chitanda commited on
Commit
596e405
·
verified ·
1 Parent(s): 1b058a0

Add files using upload-large-folder tool

Browse files
llava-v1.6-13b-unk-vqa-v1.1/checkpoint-200/config.json ADDED
@@ -0,0 +1,74 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "../pretrained-models/llava-v1.6-vicuna-13b",
3
+ "architectures": [
4
+ "LlavaLlamaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "bos_token_id": 1,
9
+ "eos_token_id": 2,
10
+ "freeze_mm_mlp_adapter": false,
11
+ "freeze_mm_vision_resampler": false,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 5120,
14
+ "image_aspect_ratio": "pad",
15
+ "image_crop_resolution": 224,
16
+ "image_grid_pinpoints": [
17
+ [
18
+ 336,
19
+ 672
20
+ ],
21
+ [
22
+ 672,
23
+ 336
24
+ ],
25
+ [
26
+ 672,
27
+ 672
28
+ ],
29
+ [
30
+ 1008,
31
+ 336
32
+ ],
33
+ [
34
+ 336,
35
+ 1008
36
+ ]
37
+ ],
38
+ "image_split_resolution": 224,
39
+ "initializer_range": 0.02,
40
+ "intermediate_size": 13824,
41
+ "max_length": 4096,
42
+ "max_position_embeddings": 4096,
43
+ "mm_hidden_size": 1024,
44
+ "mm_patch_merge_type": "flat",
45
+ "mm_projector_lr": null,
46
+ "mm_projector_type": "mlp2x_gelu",
47
+ "mm_resampler_type": null,
48
+ "mm_use_im_patch_token": false,
49
+ "mm_use_im_start_end": