "step": 178
1077
+ },
1078
+ {
1079
+ "epoch": 0.76,
1080
+ "learning_rate": 0.00017503847703917455,
1081
+ "loss": 1.3086,
1082
+ "step": 179
1083
+ },
1084
+ {
1085
+ "epoch": 0.77,
1086
+ "learning_rate": 0.0001747336437869176,
1087
+ "loss": 1.2584,
1088
+ "step": 180
1089
+ },
1090
+ {
1091
+ "epoch": 0.77,
1092
+ "learning_rate": 0.00017442722938226647,
1093
+ "loss": 1.242,
1094
+ "step": 181
1095
+ },
1096
+ {
1097
+ "epoch": 0.77,
1098
+ "learning_rate": 0.00017411924030808284,
1099
+ "loss": 1.1669,
1100
+ "step": 182
1101
+ },
1102
+ {
1103
+ "epoch": 0.78,
1104
+ "learning_rate": 0.00017380968308054385,
1105
+ "loss": 1.2288,
1106
+ "step": 183
1107
+ },
1108
+ {
1109
+ "epoch": 0.78,
1110
+ "learning_rate": 0.0001734985642490043,
1111
+ "loss": 1.2583,
1112
+ "step": 184
1113
+ },
1114
+ {
1115
+ "epoch": 0.79,
1116
+ "learning_rate": 0.00017318589039585816,
1117
+ "loss": 1.2701,
1118
+ "step": 185
1119
+ },
1120
+ {
1121
+ "epoch": 0.79,
1122
+ "learning_rate": 0.0001728716681363993,
1123
+ "loss": 1.2294,
1124
+ "step": 186
1125
+ },
1126
+ {
1127
+ "epoch": 0.8,
1128
+ "learning_rate": 0.00017255590411868136,
1129
+ "loss": 1.1843,
1130
+ "step": 187
1131
+ },
1132
+ {
1133
+ "epoch": 0.8,
1134
+ "learning_rate": 0.00017223860502337733,
1135
+ "loss": 1.2367,
1136
+ "step": 188
1137
+ },
1138
+ {
1139
+ "epoch": 0.8,
1140
+ "learning_rate": 0.00017191977756363808,
1141
+ "loss": 1.2128,
1142
+ "step": 189
1143
+ },
1144
+ {
1145
+ "epoch": 0.81,
1146
+ "learning_rate": 0.00017159942848495025,
1147
+ "loss": 1.1754,
1148
+ "step": 190
1149
+ },
1150
+ {
1151
+ "epoch": 0.81,
1152
+ "learning_rate": 0.00017127756456499372,
1153
+ "loss": 1.1917,
1154
+ "step": 191
1155
+ },
1156
+ {
1157
+ "epoch": 0.82,
1158
+ "learning_rate": 0.0001709541926134982,
1159
+ "loss": 1.2122,
1160
+ "step": 192
1161
+ },
1162
+ {
1163
+ "epoch": 0.82,
1164
+ "learning_rate": 0.0001706293194720989,
1165
+ "loss": 1.2336,
1166
+ "step": 193
1167
+ },
1168
+ {
1169
+ "epoch": 0.83,
1170
+ "learning_rate": 0.00017030295201419206,
1171
+ "loss": 1.2285,
1172
+ "step": 194
1173
+ },
1174
+ {
1175
+ "epoch": 0.83,
1176
+ "learning_rate": 0.00016997509714478944,
1177
+ "loss": 1.2509,
1178
+ "step": 195
1179
+ },
1180
+ {
1181
+ "epoch": 0.83,
1182
+ "learning_rate": 0.00016964576180037217,
1183
+ "loss": 1.217,
1184
+ "step": 196
1185
+ },
1186
+ {
1187
+ "epoch": 0.84,
1188
+ "learning_rate": 0.00016931495294874408,
1189
+ "loss": 1.2189,
1190
+ "step": 197
1191
+ },
1192
+ {
1193
+ "epoch": 0.84,
1194
+ "learning_rate": 0.00016898267758888423,
1195
+ "loss": 1.1784,
1196
+ "step": 198
1197
+ },
1198
+ {
1199
+ "epoch": 0.85,
1200
+ "learning_rate": 0.00016864894275079882,
1201
+ "loss": 1.2197,
1202
+ "step": 199
1203
+ },
1204
+ {
1205
+ "epoch": 0.85,
1206
+ "learning_rate": 0.00016831375549537252,
1207
+ "loss": 1.2552,
1208
+ "step": 200
1209
+ },
1210
+ {
1211
+ "epoch": 0.86,
1212
+ "learning_rate": 0.00016797712291421904,
1213
+ "loss": 1.1743,
1214
+ "step": 201
1215
+ },
1216
+ {
1217
+ "epoch": 0.86,
1218
+ "learning_rate": 0.00016763905212953102,
1219
+ "loss": 1.2396,
1220
+ "step": 202
1221
+ },
1222
+ {
1223
+ "epoch": 0.86,
1224
+ "learning_rate": 0.0001672995502939295,
1225
+ "loss": 1.2283,
1226
+ "step": 203
1227
+ },
1228
+ {
1229
+ "epoch": 0.87,
1230
+ "learning_rate": 0.00016695862459031248,
1231
+ "loss": 1.1892,
1232
+ "step": 204
1233
+ },
1234
+ {
1235
+ "epoch": 0.87,
1236
+ "learning_rate": 0.00016661628223170295,
1237
+ "loss": 1.1739,
1238
+ "step": 205
1239
+ },
1240
+ {
1241
+ "epoch": 0.88,
1242
+ "learning_rate": 0.00016627253046109638,
1243
+ "loss": 1.1211,
1244
+ "step": 206
1245
+ },
1246
+ {
1247
+ "epoch": 0.88,
1248
+ "learning_rate": 0.0001659273765513073,
1249
+ "loss": 1.1466,
1250
+ "step": 207
1251
+ },
1252
+ {
1253
+ "epoch": 0.89,
1254
+ "learning_rate": 0.00016558082780481563,
1255
+ "loss": 1.1761,
1256
+ "step": 208
1257
+ },
1258
+ {
1259
+ "epoch": 0.89,
1260
+ "learning_rate": 0.00016523289155361204,
1261
+ "loss": 1.2146,
1262
+ "step": 209
1263
+ },
1264
+ {
1265
+ "epoch": 0.89,
1266
+ "learning_rate": 0.00016488357515904295,
1267
+ "loss": 1.2032,
1268
+ "step": 210
1269
+ },
1270
+ {
1271
+ "epoch": 0.9,
1272
+ "learning_rate": 0.0001645328860116546,
1273
+ "loss": 1.2129,
1274
+ "step": 211
1275
+ },
1276
+ {
1277
+ "epoch": 0.9,
1278
+ "learning_rate": 0.00016418083153103683,
1279
+ "loss": 1.2211,
1280
+ "step": 212
1281
+ },
1282
+ {
1283
+ "epoch": 0.91,
1284
+ "learning_rate": 0.0001638274191656661,
1285
+ "loss": 1.2022,
1286
+ "step": 213
1287
+ },
1288
+ {
1289
+ "epoch": 0.91,
1290
+ "learning_rate": 0.00016347265639274778,
1291
+ "loss": 1.2057,
1292
+ "step": 214
1293
+ },
1294
+ {
1295
+ "epoch": 0.91,
1296
+ "learning_rate": 0.00016311655071805822,
1297
+ "loss": 1.229,
1298
+ "step": 215
1299
+ },
1300
+ {
1301
+ "epoch": 0.92,
1302
+ "learning_rate": 0.00016275910967578558,
1303
+ "loss": 1.1658,
1304
+ "step": 216
1305
+ },
1306
+ {
1307
+ "epoch": 0.92,
1308
+ "learning_rate": 0.00016240034082837078,
1309
+ "loss": 1.2102,
1310
+ "step": 217
1311
+ },
1312
+ {
1313
+ "epoch": 0.93,
1314
+ "learning_rate": 0.00016204025176634712,
1315
+ "loss": 1.2057,
1316
+ "step": 218
1317
+ },
1318
+ {
1319
+ "epoch": 0.93,
1320
+ "learning_rate": 0.00016167885010818017,
1321
+ "loss": 1.222,
1322
+ "step": 219
1323
+ },
1324
+ {
1325
+ "epoch": 0.94,
1326
+ "learning_rate": 0.00016131614350010614,
1327
+ "loss": 1.1824,
1328
+ "step": 220
1329
+ },
1330
+ {
1331
+ "epoch": 0.94,
1332
+ "learning_rate": 0.00016095213961597033,
1333
+ "loss": 1.2212,
1334
+ "step": 221
1335
+ },
1336
+ {
1337
+ "epoch": 0.94,
1338
+ "learning_rate": 0.00016058684615706477,
1339
+ "loss": 1.2101,
1340
+ "step": 222
1341
+ },
1342
+ {
1343
+ "epoch": 0.95,
1344
+ "learning_rate": 0.00016022027085196516,
1345
+ "loss": 1.153,
1346
+ "step": 223
1347
+ },
1348
+ {
1349
+ "epoch": 0.95,
1350
+ "learning_rate": 0.0001598524214563675,
1351
+ "loss": 1.1471,
1352
+ "step": 224
1353
+ },
1354
+ {
1355
+ "epoch": 0.96,
1356
+ "learning_rate": 0.00015948330575292401,
1357
+ "loss": 1.1859,
1358
+ "step": 225
1359
+ },
1360
+ {
1361
+ "epoch": 0.96,
1362
+ "learning_rate": 0.0001591129315510782,
1363
+ "loss": 1.1612,
1364
+ "step": 226
1365
+ },
1366
+ {
1367
+ "epoch": 0.97,
1368
+ "learning_rate": 0.00015874130668690003,
1369
+ "loss": 1.1691,
1370
+ "step": 227
1371
+ },
1372
+ {
1373
+ "epoch": 0.97,
1374
+ "learning_rate": 0.00015836843902291984,
1375
+ "loss": 1.1859,
1376
+ "step": 228
1377
+ },
1378
+ {
1379
+ "epoch": 0.97,
1380
+ "learning_rate": 0.00015799433644796216,
1381
+ "loss": 1.1831,
1382
+ "step": 229
1383
+ },
1384
+ {
1385
+ "epoch": 0.98,
1386
+ "learning_rate": 0.00015761900687697865,
1387
+ "loss": 1.226,
1388
+ "step": 230
1389
+ },
1390
+ {
1391
+ "epoch": 0.98,
1392
+ "learning_rate": 0.00015724245825088086,
1393
+ "loss": 1.233,
1394
+ "step": 231
1395
+ },
1396
+ {
1397
+ "epoch": 0.99,
1398
+ "learning_rate": 0.00015686469853637192,
1399
+ "loss": 1.2146,
1400
+ "step": 232
1401
+ },
1402
+ {
1403
+ "epoch": 0.99,
1404
+ "learning_rate": 0.00015648573572577839,
1405
+ "loss": 1.1979,
1406
+ "step": 233
1407
+ },
1408
+ {
1409
+ "epoch": 1.0,
1410
+ "learning_rate": 0.0001561055778368807,
1411
+ "loss": 1.1544,
1412
+ "step": 234
1413
+ },
1414
+ {
1415
+ "epoch": 1.0,
1416
+ "learning_rate": 0.00015572423291274393,
1417
+ "loss": 1.1036,
1418
+ "step": 235
1419
+ },
1420
+ {
1421
+ "epoch": 1.0,
1422
+ "learning_rate": 0.00015534170902154742,
1423
+ "loss": 0.9431,
1424
+ "step": 236
1425
+ },
1426
+ {
1427
+ "epoch": 1.01,
1428
+ "learning_rate": 0.00015495801425641407,
1429
+ "loss": 0.9798,
1430
+ "step": 237
1431
+ },
1432
+ {
1433
+ "epoch": 1.01,
1434
+ "learning_rate": 0.0001545731567352392,
1435
+ "loss": 0.9381,
1436
+ "step": 238
1437
+ },
1438
+ {
1439
+ "epoch": 1.02,
1440
+ "learning_rate": 0.00015418714460051875,
1441
+ "loss": 0.9564,
1442
+ "step": 239
1443
+ },
1444
+ {
1445
+ "epoch": 1.02,
1446
+ "learning_rate": 0.00015379998601917704,
1447
+ "loss": 0.9039,
1448
+ "step": 240
1449
+ },
1450
+ {
1451
+ "epoch": 1.03,
1452
+ "learning_rate": 0.0001534116891823939,
1453
+ "loss": 0.8984,
1454
+ "step": 241
1455
+ },
1456
+ {
1457
+ "epoch": 1.03,
1458
+ "learning_rate": 0.00015302226230543147,
1459
+ "loss": 0.952,
1460
+ "step": 242
1461
+ },
1462
+ {
1463
+ "epoch": 1.03,
1464
+ "learning_rate": 0.00015263171362746026,
1465
+ "loss": 0.94,
1466
+ "step": 243
1467
+ },
1468
+ {
1469
+ "epoch": 1.04,
1470
+ "learning_rate": 0.0001522400514113851,
1471
+ "loss": 0.9551,
1472
+ "step": 244
1473
+ },
1474
+ {
1475
+ "epoch": 1.04,
1476
+ "learning_rate": 0.00015184728394366988,
1477
+ "loss": 1.0139,
1478
+ "step": 245
1479
+ },
1480
+ {
1481
+ "epoch": 1.05,
1482
+ "learning_rate": 0.00015145341953416271,
1483
+ "loss": 0.9337,
1484
+ "step": 246
1485
+ },
1486
+ {
1487
+ "epoch": 1.05,
1488
+ "learning_rate": 0.0001510584665159198,
1489
+ "loss": 0.9504,
1490
+ "step": 247
1491
+ },
1492
+ {
1493
+ "epoch": 1.06,
1494
+ "learning_rate": 0.00015066243324502918,
1495
+ "loss": 0.9361,
1496
+ "step": 248
1497
+ },
1498
+ {
1499
+ "epoch": 1.06,
1500
+ "learning_rate": 0.00015026532810043407,
1501
+ "loss": 0.9395,
1502
+ "step": 249
1503
+ },
1504
+ {
1505
+ "epoch": 1.06,
1506
+ "learning_rate": 0.00014986715948375542,
1507
+ "loss": 0.9442,
1508
+ "step": 250
1509
+ },
1510
+ {
1511
+ "epoch": 1.07,
1512
+ "learning_rate": 0.00014946793581911428,
1513
+ "loss": 0.9664,
1514
+ "step": 251
1515
+ },
1516
+ {
1517
+ "epoch": 1.07,
1518
+ "learning_rate": 0.00014906766555295358,
1519
+ "loss": 0.8896,
1520
+ "step": 252
1521
+ },
1522
+ {
1523
+ "epoch": 1.08,
1524
+ "learning_rate": 0.00014866635715385927,
1525
+ "loss": 0.9328,
1526
+ "step": 253
1527
+ },
1528
+ {
1529
+ "epoch": 1.08,
1530
+ "learning_rate": 0.0001482640191123813,
1531
+ "loss": 0.8727,
1532
+ "step": 254
1533
+ },
1534
+ {
1535
+ "epoch": 1.09,
1536
+ "learning_rate": 0.00014786065994085396,
1537
+ "loss": 0.8948,
1538
+ "step": 255
1539
+ },
1540
+ {
1541
+ "epoch": 1.09,
1542
+ "learning_rate": 0.00014745628817321578,
1543
+ "loss": 0.9564,
1544
+ "step": 256
1545
+ },
1546
+ {
1547
+ "epoch": 1.09,
1548
+ "learning_rate": 0.00014705091236482887,
1549
+ "loss": 0.9485,
1550
+ "step": 257
1551
+ },
1552
+ {
1553
+ "epoch": 1.1,
1554
+ "learning_rate": 0.00014664454109229808,
1555
+ "loss": 0.931,
1556
+ "step": 258
1557
+ },
1558
+ {
1559
+ "epoch": 1.1,
1560
+ "learning_rate": 0.00014623718295328944,
1561
+ "loss": 0.9614,
1562
+ "step": 259
1563
+ },
1564
+ {
1565
+ "epoch": 1.11,
1566
+ "learning_rate": 0.00014582884656634827,
1567
+ "loss": 0.9347,
1568
+ "step": 260
1569
+ },
1570
+ {
1571
+ "epoch": 1.11,
1572
+ "learning_rate": 0.00014541954057071692,
1573
+ "loss": 0.8861,
1574
+ "step": 261
1575
+ },
1576
+ {
1577
+ "epoch": 1.11,
1578
+ "learning_rate": 0.00014500927362615177,
1579
+ "loss": 0.9045,
1580
+ "step": 262
1581
+ },
1582
+ {
1583
+ "epoch": 1.12,
1584
+ "learning_rate": 0.00014459805441274028,
1585
+ "loss": 0.9271,
1586
+ "step": 263
1587
+ },
1588
+ {
1589
+ "epoch": 1.12,
1590
+ "learning_rate": 0.00014418589163071722,
1591
+ "loss": 0.8829,
1592
+ "step": 264
1593
+ },
1594
+ {
1595
+ "epoch": 1.13,
1596
+ "learning_rate": 0.00014377279400028053,
1597
+ "loss": 0.9252,
1598
+ "step": 265
1599
+ },
1600
+ {
1601
+ "epoch": 1.13,
1602
+ "learning_rate": 0.00014335877026140688,
1603
+ "loss": 0.9299,
1604
+ "step": 266
1605
+ },
1606
+ {
1607
+ "epoch": 1.14,
1608
+ "learning_rate": 0.000142943829173667,
1609
+ "loss": 0.8597,
1610
+ "step": 267
1611
+ },
1612
+ {
1613
+ "epoch": 1.14,
1614
+ "learning_rate": 0.00014252797951603977,
1615
+ "loss": 0.8919,
1616
+ "step": 268
1617
+ },
1618
+ {
1619
+ "epoch": 1.14,
1620
+ "learning_rate": 0.00014211123008672712,
1621
+ "loss": 0.9356,
1622
+ "step": 269
1623
+ },
1624
+ {
1625
+ "epoch": 1.15,
1626
+ "learning_rate": 0.0001416935897029675,
1627
+ "loss": 0.8959,
1628
+ "step": 270
1629
+ },
1630
+ {
1631
+ "epoch": 1.15,
1632
+ "learning_rate": 0.0001412750672008494,
1633
+ "loss": 0.8936,
1634
+ "step": 271
1635
+ },
1636
+ {
1637
+ "epoch": 1.16,
1638
+ "learning_rate": 0.00014085567143512457,
1639
+ "loss": 0.858,
1640
+ "step": 272
1641
+ },
1642
+ {
1643
+ "epoch": 1.16,
1644
+ "learning_rate": 0.00014043541127902037,
1645
+ "loss": 0.8618,
1646
+ "step": 273
1647
+ },
1648
+ {
1649
+ "epoch": 1.17,
1650
+ "learning_rate": 0.00014001429562405225,
1651
+ "loss": 0.89,
1652
+ "step": 274
1653
+ },
1654
+ {
1655
+ "epoch": 1.17,
1656
+ "learning_rate": 0.00013959233337983582,
1657
+ "loss": 0.9272,
1658
+ "step": 275
1659
+ },
1660
+ {
1661
+ "epoch": 1.17,
1662
+ "learning_rate": 0.00013916953347389776,
1663
+ "loss": 0.8889,
1664
+ "step": 276
1665
+ },
1666
+ {
1667
+ "epoch": 1.18,
1668
+ "learning_rate": 0.0001387459048514876,
1669
+ "loss": 0.8864,
1670
+ "step": 277
1671
+ },
1672
+ {
1673
+ "epoch": 1.18,
1674
+ "learning_rate": 0.00013832145647538799,
1675
+ "loss": 0.9299,
1676
+ "step": 278
1677
+ },
1678
+ {
1679
+ "epoch": 1.19,
1680
+ "learning_rate": 0.00013789619732572538,
1681
+ "loss": 0.9379,
1682
+ "step": 279
1683
+ },
1684
+ {
1685
+ "epoch": 1.19,
1686
+ "learning_rate": 0.00013747013639977973,
1687
+ "loss": 0.8988,
1688
+ "step": 280
1689
+ },
1690
+ {
1691
+ "epoch": 1.2,
1692
+ "learning_rate": 0.0001370432827117945,
1693
+ "loss": 0.8828,
1694
+ "step": 281
1695
+ },
1696
+ {
1697
+ "epoch": 1.2,
1698
+ "learning_rate": 0.0001366156452927856,
1699
+ "loss": 0.8929,
1700
+ "step": 282
1701
+ },
1702
+ {
1703
+ "epoch": 1.2,
1704
+ "learning_rate": 0.00013618723319035056,
1705
+ "loss": 0.903,
1706
+ "step": 283
1707
+ },
1708
+ {
1709
+ "epoch": 1.21,
1710
+ "learning_rate": 0.000135758055468477,
1711
+ "loss": 0.946,
1712
+ "step": 284
1713
+ },
1714
+ {
1715
+ "epoch": 1.21,
1716
+ "learning_rate": 0.00013532812120735087,
1717
+ "loss": 0.8985,
1718
+ "step": 285
1719
+ },
1720
+ {
1721
+ "epoch": 1.22,
1722
+ "learning_rate": 0.0001348974395031643,
1723
+ "loss": 0.9617,
1724
+ "step": 286
1725
+ },
1726
+ {
1727
+ "epoch": 1.22,
1728
+ "learning_rate": 0.00013446601946792334,
1729
+ "loss": 0.8937,
1730
+ "step": 287
1731
+ },
1732
+ {
1733
+ "epoch": 1.23,
1734
+ "learning_rate": 0.00013403387022925488,
1735
+ "loss": 0.9223,
1736
+ "step": 288
1737
+ },
1738
+ {
1739
+ "epoch": 1.23,
1740
+ "learning_rate": 0.00013360100093021376,
1741
+ "loss": 0.8876,
1742
+ "step": 289
1743
+ },
1744
+ {
1745
+ "epoch": 1.23,
1746
+ "learning_rate": 0.00013316742072908927,
1747
+ "loss": 0.8735,
1748
+ "step": 290
1749
+ },
1750
+ {
1751
+ "epoch": 1.24,
1752
+ "learning_rate": 0.0001327331387992114,
1753
+ "loss": 0.8766,
1754
+ "step": 291
1755
+ },
1756
+ {
1757
+ "epoch": 1.24,
1758
+ "learning_rate": 0.00013229816432875664,
1759
+ "loss": 0.8885,
1760
+ "step": 292
1761
+ },
1762
+ {
1763
+ "epoch": 1.25,
1764
+ "learning_rate": 0.00013186250652055378,
1765
+ "loss": 0.8816,
1766
+ "step": 293
1767
+ },
1768
+ {
1769
+ "epoch": 1.25,
1770
+ "learning_rate": 0.00013142617459188899,
1771
+ "loss": 0.8838,
1772
+ "step": 294
1773
+ },
1774
+ {
1775
+ "epoch": 1.26,
1776
+ "learning_rate": 0.0001309891777743111,
1777
+ "loss": 0.8981,
1778
+ "step": 295
1779
+ },
1780
+ {
1781
+ "epoch": 1.26,
1782
+ "learning_rate": 0.00013055152531343592,
1783
+ "loss": 0.8672,
1784
+ "step": 296
1785
+ },
1786
+ {
1787
+ "epoch": 1.26,
1788
+ "learning_rate": 0.00013011322646875088,
1789
+ "loss": 0.9303,
1790
+ "step": 297
1791
+ },
1792
+ {
1793
+ "epoch": 1.27,
1794
+ "learning_rate": 0.00012967429051341913,
1795
+ "loss": 0.8727,
1796
+ "step": 298
1797
+ },
1798
+ {
1799
+ "epoch": 1.27,
1800
+ "learning_rate": 0.00012923472673408322,
1801
+ "loss": 0.896,
1802
+ "step": 299
1803
+ },
1804
+ {
1805
+ "epoch": 1.28,
1806
+ "learning_rate": 0.0001287945444306686,
1807
+ "loss": 0.9047,
1808
+ "step": 300
1809
+ },
1810
+ {
1811
+ "epoch": 1.28,
1812
+ "learning_rate": 0.00012835375291618716,
1813
+ "loss": 0.8641,
1814
+ "step": 301
1815
+ },
1816
+ {
1817
+ "epoch": 1.29,
1818
+ "learning_rate": 0.00012791236151653973,
1819
+ "loss": 0.8577,
1820
+ "step": 302
1821
+ },
1822
+ {
1823
+ "epoch": 1.29,
1824
+ "learning_rate": 0.00012747037957031916,
1825
+ "loss": 0.8815,
1826
+ "step": 303
1827
+ },
1828
+ {
1829
+ "epoch": 1.29,
1830
+ "learning_rate": 0.00012702781642861253,
1831
+ "loss": 0.9196,
1832
+ "step": 304
1833
+ },
1834
+ {
1835
+ "epoch": 1.3,
1836
+ "learning_rate": 0.00012658468145480337,
1837
+ "loss": 0.8782,
1838
+ "step": 305
1839
+ },
1840
+ {
1841
+ "epoch": 1.3,
1842
+ "learning_rate": 0.00012614098402437366,
1843
+ "loss": 0.953,
1844
+ "step": 306
1845
+ },
1846
+ {
1847
+ "epoch": 1.31,
1848
+ "learning_rate": 0.00012569673352470523,
1849
+ "loss": 0.9381,
1850
+ "step": 307
1851
+ },
1852
+ {
1853
+ "epoch": 1.31,
1854
+ "learning_rate": 0.00012525193935488137,
1855
+ "loss": 0.8818,
1856
+ "step": 308
1857
+ },
1858
+ {
1859
+ "epoch": 1.31,
1860
+ "learning_rate": 0.00012480661092548786,
1861
+ "loss": 0.9057,
1862
+ "step": 309
1863
+ },
1864
+ {
1865
+ "epoch": 1.32,
1866
+ "learning_rate": 0.00012436075765841396,
1867
+ "loss": 0.8975,
1868
+ "step": 310
1869
+ },
1870
+ {
1871
+ "epoch": 1.32,
1872
+ "learning_rate": 0.00012391438898665287,
1873
+ "loss": 0.918,
1874
+ "step": 311
1875
+ },
1876
+ {
1877
+ "epoch": 1.33,
1878
+ "learning_rate": 0.00012346751435410248,
1879
+ "loss": 0.9211,
1880
+ "step": 312
1881
+ },
1882
+ {
1883
+ "epoch": 1.33,
1884
+ "learning_rate": 0.0001230201432153653,
1885
+ "loss": 0.9608,
1886
+ "step": 313
1887
+ },
1888
+ {
1889
+ "epoch": 1.34,
1890
+ "learning_rate": 0.00012257228503554835,
1891
+ "loss": 0.8869,
1892
+ "step": 314
1893
+ },
1894
+ {
1895
+ "epoch": 1.34,
1896
+ "learning_rate": 0.00012212394929006336,
1897
+ "loss": 0.897,
1898
+ "step": 315
1899
+ },
1900
+ {
1901
+ "epoch": 1.34,
1902
+ "learning_rate": 0.00012167514546442576,
1903
+ "loss": 0.8858,
1904
+ "step": 316
1905
+ },
1906
+ {
1907
+ "epoch": 1.35,
1908
+ "learning_rate": 0.00012122588305405434,
1909
+ "loss": 0.8305,
1910
+ "step": 317
1911
+ },
1912
+ {
1913
+ "epoch": 1.35,
1914
+ "learning_rate": 0.0001207761715640702,
1915
+ "loss": 0.9205,
1916
+ "step": 318
1917
+ },
1918
+ {
1919
+ "epoch": 1.36,
1920
+ "learning_rate": 0.00012032602050909574,
1921
+ "loss": 0.922,
1922
+ "step": 319
1923
+ },
1924
+ {
1925
+ "epoch": 1.36,
1926
+ "learning_rate": 0.00011987543941305321,
1927
+ "loss": 0.9019,
1928
+ "step": 320
1929
+ },
1930
+ {
1931
+ "epoch": 1.37,
1932
+ "learning_rate": 0.00011942443780896351,
1933
+ "loss": 0.9028,
1934
+ "step": 321
1935
+ },
1936
+ {
1937
+ "epoch": 1.37,
1938
+ "learning_rate": 0.00011897302523874405,
1939
+ "loss": 0.9307,
1940
+ "step": 322
1941
+ },
1942
+ {
1943
+ "epoch": 1.37,
1944
+ "learning_rate": 0.0001185212112530073,
1945
+ "loss": 0.9475,
1946
+ "step": 323
1947
+ },
1948
+ {
1949
+ "epoch": 1.38,
1950
+ "learning_rate": 0.0001180690054108585,
1951
+ "loss": 0.9168,
1952
+ "step": 324
1953
+ },
1954
+ {
1955
+ "epoch": 1.38,
1956
+ "learning_rate": 0.00011761641727969343,
1957
+ "loss": 0.8799,
1958
+ "step": 325
1959
+ },
1960
+ {
1961
+ "epoch": 1.39,
1962
+ "learning_rate": 0.00011716345643499608,
1963
+ "loss": 0.9363,
1964
+ "step": 326
1965
+ },
1966
+ {
1967
+ "epoch": 1.39,
1968
+ "learning_rate": 0.00011671013246013596,
1969
+ "loss": 0.852,
1970
+ "step": 327
1971
+ },
1972
+ {
1973
+ "epoch": 1.4,
1974
+ "learning_rate": 0.00011625645494616535,
1975
+ "loss": 0.9343,
1976
+ "step": 328
1977
+ },
1978
+ {
1979
+ "epoch": 1.4,
1980
+ "learning_rate": 0.0001158024334916165,
1981
+ "loss": 0.916,
1982
+ "step": 329
1983
+ },
1984
+ {
1985
+ "epoch": 1.4,
1986
+ "learning_rate": 0.00011534807770229845,
1987
+ "loss": 0.9824,
1988
+ "step": 330
1989
+ },
1990
+ {
1991
+ "epoch": 1.41,
1992
+ "learning_rate": 0.00011489339719109378,
1993
+ "loss": 0.9317,
1994
+ "step": 331
1995
+ },
1996
+ {
1997
+ "epoch": 1.41,
1998
+ "learning_rate": 0.00011443840157775527,
1999
+ "loss": 0.9233,
2000
+ "step": 332
2001
+ },
2002
+ {
2003
+ "epoch": 1.42,
2004
+ "learning_rate": 0.00011398310048870247,
2005
+ "loss": 0.9104,
2006
+ "step": 333
2007
+ },
2008
+ {
2009
+ "epoch": 1.42,
2010
+ "learning_rate": 0.00011352750355681772,
2011
+ "loss": 0.9006,
2012
+ "step": 334
2013
+ },
2014
+ {
2015
+ "epoch": 1.43,
2016
+ "learning_rate": 0.00011307162042124277,
2017
+ "loss": 0.926,
2018
+ "step": 335
2019
+ },
2020
+ {
2021
+ "epoch": 1.43,
2022
+ "learning_rate": 0.00011261546072717454,
2023
+ "loss": 0.9303,
2024
+ "step": 336
2025
+ },
2026
+ {
2027
+ "epoch": 1.43,
2028
+ "learning_rate": 0.00011215903412566111,
2029
+ "loss": 0.8762,
2030
+ "step": 337
2031
+ },
2032
+ {
2033
+ "epoch": 1.44,
2034
+ "learning_rate": 0.00011170235027339766,
2035
+ "loss": 0.8965,
2036
+ "step": 338
2037
+ },
2038
+ {
2039
+ "epoch": 1.44,
2040
+ "learning_rate": 0.00011124541883252198,
2041
+ "loss": 0.9352,
2042
+ "step": 339
2043
+ },
2044
+ {
2045
+ "epoch": 1.45,
2046
+ "learning_rate": 0.00011078824947041016,
2047
+ "loss": 0.9081,
2048
+ "step": 340
2049
+ },
2050
+ {
2051
+ "epoch": 1.45,
2052
+ "learning_rate": 0.00011033085185947208,
2053
+ "loss": 0.9391,
2054
+ "step": 341
2055
+ },
2056
+ {
2057
+ "epoch": 1.46,
2058
+ "learning_rate": 0.00010987323567694661,
2059
+ "loss": 0.8827,
2060
+ "step": 342
2061
+ },
2062
+ {
2063
+ "epoch": 1.46,
2064
+ "learning_rate": 0.00010941541060469712,
2065
+ "loss": 0.9139,
2066
+ "step": 343
2067
+ },
2068
+ {
2069
+ "epoch": 1.46,
2070
+ "learning_rate": 0.00010895738632900636,
2071
+ "loss": 0.8791,
2072
+ "step": 344
2073
+ },
2074
+ {
2075
+ "epoch": 1.47,
2076
+ "learning_rate": 0.00010849917254037174,
2077
+ "loss": 0.8964,
2078
+ "step": 345
2079
+ },
2080
+ {
2081
+ "epoch": 1.47,
2082
+ "learning_rate": 0.00010804077893330023,
2083
+ "loss": 0.924,
2084
+ "step": 346
2085
+ },
2086
+ {
2087
+ "epoch": 1.48,
2088
+ "learning_rate": 0.00010758221520610321,
2089
+ "loss": 0.8887,
2090
+ "step": 347
2091
+ },
2092
+ {
2093
+ "epoch": 1.48,
2094
+ "learning_rate": 0.00010712349106069131,
2095
+ "loss": 0.9306,
2096
+ "step": 348
2097
+ },
2098
+ {
2099
+ "epoch": 1.49,
2100
+ "learning_rate": 0.00010666461620236922,
2101
+ "loss": 0.8802,
2102
+ "step": 349
2103
+ },
2104
+ {
2105
+ "epoch": 1.49,
2106
+ "learning_rate": 0.00010620560033963025,
2107
+ "loss": 0.897,
2108
+ "step": 350
2109
+ },
2110
+ {
2111
+ "epoch": 1.49,
2112
+ "learning_rate": 0.00010574645318395095,
2113
+ "loss": 0.9017,
2114
+ "step": 351
2115
+ },
2116
+ {
2117
+ "epoch": 1.5,
2118
+ "learning_rate": 0.00010528718444958567,
2119
+ "loss": 0.8665,
2120
+ "step": 352
2121
+ },
2122
+ {
2123
+ "epoch": 1.5,
2124
+ "learning_rate": 0.00010482780385336106,
2125
+ "loss": 0.9159,
2126
+ "step": 353
2127
+ },
2128
+ {
2129
+ "epoch": 1.51,
2130
+ "learning_rate": 0.00010436832111447034,
2131
+ "loss": 0.8568,
2132
+ "step": 354
2133
+ },
2134
+ {
2135
+ "epoch": 1.51,
2136
+ "learning_rate": 0.00010390874595426794,
2137
+ "loss": 0.9258,
2138
+ "step": 355
2139
+ },
2140
+ {
2141
+ "epoch": 1.51,
2142
+ "learning_rate": 0.00010344908809606353,
2143
+ "loss": 0.8684,
2144
+ "step": 356
2145
+ },
2146
+ {
2147
+ "epoch": 1.52,
2148
+ "learning_rate": 0.00010298935726491648,
2149
+ "loss": 0.8991,
2150
+ "step": 357
2151
+ },
2152
+ {
2153
+ "epoch": 1.52,
2154
+ "learning_rate": 0.00010252956318743006,
2155
+ "loss": 0.8716,
2156
+ "step": 358
2157
+ },
2158
+ {
2159
+ "epoch": 1.53,
2160
+ "learning_rate": 0.0001020697155915457,
2161
+ "loss": 0.9192,
2162
+ "step": 359
2163
+ },
2164
+ {
2165
+ "epoch": 1.53,
2166
+ "learning_rate": 0.000101609824206337,
2167
+ "loss": 0.8241,
2168
+ "step": 360
2169
+ },
2170
+ {
2171
+ "epoch": 1.54,
2172
+ "learning_rate": 0.00010114989876180423,
2173
+ "loss": 0.8991,
2174
+ "step": 361
2175
+ },
2176
+ {
2177
+ "epoch": 1.54,
2178
+ "learning_rate": 0.00010068994898866804,
2179
+ "loss": 0.8729,
2180
+ "step": 362
2181
+ },
2182
+ {
2183
+ "epoch": 1.54,
2184
+ "learning_rate": 0.00010022998461816389,
2185
+ "loss": 0.9218,
2186
+ "step": 363
2187
+ },
2188
+ {
2189
+ "epoch": 1.55,
2190
+ "learning_rate": 9.977001538183616e-05,
2191
+ "loss": 0.8121,
2192
+ "step": 364
2193
+ },
2194
+ {
2195
+ "epoch": 1.55,
2196
+ "learning_rate": 9.9310051011332e-05,
2197
+ "loss": 0.8794,
2198
+ "step": 365
2199
+ },
2200
+ {
2201
+ "epoch": 1.56,
2202
+ "learning_rate": 9.88501012381958e-05,
2203
+ "loss": 0.9085,
2204
+ "step": 366
2205
+ },
2206
+ {
2207
+ "epoch": 1.56,
2208
+ "learning_rate": 9.839017579366299e-05,
2209
+ "loss": 0.9281,
2210
+ "step": 367
2211
+ },
2212
+ {
2213
+ "epoch": 1.57,
2214
+ "learning_rate": 9.793028440845434e-05,
2215
+ "loss": 0.8583,
2216
+ "step": 368
2217
+ },
2218
+ {
2219
+ "epoch": 1.57,
2220
+ "learning_rate": 9.747043681256996e-05,
2221
+ "loss": 0.8277,
2222
+ "step": 369
2223
+ },
2224
+ {
2225
+ "epoch": 1.57,
2226
+ "learning_rate": 9.701064273508356e-05,
2227
+ "loss": 0.8954,
2228
+ "step": 370
2229
+ },
2230
+ {
2231
+ "epoch": 1.58,
2232
+ "learning_rate": 9.65509119039365e-05,
2233
+ "loss": 0.9306,
2234
+ "step": 371
2235
+ },
2236
+ {
2237
+ "epoch": 1.58,
2238
+ "learning_rate": 9.609125404573211e-05,
2239
+ "loss": 0.9431,
2240
+ "step": 372
2241
+ },
2242
+ {
2243
+ "epoch": 1.59,
2244
+ "learning_rate": 9.563167888552968e-05,
2245
+ "loss": 0.9396,
2246
+ "step": 373
2247
+ },
2248
+ {
2249
+ "epoch": 1.59,
2250
+ "learning_rate": 9.517219614663896e-05,
2251
+ "loss": 0.8953,
2252
+ "step": 374
2253
+ },
2254
+ {
2255
+ "epoch": 1.6,
2256
+ "learning_rate": 9.471281555041432e-05,
2257
+ "loss": 0.8922,
2258
+ "step": 375
2259
+ },
2260
+ {
2261
+ "epoch": 1.6,
2262
+ "learning_rate": 9.425354681604907e-05,
2263
+ "loss": 0.8737,
2264
+ "step": 376
2265
+ },
2266
+ {
2267
+ "epoch": 1.6,
2268
+ "learning_rate": 9.379439966036977e-05,
2269
+ "loss": 0.8988,
2270
+ "step": 377
2271
+ },
2272
+ {
2273
+ "epoch": 1.61,
2274
+ "learning_rate": 9.333538379763079e-05,
2275
+ "loss": 0.8868,
2276
+ "step": 378
2277
+ },
2278
+ {
2279
+ "epoch": 1.61,
2280
+ "learning_rate": 9.28765089393087e-05,
2281
+ "loss": 0.9303,
2282
+ "step": 379
2283
+ },
2284
+ {
2285
+ "epoch": 1.62,
2286
+ "learning_rate": 9.241778479389683e-05,
2287
+ "loss": 0.934,
2288
+ "step": 380
2289
+ },
2290
+ {
2291
+ "epoch": 1.62,
2292
+ "learning_rate": 9.195922106669981e-05,
2293
+ "loss": 0.8812,
2294
+ "step": 381
2295
+ },
2296
+ {
2297
+ "epoch": 1.63,
2298
+ "learning_rate": 9.150082745962828e-05,
2299
+ "loss": 0.867,
2300
+ "step": 382
2301
+ },
2302
+ {
2303
+ "epoch": 1.63,
2304
+ "learning_rate": 9.104261367099365e-05,
2305
+ "loss": 0.8649,
2306
+ "step": 383
2307
+ },
2308
+ {
2309
+ "epoch": 1.63,
2310
+ "learning_rate": 9.058458939530295e-05,
2311
+ "loss": 0.9221,
2312
+ "step": 384
2313
+ },
2314
+ {
2315
+ "epoch": 1.64,
2316
+ "learning_rate": 9.01267643230534e-05,
2317
+ "loss": 0.859,
2318
+ "step": 385
2319
+ },
2320
+ {
2321
+ "epoch": 1.64,
2322
+ "learning_rate": 8.966914814052796e-05,
2323
+ "loss": 0.9446,
2324
+ "step": 386
2325
+ },
2326
+ {
2327
+ "epoch": 1.65,
2328
+ "learning_rate": 8.921175052958985e-05,
2329
+ "loss": 0.885,
2330
+ "step": 387
2331
+ },
2332
+ {
2333
+ "epoch": 1.65,
2334
+ "learning_rate": 8.875458116747806e-05,
2335
+ "loss": 0.8868,
2336
+ "step": 388
2337
+ },
2338
+ {
2339
+ "epoch": 1.66,
2340
+ "learning_rate": 8.829764972660237e-05,
2341
+ "loss": 0.9057,
2342
+ "step": 389
2343
+ },
2344
+ {
2345
+ "epoch": 1.66,
2346
+ "learning_rate": 8.78409658743389e-05,
2347
+ "loss": 0.8473,
2348
+ "step": 390
2349
+ },
2350
+ {
2351
+ "epoch": 1.66,
2352
+ "learning_rate": 8.738453927282548e-05,
2353
+ "loss": 0.9346,
2354
+ "step": 391
2355
+ },
2356
+ {
2357
+ "epoch": 1.67,
2358
+ "learning_rate": 8.692837957875725e-05,
2359
+ "loss": 0.9158,
2360
+ "step": 392
2361
+ },
2362
+ {
2363
+ "epoch": 1.67,
2364
+ "learning_rate": 8.647249644318232e-05,
2365
+ "loss": 0.8827,
2366
+ "step": 393
2367
+ },
2368
+ {
2369
+ "epoch": 1.68,
2370
+ "learning_rate": 8.601689951129757e-05,
2371
+ "loss": 0.8713,
2372
+ "step": 394
2373
+ },
2374
+ {
2375
+ "epoch": 1.68,
2376
+ "learning_rate": 8.556159842224472e-05,
2377
+ "loss": 0.923,
2378
+ "step": 395
2379
+ },
2380
+ {
2381
+ "epoch": 1.69,
2382
+ "learning_rate": 8.510660280890625e-05,
2383
+ "loss": 0.8329,
2384
+ "step": 396
2385
+ },
2386
+ {
2387
+ "epoch": 1.69,
2388
+ "learning_rate": 8.465192229770156e-05,
2389
+ "loss": 0.8489,
2390
+ "step": 397
2391
+ },
2392
+ {
2393
+ "epoch": 1.69,
2394
+ "learning_rate": 8.41975665083835e-05,
2395
+ "loss": 0.8915,
2396
+ "step": 398
2397
+ },
2398
+ {
2399
+ "epoch": 1.7,
2400
+ "learning_rate": 8.374354505383467e-05,
2401
+ "loss": 0.8903,
2402
+ "step": 399
2403
+ },
2404
+ {
2405
+ "epoch": 1.7,
2406
+ "learning_rate": 8.328986753986409e-05,
2407
+ "loss": 0.913,
2408
+ "step": 400
2409
+ },
2410
+ {
2411
+ "epoch": 1.71,
2412
+ "learning_rate": 8.283654356500394e-05,
2413
+ "loss": 0.8553,
2414
+ "step": 401
2415
+ },
2416
+ {
2417
+ "epoch": 1.71,
2418
+ "learning_rate": 8.238358272030658e-05,
2419
+ "loss": 0.905,
2420
+ "step": 402
2421
+ },
2422
+ {
2423
+ "epoch": 1.71,
2424
+ "learning_rate": 8.193099458914148e-05,
2425
+ "loss": 0.8946,
2426
+ "step": 403
2427
+ },
2428
+ {
2429
+ "epoch": 1.72,
2430
+ "learning_rate": 8.147878874699274e-05,
2431
+ "loss": 0.9124,
2432
+ "step": 404
2433
+ },
2434
+ {
2435
+ "epoch": 1.72,
2436
+ "learning_rate": 8.102697476125597e-05,
2437
+ "loss": 0.8419,
2438
+ "step": 405
2439
+ },
2440
+ {
2441
+ "epoch": 1.73,
2442
+ "learning_rate": 8.057556219103653e-05,
2443
+ "loss": 0.905,
2444
+ "step": 406
2445
+ },
2446
+ {
2447
+ "epoch": 1.73,
2448
+ "learning_rate": 8.012456058694678e-05,
2449
+ "loss": 0.8736,
2450
+ "step": 407
2451
+ },
2452
+ {
2453
+ "epoch": 1.74,
2454
+ "learning_rate": 7.967397949090431e-05,
2455
+ "loss": 0.8963,
2456
+ "step": 408
2457
+ },
2458
+ {
2459
+ "epoch": 1.74,
2460
+ "learning_rate": 7.922382843592984e-05,
2461
+ "loss": 0.9139,
2462
+ "step": 409
2463
+ },
2464
+ {
2465
+ "epoch": 1.74,
2466
+ "learning_rate": 7.87741169459457e-05,
2467
+ "loss": 0.8537,
2468
+ "step": 410
2469
+ },
2470
+ {
2471
+ "epoch": 1.75,
2472
+ "learning_rate": 7.832485453557424e-05,
2473
+ "loss": 0.8808,
2474
+ "step": 411
2475
+ },
2476
+ {
2477
+ "epoch": 1.75,
2478
+ "learning_rate": 7.787605070993668e-05,
2479
+ "loss": 0.867,
2480
+ "step": 412
2481
+ },
2482
+ {
2483
+ "epoch": 1.76,
2484
+ "learning_rate": 7.742771496445167e-05,
2485
+ "loss": 0.8718,
2486
+ "step": 413
2487
+ },
2488
+ {
2489
+ "epoch": 1.76,
2490
+ "learning_rate": 7.697985678463476e-05,
2491
+ "loss": 0.8623,
2492
+ "step": 414
2493
+ },
2494
+ {
2495
+ "epoch": 1.77,
2496
+ "learning_rate": 7.653248564589751e-05,
2497
+ "loss": 0.8879,
2498
+ "step": 415
2499
+ },
2500
+ {
2501
+ "epoch": 1.77,
2502
+ "learning_rate": 7.608561101334714e-05,
2503
+ "loss": 0.854,
2504
+ "step": 416
2505
+ },
2506
+ {
2507
+ "epoch": 1.77,
2508
+ "learning_rate": 7.563924234158607e-05,
2509
+ "loss": 0.9452,
2510
+ "step": 417
2511
+ },
2512
+ {
2513
+ "epoch": 1.78,
2514
+ "learning_rate": 7.519338907451215e-05,
2515
+ "loss": 0.8661,
2516
+ "step": 418
2517
+ },
2518
+ {
2519
+ "epoch": 1.78,
2520
+ "learning_rate": 7.474806064511863e-05,
2521
+ "loss": 0.8852,
2522
+ "step": 419
2523
+ },
2524
+ {
2525
+ "epoch": 1.79,
2526
+ "learning_rate": 7.43032664752948e-05,
2527
+ "loss": 0.8795,
2528
+ "step": 420
2529
+ },
2530
+ {
2531
+ "epoch": 1.79,
2532
+ "learning_rate": 7.385901597562637e-05,
2533
+ "loss": 0.8973,
2534
+ "step": 421
2535
+ },
2536
+ {
2537
+ "epoch": 1.8,
2538
+ "learning_rate": 7.341531854519664e-05,
2539
+ "loss": 0.8834,
2540
+ "step": 422
2541
+ },
2542
+ {
2543
+ "epoch": 1.8,
2544
+ "learning_rate": 7.29721835713875e-05,
2545
+ "loss": 0.8979,
2546
+ "step": 423
2547
+ },
2548
+ {
2549
+ "epoch": 1.8,
2550
+ "learning_rate": 7.25296204296809e-05,
2551
+ "loss": 0.8892,
2552
+ "step": 424
2553
+ },
2554
+ {
2555
+ "epoch": 1.81,
2556
+ "learning_rate": 7.208763848346029e-05,
2557
+ "loss": 0.8541,
2558
+ "step": 425
2559
+ },
2560
+ {
2561
+ "epoch": 1.81,
2562
+ "learning_rate": 7.164624708381285e-05,
2563
+ "loss": 0.8822,
2564
+ "step": 426
2565
+ },
2566
+ {
2567
+ "epoch": 1.82,
2568
+ "learning_rate": 7.120545556933138e-05,
2569
+ "loss": 0.854,
2570
+ "step": 427
2571
+ },
2572
+ {
2573
+ "epoch": 1.82,
2574
+ "learning_rate": 7.076527326591682e-05,
2575
+ "loss": 0.8699,
2576
+ "step": 428
2577
+ },
2578
+ {
2579
+ "epoch": 1.83,
2580
+ "learning_rate": 7.03257094865809e-05,
2581
+ "loss": 0.9127,
2582
+ "step": 429
2583
+ },
2584
+ {
2585
+ "epoch": 1.83,
2586
+ "learning_rate": 6.988677353124913e-05,
2587
+ "loss": 0.894,
2588
+ "step": 430
2589
+ },
2590
+ {
2591
+ "epoch": 1.83,
2592
+ "learning_rate": 6.94484746865641e-05,
2593
+ "loss": 0.9312,
2594
+ "step": 431
2595
+ },
2596
+ {
2597
+ "epoch": 1.84,
2598
+ "learning_rate": 6.901082222568895e-05,
2599
+ "loss": 0.9396,
2600
+ "step": 432
2601
+ },
2602
+ {
2603
+ "epoch": 1.84,
2604
+ "learning_rate": 6.857382540811101e-05,
2605
+ "loss": 0.8763,
2606
+ "step": 433
2607
+ },
2608
+ {
2609
+ "epoch": 1.85,
2610
+ "learning_rate": 6.813749347944625e-05,
2611
+ "loss": 0.8556,
2612
+ "step": 434
2613
+ },
2614
+ {
2615
+ "epoch": 1.85,
2616
+ "learning_rate": 6.770183567124337e-05,
2617
+ "loss": 0.8683,
2618
+ "step": 435
2619
+ },
2620
+ {
2621
+ "epoch": 1.86,
2622
+ "learning_rate": 6.726686120078862e-05,
2623
+ "loss": 0.8735,
2624
+ "step": 436
2625
+ },
2626
+ {
2627
+ "epoch": 1.86,
2628
+ "learning_rate": 6.683257927091074e-05,
2629
+ "loss": 0.9203,
2630
+ "step": 437
2631
+ },
2632
+ {
2633
+ "epoch": 1.86,
2634
+ "learning_rate": 6.639899906978626e-05,
2635
+ "loss": 0.9083,
2636
+ "step": 438
2637
+ },
2638
+ {
2639
+ "epoch": 1.87,
2640
+ "learning_rate": 6.596612977074515e-05,
2641
+ "loss": 0.9017,
2642
+ "step": 439
2643
+ },
2644
+ {
2645
+ "epoch": 1.87,
2646
+ "learning_rate": 6.553398053207671e-05,
2647
+ "loss": 0.8662,
2648
+ "step": 440
2649
+ },
2650
+ {
2651
+ "epoch": 1.88,
2652
+ "learning_rate": 6.510256049683571e-05,
2653
+ "loss": 0.8486,
2654
+ "step": 441
2655
+ },
2656
+ {
2657
+ "epoch": 1.88,
2658
+ "learning_rate": 6.467187879264916e-05,
2659
+ "loss": 0.8897,
2660
+ "step": 442
2661
+ },
2662
+ {
2663
+ "epoch": 1.89,
2664
+ "learning_rate": 6.4241944531523e-05,
2665
+ "loss": 0.9247,
2666
+ "step": 443
2667
+ },
2668
+ {
2669
+ "epoch": 1.89,
2670
+ "learning_rate": 6.381276680964947e-05,
2671
+ "loss": 0.922,
2672
+ "step": 444
2673
+ },
2674
+ {
2675
+ "epoch": 1.89,
2676
+ "learning_rate": 6.338435470721442e-05,
2677
+ "loss": 0.897,
2678
+ "step": 445
2679
+ },
2680
+ {
2681
+ "epoch": 1.9,
2682
+ "learning_rate": 6.295671728820553e-05,
2683
+ "loss": 0.898,
2684
+ "step": 446
2685
+ },
2686
+ {
2687
+ "epoch": 1.9,
2688
+ "learning_rate": 6.252986360022029e-05,
2689
+ "loss": 0.8948,
2690
+ "step": 447
2691
+ },
2692
+ {
2693
+ "epoch": 1.91,
2694
+ "learning_rate": 6.210380267427467e-05,
2695
+ "loss": 0.8923,
2696
+ "step": 448
2697
+ },
2698
+ {
2699
+ "epoch": 1.91,
2700
+ "learning_rate": 6.167854352461202e-05,
2701
+ "loss": 0.8976,
2702
+ "step": 449
2703
+ },
2704
+ {
2705
+ "epoch": 1.91,
2706
+ "learning_rate": 6.125409514851243e-05,
2707
+ "loss": 0.9255,
2708
+ "step": 450
2709
+ },
2710
+ {
2711
+ "epoch": 1.92,
2712
+ "learning_rate": 6.083046652610224e-05,
2713
+ "loss": 0.8501,
2714
+ "step": 451
2715
+ },
2716
+ {
2717
+ "epoch": 1.92,
2718
+ "learning_rate": 6.040766662016424e-05,
2719
+ "loss": 0.9446,
2720
+ "step": 452
2721
+ },
2722
+ {
2723
+ "epoch": 1.93,
2724
+ "learning_rate": 5.998570437594775e-05,
2725
+ "loss": 0.8504,
2726
+ "step": 453
2727
+ },
2728
+ {
2729
+ "epoch": 1.93,
2730
+ "learning_rate": 5.9564588720979655e-05,
2731
+ "loss": 0.8629,
2732
+ "step": 454
2733
+ },
2734
+ {
2735
+ "epoch": 1.94,
2736
+ "learning_rate": 5.914432856487544e-05,
2737
+ "loss": 0.826,
2738
+ "step": 455
2739
+ },
2740
+ {
2741
+ "epoch": 1.94,
2742
+ "learning_rate": 5.8724932799150586e-05,
2743
+ "loss": 0.883,
2744
+ "step": 456
2745
+ },
2746
+ {
2747
+ "epoch": 1.94,
2748
+ "learning_rate": 5.830641029703254e-05,
2749
+ "loss": 0.8657,
2750
+ "step": 457
2751
+ },
2752
+ {
2753
+ "epoch": 1.95,
2754
+ "learning_rate": 5.788876991327288e-05,
2755
+ "loss": 0.893,
2756
+ "step": 458
2757
+ },
2758
+ {
2759
+ "epoch": 1.95,
2760
+ "learning_rate": 5.747202048396023e-05,
2761
+ "loss": 0.8951,
2762
+ "step": 459
2763
+ },
2764
+ {
2765
+ "epoch": 1.96,
2766
+ "learning_rate": 5.705617082633306e-05,
2767
+ "loss": 0.8968,
2768
+ "step": 460
2769
+ },
2770
+ {
2771
+ "epoch": 1.96,
2772
+ "learning_rate": 5.664122973859313e-05,
2773
+ "loss": 0.849,
2774
+ "step": 461
2775
+ },
2776
+ {
2777
+ "epoch": 1.97,
2778
+ "learning_rate": 5.622720599971952e-05,
2779
+ "loss": 0.8614,
2780
+ "step": 462
2781
+ },
2782
+ {
2783
+ "epoch": 1.97,
2784
+ "learning_rate": 5.5814108369282824e-05,
2785
+ "loss": 0.8236,
2786
+ "step": 463
2787
+ },
2788
+ {
2789
+ "epoch": 1.97,
2790
+ "learning_rate": 5.540194558725973e-05,
2791
+ "loss": 0.8196,
2792
+ "step": 464
2793
+ },
2794
+ {
2795
+ "epoch": 1.98,
2796
+ "learning_rate": 5.4990726373848243e-05,
2797
+ "loss": 0.8639,
2798
+ "step": 465
2799
+ },
2800
+ {
2801
+ "epoch": 1.98,
2802
+ "learning_rate": 5.458045942928309e-05,
2803
+ "loss": 0.8689,
2804
+ "step": 466
2805
+ },
2806
+ {
2807
+ "epoch": 1.99,
2808
+ "learning_rate": 5.417115343365171e-05,
2809
+ "loss": 0.8379,
2810
+ "step": 467
2811
+ },
2812
+ {
2813
+ "epoch": 1.99,
2814
+ "learning_rate": 5.37628170467106e-05,
2815
+ "loss": 0.8694,
2816
+ "step": 468
2817
+ },
2818
+ {
2819
+ "epoch": 2.0,
2820
+ "learning_rate": 5.3355458907701925e-05,
2821
+ "loss": 0.8661,
2822
+ "step": 469
2823
+ },
2824
+ {
2825
+ "epoch": 2.0,
2826
+ "learning_rate": 5.2949087635171144e-05,
2827
+ "loss": 0.6885,
2828
+ "step": 470
2829
+ },
2830
+ {
2831
+ "epoch": 2.0,
2832
+ "learning_rate": 5.254371182678424e-05,
2833
+ "loss": 0.6118,
2834
+ "step": 471
2835
+ },
2836
+ {
2837
+ "epoch": 2.01,
2838
+ "learning_rate": 5.213934005914607e-05,
2839
+ "loss": 0.6482,
2840
+ "step": 472
2841
+ },
2842
+ {
2843
+ "epoch": 2.01,
2844
+ "learning_rate": 5.173598088761874e-05,
2845
+ "loss": 0.5909,
2846
+ "step": 473
2847
+ },
2848
+ {
2849
+ "epoch": 2.02,
2850
+ "learning_rate": 5.133364284614077e-05,
2851
+ "loss": 0.6283,
2852
+ "step": 474
2853
+ },
2854
+ {
2855
+ "epoch": 2.02,
2856
+ "learning_rate": 5.093233444704641e-05,
2857
+ "loss": 0.5609,
2858
+ "step": 475
2859
+ },
2860
+ {
2861
+ "epoch": 2.03,
2862
+ "learning_rate": 5.053206418088572e-05,
2863
+ "loss": 0.5663,
2864
+ "step": 476
2865
+ },
2866
+ {
2867
+ "epoch": 2.03,
2868
+ "learning_rate": 5.0132840516244604e-05,
2869
+ "loss": 0.5876,
2870
+ "step": 477
2871
+ },
2872
+ {
2873
+ "epoch": 2.03,
2874
+ "learning_rate": 4.9734671899565955e-05,
2875
+ "loss": 0.6185,
2876
+ "step": 478
2877
+ },
2878
+ {
2879
+ "epoch": 2.04,
2880
+ "learning_rate": 4.933756675497082e-05,
2881
+ "loss": 0.6254,
2882
+ "step": 479
2883
+ },
2884
+ {
2885
+ "epoch": 2.04,
2886
+ "learning_rate": 4.894153348408021e-05,
2887
+ "loss": 0.5577,
2888
+ "step": 480
2889
+ },
2890
+ {
2891
+ "epoch": 2.05,
2892
+ "learning_rate": 4.8546580465837274e-05,
2893
+ "loss": 0.6386,
2894
+ "step": 481
2895
+ },
2896
+ {
2897
+ "epoch": 2.05,
2898
+ "learning_rate": 4.815271605633012e-05,
2899
+ "loss": 0.5778,
2900
+ "step": 482
2901
+ },
2902
+ {
2903
+ "epoch": 2.06,
2904
+ "learning_rate": 4.775994858861492e-05,
2905
+ "loss": 0.6156,
2906
+ "step": 483
2907
+ },
2908
+ {
2909
+ "epoch": 2.06,
2910
+ "learning_rate": 4.7368286372539775e-05,
2911
+ "loss": 0.6072,
2912
+ "step": 484
2913
+ },
2914
+ {
2915
+ "epoch": 2.06,
2916
+ "learning_rate": 4.697773769456859e-05,
2917
+ "loss": 0.6396,
2918
+ "step": 485
2919
+ },
2920
+ {
2921
+ "epoch": 2.07,
2922
+ "learning_rate": 4.658831081760614e-05,
2923
+ "loss": 0.585,
2924
+ "step": 486
2925
+ },
2926
+ {
2927
+ "epoch": 2.07,
2928
+ "learning_rate": 4.6200013980822954e-05,
2929
+ "loss": 0.5711,
2930
+ "step": 487
2931
+ },
2932
+ {
2933
+ "epoch": 2.08,
2934
+ "learning_rate": 4.5812855399481256e-05,
2935
+ "loss": 0.6314,
2936
+ "step": 488
2937
+ },
2938
+ {
2939
+ "epoch": 2.08,
2940
+ "learning_rate": 4.542684326476082e-05,
2941
+ "loss": 0.6257,
2942
+ "step": 489
2943
+ },
2944
+ {
2945
+ "epoch": 2.09,
2946
+ "learning_rate": 4.504198574358596e-05,
2947
+ "loss": 0.6016,
2948
+ "step": 490
2949
+ },
2950
+ {
2951
+ "epoch": 2.09,
2952
+ "learning_rate": 4.465829097845261e-05,
2953
+ "loss": 0.5966,
2954
+ "step": 491
2955
+ },
2956
+ {
2957
+ "epoch": 2.09,
2958
+ "learning_rate": 4.427576708725609e-05,
2959
+ "loss": 0.5442,
2960
+ "step": 492
2961
+ },
2962
+ {
2963
+ "epoch": 2.1,
2964
+ "learning_rate": 4.389442216311933e-05,
2965
+ "loss": 0.6366,
2966
+ "step": 493
2967
+ },
2968
+ {
2969
+ "epoch": 2.1,
2970
+ "learning_rate": 4.351426427422165e-05,
2971
+ "loss": 0.5575,
2972
+ "step": 494
2973
+ },
2974
+ {
2975
+ "epoch": 2.11,
2976
+ "learning_rate": 4.313530146362809e-05,
2977
+ "loss": 0.6181,
2978
+ "step": 495
2979
+ },
2980
+ {
2981
+ "epoch": 2.11,
2982
+ "learning_rate": 4.275754174911921e-05,
2983
+ "loss": 0.6203,
2984
+ "step": 496
2985
+ },
2986
+ {
2987
+ "epoch": 2.11,
2988
+ "learning_rate": 4.2380993123021385e-05,
2989
+ "loss": 0.5802,
2990
+ "step": 497
2991
+ },
2992
+ {
2993
+ "epoch": 2.12,
2994
+ "learning_rate": 4.200566355203784e-05,
2995
+ "loss": 0.5784,
2996
+ "step": 498
2997
+ },
2998
+ {
2999
+ "epoch": 2.12,
3000
+ "learning_rate": 4.163156097708014e-05,
3001
+ "loss": 0.6171,
3002
+ "step": 499
3003
+ },
3004
+ {
3005
+ "epoch": 2.13,
3006
+ "learning_rate": 4.1258693313099996e-05,
3007
+ "loss": 0.5918,
3008
+ "step": 500
3009
+ },
3010
+ {
3011
+ "epoch": 2.13,
3012
+ "learning_rate": 4.088706844892182e-05,
3013
+ "loss": 0.5546,
3014
+ "step": 501
3015
+ },
3016
+ {
3017
+ "epoch": 2.14,
3018
+ "learning_rate": 4.051669424707602e-05,
3019
+ "loss": 0.6072,
3020
+ "step": 502
3021
+ },
3022
+ {
3023
+ "epoch": 2.14,
3024
+ "learning_rate": 4.014757854363249e-05,
3025
+ "loss": 0.618,
3026
+ "step": 503
3027
+ },
3028
+ {
3029
+ "epoch": 2.14,
3030
+ "learning_rate": 3.977972914803486e-05,
3031
+ "loss": 0.5796,
3032
+ "step": 504
3033
+ },
3034
+ {
3035
+ "epoch": 2.15,
3036
+ "learning_rate": 3.9413153842935255e-05,
3037
+ "loss": 0.5466,
3038
+ "step": 505
3039
+ },
3040
+ {
3041
+ "epoch": 2.15,
3042
+ "learning_rate": 3.9047860384029675e-05,
3043
+ "loss": 0.5826,
3044
+ "step": 506
3045
+ },
3046
+ {
3047
+ "epoch": 2.16,
3048
+ "learning_rate": 3.868385649989388e-05,
3049
+ "loss": 0.5873,
3050
+ "step": 507
3051
+ },
3052
+ {
3053
+ "epoch": 2.16,
3054
+ "learning_rate": 3.832114989181988e-05,
3055
+ "loss": 0.5975,
3056
+ "step": 508
3057
+ },
3058
+ {
3059
+ "epoch": 2.17,
3060
+ "learning_rate": 3.795974823365287e-05,
3061
+ "loss": 0.6079,
3062
+ "step": 509
3063
+ },
3064
+ {
3065
+ "epoch": 2.17,
3066
+ "learning_rate": 3.759965917162925e-05,
3067
+ "loss": 0.5623,
3068
+ "step": 510
3069
+ },
3070
+ {
3071
+ "epoch": 2.17,
3072
+ "learning_rate": 3.724089032421441e-05,
3073
+ "loss": 0.5978,
3074
+ "step": 511
3075
+ },
3076
+ {
3077
+ "epoch": 2.18,
3078
+ "learning_rate": 3.688344928194181e-05,
3079
+ "loss": 0.5926,
3080
+ "step": 512
3081
+ },
3082
+ {
3083
+ "epoch": 2.18,
3084
+ "learning_rate": 3.652734360725224e-05,
3085
+ "loss": 0.5724,
3086
+ "step": 513
3087
+ },
3088
+ {
3089
+ "epoch": 2.19,
3090
+ "learning_rate": 3.617258083433396e-05,
3091
+ "loss": 0.5538,
3092
+ "step": 514
3093
+ },
3094
+ {
3095
+ "epoch": 2.19,
3096
+ "learning_rate": 3.581916846896318e-05,
3097
+ "loss": 0.5624,
3098
+ "step": 515
3099
+ },
3100
+ {
3101
+ "epoch": 2.2,
3102
+ "learning_rate": 3.546711398834543e-05,
3103
+ "loss": 0.5711,
3104
+ "step": 516
3105
+ },
3106
+ {
3107
+ "epoch": 2.2,
3108
+ "learning_rate": 3.5116424840957065e-05,
3109
+ "loss": 0.5629,
3110
+ "step": 517
3111
+ },
3112
+ {
3113
+ "epoch": 2.2,
3114
+ "learning_rate": 3.476710844638795e-05,
3115
+ "loss": 0.5501,
3116
+ "step": 518
3117
+ },
3118
+ {
3119
+ "epoch": 2.21,
3120
+ "learning_rate": 3.441917219518438e-05,
3121
+ "loss": 0.6033,
3122
+ "step": 519
3123
+ },
3124
+ {
3125
+ "epoch": 2.21,
3126
+ "learning_rate": 3.407262344869272e-05,
3127
+ "loss": 0.5593,
3128
+ "step": 520
3129
+ },
3130
+ {
3131
+ "epoch": 2.22,
3132
+ "learning_rate": 3.3727469538903646e-05,
3133
+ "loss": 0.5355,
3134
+ "step": 521
3135
+ },
3136
+ {
3137
+ "epoch": 2.22,
3138
+ "learning_rate": 3.338371776829705e-05,
3139
+ "loss": 0.5959,
3140
+ "step": 522
3141
+ },
3142
+ {
3143
+ "epoch": 2.23,
3144
+ "learning_rate": 3.3041375409687526e-05,
3145
+ "loss": 0.5828,
3146
+ "step": 523
3147
+ },
3148
+ {
3149
+ "epoch": 2.23,
3150
+ "learning_rate": 3.2700449706070534e-05,
3151
+ "loss": 0.5922,
3152
+ "step": 524
3153
+ },
3154
+ {
3155
+ "epoch": 2.23,
3156
+ "learning_rate": 3.236094787046901e-05,
3157
+ "loss": 0.6114,
3158
+ "step": 525
3159
+ },
3160
+ {
3161
+ "epoch": 2.24,
3162
+ "learning_rate": 3.202287708578097e-05,
3163
+ "loss": 0.605,
3164
+ "step": 526
3165
+ },
3166
+ {
3167
+ "epoch": 2.24,
3168
+ "learning_rate": 3.168624450462746e-05,
3169
+ "loss": 0.6079,
3170
+ "step": 527
3171
+ },
3172
+ {
3173
+ "epoch": 2.25,
3174
+ "learning_rate": 3.13510572492012e-05,
3175
+ "loss": 0.5634,
3176
+ "step": 528
3177
+ },
3178
+ {
3179
+ "epoch": 2.25,
3180
+ "learning_rate": 3.10173224111158e-05,
3181
+ "loss": 0.4974,
3182
+ "step": 529
3183
+ },
3184
+ {
3185
+ "epoch": 2.26,
3186
+ "learning_rate": 3.0685047051255946e-05,
3187
+ "loss": 0.551,
3188
+ "step": 530
3189
+ },
3190
+ {
3191
+ "epoch": 2.26,
3192
+ "learning_rate": 3.035423819962785e-05,
3193
+ "loss": 0.6034,
3194
+ "step": 531
3195
+ },
3196
+ {
3197
+ "epoch": 2.26,
3198
+ "learning_rate": 3.002490285521059e-05,
3199
+ "loss": 0.5759,
3200
+ "step": 532
3201
+ },
3202
+ {
3203
+ "epoch": 2.27,
3204
+ "learning_rate": 2.9697047985807958e-05,
3205
+ "loss": 0.5502,
3206
+ "step": 533
3207
+ },
3208
+ {
3209
+ "epoch": 2.27,
3210
+ "learning_rate": 2.9370680527901116e-05,
3211
+ "loss": 0.5722,
3212
+ "step": 534
3213
+ },
3214
+ {
3215
+ "epoch": 2.28,
3216
+ "learning_rate": 2.904580738650181e-05,
3217
+ "loss": 0.559,
3218
+ "step": 535
3219
+ },
3220
+ {
3221
+ "epoch": 2.28,
3222
+ "learning_rate": 2.872243543500629e-05,
3223
+ "loss": 0.5352,
3224
+ "step": 536
3225
+ },
3226
+ {
3227
+ "epoch": 2.29,
3228
+ "learning_rate": 2.840057151504979e-05,
3229
+ "loss": 0.5857,
3230
+ "step": 537
3231
+ },
3232
+ {
3233
+ "epoch": 2.29,
3234
+ "learning_rate": 2.8080222436361934e-05,
3235
+ "loss": 0.5819,
3236
+ "step": 538
3237
+ },
3238
+ {
3239
+ "epoch": 2.29,
3240
+ "learning_rate": 2.7761394976622658e-05,
3241
+ "loss": 0.5741,
3242
+ "step": 539
3243
+ },
3244
+ {
3245
+ "epoch": 2.3,
3246
+ "learning_rate": 2.7444095881318656e-05,
3247
+ "loss": 0.5694,
3248
+ "step": 540
3249
+ },
3250
+ {
3251
+ "epoch": 2.3,
3252
+ "learning_rate": 2.712833186360072e-05,
3253
+ "loss": 0.5797,
3254
+ "step": 541
3255
+ },
3256
+ {
3257
+ "epoch": 2.31,
3258
+ "learning_rate": 2.6814109604141848e-05,
3259
+ "loss": 0.5583,
3260
+ "step": 542
3261
+ },
3262
+ {
3263
+ "epoch": 2.31,
3264
+ "learning_rate": 2.6501435750995727e-05,
3265
+ "loss": 0.5771,
3266
+ "step": 543
3267
+ },
3268
+ {
3269
+ "epoch": 2.31,
3270
+ "learning_rate": 2.619031691945618e-05,
3271
+ "loss": 0.5822,
3272
+ "step": 544
3273
+ },
3274
+ {
3275
+ "epoch": 2.32,
3276
+ "learning_rate": 2.588075969191718e-05,
3277
+ "loss": 0.5831,
3278
+ "step": 545
3279
+ },
3280
+ {
3281
+ "epoch": 2.32,
3282
+ "learning_rate": 2.5572770617733544e-05,
3283
+ "loss": 0.5801,
3284
+ "step": 546
3285
+ },
3286
+ {
3287
+ "epoch": 2.33,
3288
+ "learning_rate": 2.5266356213082433e-05,
3289
+ "loss": 0.5578,
3290
+ "step": 547
3291
+ },
3292
+ {
3293
+ "epoch": 2.33,
3294
+ "learning_rate": 2.496152296082548e-05,
3295
+ "loss": 0.6124,
3296
+ "step": 548
3297
+ },
3298
+ {
3299
+ "epoch": 2.34,
3300
+ "learning_rate": 2.465827731037147e-05,
3301
+ "loss": 0.6306,
3302
+ "step": 549
3303
+ },
3304
+ {
3305
+ "epoch": 2.34,
3306
+ "learning_rate": 2.4356625677540233e-05,
3307
+ "loss": 0.5585,
3308
+ "step": 550
3309
+ },
3310
+ {
3311
+ "epoch": 2.34,
3312
+ "learning_rate": 2.405657444442657e-05,
3313
+ "loss": 0.5284,
3314
+ "step": 551
3315
+ },
3316
+ {
3317
+ "epoch": 2.35,
3318
+ "learning_rate": 2.3758129959265407e-05,
3319
+ "loss": 0.6024,
3320
+ "step": 552
3321
+ },
3322
+ {
3323
+ "epoch": 2.35,
3324
+ "learning_rate": 2.3461298536297328e-05,
3325
+ "loss": 0.5512,
3326
+ "step": 553
3327
+ },
3328
+ {
3329
+ "epoch": 2.36,
3330
+ "learning_rate": 2.3166086455635218e-05,
3331
+ "loss": 0.5346,
3332
+ "step": 554
3333
+ },
3334
+ {
3335
+ "epoch": 2.36,
3336
+ "learning_rate": 2.2872499963131155e-05,
3337
+ "loss": 0.5329,
3338
+ "step": 555
3339
+ },
3340
+ {
3341
+ "epoch": 2.37,
3342
+ "learning_rate": 2.258054527024451e-05,
3343
+ "loss": 0.5496,
3344
+ "step": 556
3345
+ },
3346
+ {
3347
+ "epoch": 2.37,
3348
+ "learning_rate": 2.2290228553910242e-05,
3349
+ "loss": 0.5877,
3350
+ "step": 557
3351
+ },
3352
+ {
3353
+ "epoch": 2.37,
3354
+ "learning_rate": 2.2001555956408428e-05,
3355
+ "loss": 0.5671,
3356
+ "step": 558
3357
+ },
3358
+ {
3359
+ "epoch": 2.38,
3360
+ "learning_rate": 2.1714533585234244e-05,
3361
+ "loss": 0.5922,
3362
+ "step": 559
3363
+ },
3364
+ {
3365
+ "epoch": 2.38,
3366
+ "learning_rate": 2.142916751296876e-05,
3367
+ "loss": 0.6145,
3368
+ "step": 560
3369
+ },
3370
+ {
3371
+ "epoch": 2.39,
3372
+ "learning_rate": 2.114546377715042e-05,
3373
+ "loss": 0.5264,
3374
+ "step": 561
3375
+ },
3376
+ {
3377
+ "epoch": 2.39,
3378
+ "learning_rate": 2.0863428380147344e-05,
3379
+ "loss": 0.5404,
3380
+ "step": 562
3381
+ },
3382
+ {
3383
+ "epoch": 2.4,
3384
+ "learning_rate": 2.0583067289030335e-05,
3385
+ "loss": 0.6048,
3386
+ "step": 563
3387
+ },
3388
+ {
3389
+ "epoch": 2.4,
3390
+ "learning_rate": 2.030438643544663e-05,
3391
+ "loss": 0.5456,
3392
+ "step": 564
3393
+ },
3394
+ {
3395
+ "epoch": 2.4,
3396
+ "learning_rate": 2.0027391715494347e-05,
3397
+ "loss": 0.5801,
3398
+ "step": 565
3399
+ },
3400
+ {
3401
+ "epoch": 2.41,
3402
+ "learning_rate": 1.9752088989597795e-05,
3403
+ "loss": 0.5813,
3404
+ "step": 566
3405
+ },
3406
+ {
3407
+ "epoch": 2.41,
3408
+ "learning_rate": 1.9478484082383562e-05,
3409
+ "loss": 0.5641,
3410
+ "step": 567
3411
+ },
3412
+ {
3413
+ "epoch": 2.42,
3414
+ "learning_rate": 1.9206582782557136e-05,
3415
+ "loss": 0.5711,
3416
+ "step": 568
3417
+ },
3418
+ {
3419
+ "epoch": 2.42,
3420
+ "learning_rate": 1.893639084278046e-05,
3421
+ "loss": 0.5707,
3422
+ "step": 569
3423
+ },
3424
+ {
3425
+ "epoch": 2.43,
3426
+ "learning_rate": 1.86679139795503e-05,
3427
+ "loss": 0.5731,
3428
+ "step": 570
3429
+ },
3430
+ {
3431
+ "epoch": 2.43,
3432
+ "learning_rate": 1.8401157873077257e-05,
3433
+ "loss": 0.5849,
3434
+ "step": 571
3435
+ },
3436
+ {
3437
+ "epoch": 2.43,
3438
+ "learning_rate": 1.8136128167165578e-05,
3439
+ "loss": 0.581,
3440
+ "step": 572
3441
+ },
3442
+ {
3443
+ "epoch": 2.44,
3444
+ "learning_rate": 1.787283046909376e-05,
3445
+ "loss": 0.613,
3446
+ "step": 573
3447
+ },
3448
+ {
3449
+ "epoch": 2.44,
3450
+ "learning_rate": 1.7611270349495924e-05,
3451
+ "loss": 0.59,
3452
+ "step": 574
3453
+ },
3454
+ {
3455
+ "epoch": 2.45,
3456
+ "learning_rate": 1.735145334224394e-05,
3457
+ "loss": 0.5961,
3458
+ "step": 575
3459
+ },
3460
+ {
3461
+ "epoch": 2.45,
3462
+ "learning_rate": 1.7093384944330393e-05,
3463
+ "loss": 0.5323,
3464
+ "step": 576
3465
+ },
3466
+ {
3467
+ "epoch": 2.46,
3468
+ "learning_rate": 1.6837070615752115e-05,
3469
+ "loss": 0.5652,
3470
+ "step": 577
3471
+ },
3472
+ {
3473
+ "epoch": 2.46,
3474
+ "learning_rate": 1.6582515779394968e-05,
3475
+ "loss": 0.5533,
3476
+ "step": 578
3477
+ },
3478
+ {
3479
+ "epoch": 2.46,
3480
+ "learning_rate": 1.632972582091884e-05,
3481
+ "loss": 0.5579,
3482
+ "step": 579
3483
+ },
3484
+ {
3485
+ "epoch": 2.47,
3486
+ "learning_rate": 1.6078706088643836e-05,
3487
+ "loss": 0.5813,
3488
+ "step": 580
3489
+ },
3490
+ {
3491
+ "epoch": 2.47,
3492
+ "learning_rate": 1.5829461893437015e-05,
3493
+ "loss": 0.5158,
3494
+ "step": 581
3495
+ },
3496
+ {
3497
+ "epoch": 2.48,
3498
+ "learning_rate": 1.558199850860016e-05,
3499
+ "loss": 0.5997,
3500
+ "step": 582
3501
+ },
3502
+ {
3503
+ "epoch": 2.48,
3504
+ "learning_rate": 1.533632116975814e-05,
3505
+ "loss": 0.5398,
3506
+ "step": 583
3507
+ },
3508
+ {
3509
+ "epoch": 2.49,
3510
+ "learning_rate": 1.5092435074748146e-05,
3511
+ "loss": 0.599,
3512
+ "step": 584
3513
+ },
3514
+ {
3515
+ "epoch": 2.49,
3516
+ "learning_rate": 1.48503453835097e-05,
3517
+ "loss": 0.6448,
3518
+ "step": 585
3519
+ },
3520
+ {
3521
+ "epoch": 2.49,
3522
+ "learning_rate": 1.4610057217975526e-05,
3523
+ "loss": 0.589,
3524
+ "step": 586
3525
+ },
3526
+ {
3527
+ "epoch": 2.5,
3528
+ "learning_rate": 1.4371575661963143e-05,
3529
+ "loss": 0.5743,
3530
+ "step": 587
3531
+ },
3532
+ {
3533
+ "epoch": 2.5,
3534
+ "learning_rate": 1.4134905761067329e-05,
3535
+ "loss": 0.5822,
3536
+ "step": 588
3537
+ },
3538
+ {
3539
+ "epoch": 2.51,
3540
+ "learning_rate": 1.3900052522553397e-05,
3541
+ "loss": 0.5397,
3542
+ "step": 589
3543
+ },
3544
+ {
3545
+ "epoch": 2.51,
3546
+ "learning_rate": 1.3667020915251173e-05,
3547
+ "loss": 0.5798,
3548
+ "step": 590
3549
+ },
3550
+ {
3551
+ "epoch": 2.51,
3552
+ "learning_rate": 1.3435815869449964e-05,
3553
+ "loss": 0.5432,
3554
+ "step": 591
3555
+ },
3556
+ {
3557
+ "epoch": 2.52,
3558
+ "learning_rate": 1.3206442276794207e-05,
3559
+ "loss": 0.6011,
3560
+ "step": 592
3561
+ },
3562
+ {
3563
+ "epoch": 2.52,
3564
+ "learning_rate": 1.297890499017992e-05,
3565
+ "loss": 0.5629,
3566
+ "step": 593
3567
+ },
3568
+ {
3569
+ "epoch": 2.53,
3570
+ "learning_rate": 1.2753208823652141e-05,
3571
+ "loss": 0.5176,
3572
+ "step": 594
3573
+ },
3574
+ {
3575
+ "epoch": 2.53,
3576
+ "learning_rate": 1.2529358552302972e-05,
3577
+ "loss": 0.5281,
3578
+ "step": 595
3579
+ },
3580
+ {
3581
+ "epoch": 2.54,
3582
+ "learning_rate": 1.2307358912170686e-05,
3583
+ "loss": 0.5945,
3584
+ "step": 596
3585
+ },
3586
+ {
3587
+ "epoch": 2.54,
3588
+ "learning_rate": 1.2087214600139308e-05,
3589
+ "loss": 0.551,
3590
+ "step": 597
3591
+ },
3592
+ {
3593
+ "epoch": 2.54,
3594
+ "learning_rate": 1.1868930273839473e-05,
3595
+ "loss": 0.5338,
3596
+ "step": 598
3597
+ },
3598
+ {
3599
+ "epoch": 2.55,
3600
+ "learning_rate": 1.1652510551549723e-05,
3601
+ "loss": 0.5745,
3602
+ "step": 599
3603
+ },
3604
+ {
3605
+ "epoch": 2.55,
3606
+ "learning_rate": 1.1437960012098892e-05,
3607
+ "loss": 0.5542,
3608
+ "step": 600
3609
+ },
3610
+ {
3611
+ "epoch": 2.56,
3612
+ "learning_rate": 1.1225283194769176e-05,
3613
+ "loss": 0.5808,
3614
+ "step": 601
3615
+ },
3616
+ {
3617
+ "epoch": 2.56,
3618
+ "learning_rate": 1.1014484599200125e-05,
3619
+ "loss": 0.5374,
3620
+ "step": 602
3621
+ },
3622
+ {
3623
+ "epoch": 2.57,
3624
+ "learning_rate": 1.0805568685293422e-05,
3625
+ "loss": 0.5708,
3626
+ "step": 603
3627
+ },
3628
+ {
3629
+ "epoch": 2.57,
3630
+ "learning_rate": 1.0598539873118552e-05,
3631
+ "loss": 0.5936,
3632
+ "step": 604
3633
+ },
3634
+ {
3635
+ "epoch": 2.57,
3636
+ "learning_rate": 1.0393402542819231e-05,
3637
+ "loss": 0.5621,
3638
+ "step": 605
3639
+ },
3640
+ {
3641
+ "epoch": 2.58,
3642
+ "learning_rate": 1.0190161034520795e-05,
3643
+ "loss": 0.5412,
3644
+ "step": 606
3645
+ },
3646
+ {
3647
+ "epoch": 2.58,
3648
+ "learning_rate": 9.988819648238379e-06,
3649
+ "loss": 0.5525,
3650
+ "step": 607
3651
+ },
3652
+ {
3653
+ "epoch": 2.59,
3654
+ "learning_rate": 9.789382643785895e-06,
3655
+ "loss": 0.5622,
3656
+ "step": 608
3657
+ },
3658
+ {
3659
+ "epoch": 2.59,
3660
+ "learning_rate": 9.591854240685882e-06,
3661
+ "loss": 0.5662,
3662
+ "step": 609
3663
+ },
3664
+ {
3665
+ "epoch": 2.6,
3666
+ "learning_rate": 9.396238618080322e-06,
3667
+ "loss": 0.5728,
3668
+ "step": 610
3669
+ },
3670
+ {
3671
+ "epoch": 2.6,
3672
+ "learning_rate": 9.202539914642182e-06,
3673
+ "loss": 0.5531,
3674
+ "step": 611
3675
+ },
3676
+ {
3677
+ "epoch": 2.6,
3678
+ "learning_rate": 9.010762228487813e-06,
3679
+ "loss": 0.5631,
3680
+ "step": 612
3681
+ },
3682
+ {
3683
+ "epoch": 2.61,
3684
+ "learning_rate": 8.820909617090289e-06,
3685
+ "loss": 0.5265,
3686
+ "step": 613
3687
+ },
3688
+ {
3689
+ "epoch": 2.61,
3690
+ "learning_rate": 8.632986097193573e-06,
3691
+ "loss": 0.5755,
3692
+ "step": 614
3693
+ },
3694
+ {
3695
+ "epoch": 2.62,
3696
+ "learning_rate": 8.446995644727473e-06,
3697
+ "loss": 0.5626,
3698
+ "step": 615
3699
+ },
3700
+ {
3701
+ "epoch": 2.62,
3702
+ "learning_rate": 8.2629421947236e-06,
3703
+ "loss": 0.618,
3704
+ "step": 616
3705
+ },
3706
+ {
3707
+ "epoch": 2.63,
3708
+ "learning_rate": 8.080829641232013e-06,
3709
+ "loss": 0.5434,
3710
+ "step": 617
3711
+ },
3712
+ {
3713
+ "epoch": 2.63,
3714
+ "learning_rate": 7.900661837238977e-06,
3715
+ "loss": 0.5201,
3716
+ "step": 618
3717
+ },
3718
+ {
3719
+ "epoch": 2.63,
3720
+ "learning_rate": 7.72244259458531e-06,
3721
+ "loss": 0.5886,
3722
+ "step": 619
3723
+ },
3724
+ {
3725
+ "epoch": 2.64,
3726
+ "learning_rate": 7.546175683885814e-06,
3727
+ "loss": 0.6115,
3728
+ "step": 620
3729
+ },
3730
+ {
3731
+ "epoch": 2.64,
3732
+ "learning_rate": 7.371864834449405e-06,
3733
+ "loss": 0.5531,
3734
+ "step": 621
3735
+ },
3736
+ {
3737
+ "epoch": 2.65,
3738
+ "learning_rate": 7.199513734200369e-06,
3739
+ "loss": 0.5619,
3740
+ "step": 622
3741
+ },
3742
+ {
3743
+ "epoch": 2.65,
3744
+ "learning_rate": 7.029126029600197e-06,
3745
+ "loss": 0.5434,
3746
+ "step": 623
3747
+ },
3748
+ {
3749
+ "epoch": 2.66,
3750
+ "learning_rate": 6.860705325570494e-06,
3751
+ "loss": 0.5542,
3752
+ "step": 624
3753
+ },
3754
+ {
3755
+ "epoch": 2.66,
3756
+ "learning_rate": 6.694255185416687e-06,
3757
+ "loss": 0.5529,
3758
+ "step": 625
3759
+ },
3760
+ {
3761
+ "epoch": 2.66,
3762
+ "learning_rate": 6.529779130752678e-06,
3763
+ "loss": 0.539,
3764
+ "step": 626
3765
+ },
3766
+ {
3767
+ "epoch": 2.67,
3768
+ "learning_rate": 6.3672806414262765e-06,
3769
+ "loss": 0.5577,
3770
+ "step": 627
3771
+ },
3772
+ {
3773
+ "epoch": 2.67,
3774
+ "learning_rate": 6.206763155445627e-06,
3775
+ "loss": 0.5307,
3776
+ "step": 628
3777
+ },
3778
+ {
3779
+ "epoch": 2.68,
3780
+ "learning_rate": 6.0482300689064466e-06,
3781
+ "loss": 0.5418,
3782
+ "step": 629
3783
+ },
3784
+ {
3785
+ "epoch": 2.68,
3786
+ "learning_rate": 5.891684735920167e-06,
3787
+ "loss": 0.5613,
3788
+ "step": 630
3789
+ },
3790
+ {
3791
+ "epoch": 2.69,
3792
+ "learning_rate": 5.737130468542972e-06,
3793
+ "loss": 0.5926,
3794
+ "step": 631
3795
+ },
3796
+ {
3797
+ "epoch": 2.69,
3798
+ "learning_rate": 5.58457053670578e-06,
3799
+ "loss": 0.5289,
3800
+ "step": 632
3801
+ },
3802
+ {
3803
+ "epoch": 2.69,
3804
+ "learning_rate": 5.434008168144944e-06,
3805
+ "loss": 0.5981,
3806
+ "step": 633
3807
+ },
3808
+ {
3809
+ "epoch": 2.7,
3810
+ "learning_rate": 5.2854465483340725e-06,
3811
+ "loss": 0.5581,
3812
+ "step": 634
3813
+ },
3814
+ {
3815
+ "epoch": 2.7,
3816
+ "learning_rate": 5.1388888204165875e-06,
3817
+ "loss": 0.5844,
3818
+ "step": 635
3819
+ },
3820
+ {
3821
+ "epoch": 2.71,
3822
+ "learning_rate": 4.9943380851392604e-06,
3823
+ "loss": 0.5871,
3824
+ "step": 636
3825
+ },
3826
+ {
3827
+ "epoch": 2.71,
3828
+ "learning_rate": 4.851797400786506e-06,
3829
+ "loss": 0.5835,
3830
+ "step": 637
3831
+ },
3832
+ {
3833
+ "epoch": 2.71,
3834
+ "learning_rate": 4.7112697831158126e-06,
3835
+ "loss": 0.5522,
3836
+ "step": 638
3837
+ },
3838
+ {
3839
+ "epoch": 2.72,
3840
+ "learning_rate": 4.572758205293848e-06,
3841
+ "loss": 0.6108,
3842
+ "step": 639
3843
+ },
3844
+ {
3845
+ "epoch": 2.72,
3846
+ "learning_rate": 4.4362655978336e-06,
3847
+ "loss": 0.5668,
3848
+ "step": 640
3849
+ },
3850
+ {
3851
+ "epoch": 2.73,
3852
+ "learning_rate": 4.3017948485323255e-06,
3853
+ "loss": 0.5755,
3854
+ "step": 641
3855
+ },
3856
+ {
3857
+ "epoch": 2.73,
3858
+ "learning_rate": 4.169348802410522e-06,
3859
+ "loss": 0.5658,
3860
+ "step": 642
3861
+ },
3862
+ {
3863
+ "epoch": 2.74,
3864
+ "learning_rate": 4.038930261651674e-06,
3865
+ "loss": 0.5555,
3866
+ "step": 643
3867
+ },
3868
+ {
3869
+ "epoch": 2.74,
3870
+ "learning_rate": 3.910541985543014e-06,
3871
+ "loss": 0.5277,
3872
+ "step": 644
3873
+ },
3874
+ {
3875
+ "epoch": 2.74,
3876
+ "learning_rate": 3.7841866904170798e-06,
3877
+ "loss": 0.5733,
3878
+ "step": 645
3879
+ },
3880
+ {
3881
+ "epoch": 2.75,
3882
+ "learning_rate": 3.6598670495943123e-06,
3883
+ "loss": 0.6289,
3884
+ "step": 646
3885
+ },
3886
+ {
3887
+ "epoch": 2.75,
3888
+ "learning_rate": 3.537585693326484e-06,
3889
+ "loss": 0.5355,
3890
+ "step": 647
3891
+ },
3892
+ {
3893
+ "epoch": 2.76,
3894
+ "learning_rate": 3.4173452087410187e-06,
3895
+ "loss": 0.6279,
3896
+ "step": 648
3897
+ },
3898
+ {
3899
+ "epoch": 2.76,
3900
+ "learning_rate": 3.2991481397862568e-06,
3901
+ "loss": 0.5598,
3902
+ "step": 649
3903
+ },
3904
+ {
3905
+ "epoch": 2.77,
3906
+ "learning_rate": 3.1829969871776555e-06,
3907
+ "loss": 0.5882,
3908
+ "step": 650
3909
+ },
3910
+ {
3911
+ "epoch": 2.77,
3912
+ "learning_rate": 3.0688942083448967e-06,
3913
+ "loss": 0.5646,
3914
+ "step": 651
3915
+ },
3916
+ {
3917
+ "epoch": 2.77,
3918
+ "learning_rate": 2.9568422173798294e-06,
3919
+ "loss": 0.6285,
3920
+ "step": 652
3921
+ },
3922
+ {
3923
+ "epoch": 2.78,
3924
+ "learning_rate": 2.846843384985476e-06,
3925
+ "loss": 0.5294,
3926
+ "step": 653
3927
+ },
3928
+ {
3929
+ "epoch": 2.78,
3930
+ "learning_rate": 2.7389000384257955e-06,
3931
+ "loss": 0.5308,
3932
+ "step": 654
3933
+ },
3934
+ {
3935
+ "epoch": 2.79,
3936
+ "learning_rate": 2.633014461476524e-06,
3937
+ "loss": 0.5364,
3938
+ "step": 655
3939
+ },
3940
+ {
3941
+ "epoch": 2.79,
3942
+ "learning_rate": 2.5291888943767992e-06,
3943
+ "loss": 0.5462,
3944
+ "step": 656
3945
+ },
3946
+ {
3947
+ "epoch": 2.8,
3948
+ "learning_rate": 2.427425533781746e-06,
3949
+ "loss": 0.6125,
3950
+ "step": 657
3951
+ },
3952
+ {
3953
+ "epoch": 2.8,
3954
+ "learning_rate": 2.3277265327160904e-06,
3955
+ "loss": 0.5888,
3956
+ "step": 658
3957
+ },
3958
+ {
3959
+ "epoch": 2.8,
3960
+ "learning_rate": 2.2300940005285374e-06,
3961
+ "loss": 0.5855,
3962
+ "step": 659
3963
+ },
3964
+ {
3965
+ "epoch": 2.81,
3966
+ "learning_rate": 2.134530002847146e-06,
3967
+ "loss": 0.5628,
3968
+ "step": 660
3969
+ },
3970
+ {
3971
+ "epoch": 2.81,
3972
+ "learning_rate": 2.0410365615356365e-06,
3973
+ "loss": 0.5872,
3974
+ "step": 661
3975
+ },
3976
+ {
3977
+ "epoch": 2.82,
3978
+ "learning_rate": 1.9496156546506274e-06,
3979
+ "loss": 0.5642,
3980
+ "step": 662
3981
+ },
3982
+ {
3983
+ "epoch": 2.82,
3984
+ "learning_rate": 1.8602692163997681e-06,
3985
+ "loss": 0.5872,
3986
+ "step": 663
3987
+ },
3988
+ {
3989
+ "epoch": 2.83,
3990
+ "learning_rate": 1.7729991371008502e-06,
3991
+ "loss": 0.6044,
3992
+ "step": 664
3993
+ },
3994
+ {
3995
+ "epoch": 2.83,
3996
+ "learning_rate": 1.6878072631417386e-06,
3997
+ "loss": 0.5491,
3998
+ "step": 665
3999
+ },
4000
+ {
4001
+ "epoch": 2.83,
4002
+ "learning_rate": 1.6046953969413915e-06,
4003
+ "loss": 0.5291,
4004
+ "step": 666
4005
+ },
4006
+ {
4007
+ "epoch": 2.84,
4008
+ "learning_rate": 1.5236652969116804e-06,
4009
+ "loss": 0.539,
4010
+ "step": 667
4011
+ },
4012
+ {
4013
+ "epoch": 2.84,
4014
+ "learning_rate": 1.444718677420176e-06,
4015
+ "loss": 0.5518,
4016
+ "step": 668
4017
+ },
4018
+ {
4019
+ "epoch": 2.85,
4020
+ "learning_rate": 1.367857208753931e-06,
4021
+ "loss": 0.605,
4022
+ "step": 669
4023
+ },
4024
+ {
4025
+ "epoch": 2.85,
4026
+ "learning_rate": 1.2930825170840877e-06,
4027
+ "loss": 0.5938,
4028
+ "step": 670
4029
+ },
4030
+ {
4031
+ "epoch": 2.86,
4032
+ "learning_rate": 1.2203961844315048e-06,
4033
+ "loss": 0.5656,
4034
+ "step": 671
4035
+ },
4036
+ {
4037
+ "epoch": 2.86,
4038
+ "learning_rate": 1.1497997486332513e-06,
4039
+ "loss": 0.5663,
4040
+ "step": 672
4041
+ },
4042
+ {
4043
+ "epoch": 2.86,
4044
+ "learning_rate": 1.0812947033101207e-06,
4045
+ "loss": 0.5731,
4046
+ "step": 673
4047
+ },
4048
+ {
4049
+ "epoch": 2.87,
4050
+ "learning_rate": 1.0148824978349792e-06,
4051
+ "loss": 0.5909,
4052
+ "step": 674
4053
+ },
4054
+ {
4055
+ "epoch": 2.87,
4056
+ "learning_rate": 9.505645373021455e-07,
4057
+ "loss": 0.5732,
4058
+ "step": 675
4059
+ },
4060
+ {
4061
+ "epoch": 2.88,
4062
+ "learning_rate": 8.883421824976479e-07,
4063
+ "loss": 0.5457,
4064
+ "step": 676
4065
+ },
4066
+ {
4067
+ "epoch": 2.88,
4068
+ "learning_rate": 8.282167498703918e-07,
4069
+ "loss": 0.559,
4070
+ "step": 677
4071
+ },
4072
+ {
4073
+ "epoch": 2.89,
4074
+ "learning_rate": 7.701895115043822e-07,
4075
+ "loss": 0.5706,
4076
+ "step": 678
4077
+ },
4078
+ {
4079
+ "epoch": 2.89,
4080
+ "learning_rate": 7.142616950917446e-07,
4081
+ "loss": 0.5509,
4082
+ "step": 679
4083
+ },
4084
+ {
4085
+ "epoch": 2.89,
4086
+ "learning_rate": 6.604344839068021e-07,
4087
+ "loss": 0.524,
4088
+ "step": 680
4089
+ },
4090
+ {
4091
+ "epoch": 2.9,
4092
+ "learning_rate": 6.087090167809839e-07,
4093
+ "loss": 0.5933,
4094
+ "step": 681
4095
+ },
4096
+ {
4097
+ "epoch": 2.9,
4098
+ "learning_rate": 5.590863880788111e-07,
4099
+ "loss": 0.5565,
4100
+ "step": 682
4101
+ },
4102
+ {
4103
+ "epoch": 2.91,
4104
+ "learning_rate": 5.115676476746489e-07,
4105
+ "loss": 0.5286,
4106
+ "step": 683
4107
+ },
4108
+ {
4109
+ "epoch": 2.91,
4110
+ "learning_rate": 4.661538009305577e-07,
4111
+ "loss": 0.629,
4112
+ "step": 684
4113
+ },
4114
+ {
4115
+ "epoch": 2.91,
4116
+ "learning_rate": 4.2284580867500976e-07,
4117
+ "loss": 0.5606,
4118
+ "step": 685
4119
+ },
4120
+ {
4121
+ "epoch": 2.92,
4122
+ "learning_rate": 3.8164458718255025e-07,
4123
+ "loss": 0.5308,
4124
+ "step": 686
4125
+ },
4126
+ {
4127
+ "epoch": 2.92,
4128
+ "learning_rate": 3.4255100815442365e-07,
4129
+ "loss": 0.5966,
4130
+ "step": 687
4131
+ },
4132
+ {
4133
+ "epoch": 2.93,
4134
+ "learning_rate": 3.0556589870012196e-07,
4135
+ "loss": 0.5551,
4136
+ "step": 688
4137
+ },
4138
+ {
4139
+ "epoch": 2.93,
4140
+ "learning_rate": 2.7069004131987653e-07,
4141
+ "loss": 0.5597,
4142
+ "step": 689
4143
+ },
4144
+ {
4145
+ "epoch": 2.94,
4146
+ "learning_rate": 2.379241738881377e-07,
4147
+ "loss": 0.5856,
4148
+ "step": 690
4149
+ },
4150
+ {
4151
+ "epoch": 2.94,
4152
+ "learning_rate": 2.0726898963793205e-07,
4153
+ "loss": 0.5698,
4154
+ "step": 691
4155
+ },
4156
+ {
4157
+ "epoch": 2.94,
4158
+ "learning_rate": 1.78725137146174e-07,
4159
+ "loss": 0.5685,
4160
+ "step": 692
4161
+ },
4162
+ {
4163
+ "epoch": 2.95,
4164
+ "learning_rate": 1.5229322032002115e-07,
4165
+ "loss": 0.5562,
4166
+ "step": 693
4167
+ },
4168
+ {
4169
+ "epoch": 2.95,
4170
+ "learning_rate": 1.27973798384029e-07,
4171
+ "loss": 0.5893,
4172
+ "step": 694
4173
+ },
4174
+ {
4175
+ "epoch": 2.96,
4176
+ "learning_rate": 1.0576738586831614e-07,
4177
+ "loss": 0.5793,
4178
+ "step": 695
4179
+ },
4180
+ {
4181
+ "epoch": 2.96,
4182
+ "learning_rate": 8.567445259775042e-08,
4183
+ "loss": 0.5354,
4184
+ "step": 696
4185
+ },
4186
+ {
4187
+ "epoch": 2.97,
4188
+ "learning_rate": 6.769542368190162e-08,
4189
+ "loss": 0.5889,
4190
+ "step": 697
4191
+ },
4192
+ {
4193
+ "epoch": 2.97,
4194
+ "learning_rate": 5.183067950617071e-08,
4195
+ "loss": 0.5501,
4196
+ "step": 698
4197
+ },
4198
+ {
4199
+ "epoch": 2.97,
4200
+ "learning_rate": 3.808055572362967e-08,
4201
+ "loss": 0.5596,
4202
+ "step": 699
4203
+ },
4204
+ {
4205
+ "epoch": 2.98,
4206
+ "learning_rate": 2.6445343247982755e-08,
4207
+ "loss": 0.5942,
4208
+ "step": 700
4209
+ },
4210
+ {
4211
+ "epoch": 2.98,
4212
+ "learning_rate": 1.6925288247393588e-08,
4213
+ "loss": 0.5747,
4214
+ "step": 701
4215
+ },
4216
+ {
4217
+ "epoch": 2.99,
4218
+ "learning_rate": 9.52059213927825e-09,
4219
+ "loss": 0.5802,
4220
+ "step": 702
4221
+ },
4222
+ {
4223
+ "epoch": 2.99,
4224
+ "learning_rate": 4.231411586064216e-09,
4225
+ "loss": 0.5792,
4226
+ "step": 703
4227
+ },
4228
+ {
4229
+ "epoch": 3.0,
4230
+ "learning_rate": 1.0578584918374823e-09,
4231
+ "loss": 0.5585,
4232
+ "step": 704
4233
+ },
4234
+ {
4235
+ "epoch": 3.0,
4236
+ "learning_rate": 0.0,
4237
+ "loss": 0.4152,
4238
+ "step": 705
4239
+ },
4240
+ {
4241
+ "epoch": 3.0,
4242
+ "step": 705,
4243
+ "total_flos": 59407669002240.0,
4244
+ "train_loss": 0.9080784771036594,
4245
+ "train_runtime": 4849.5227,
4246
+ "train_samples_per_second": 18.568,
4247
+ "train_steps_per_second": 0.145
4248
+ }
4249
+ ],
4250
+ "max_steps": 705,
4251
+ "num_train_epochs": 3,
4252
+ "total_flos": 59407669002240.0,
4253
+ "trial_name": null,
4254
+ "trial_params": null
4255
+ }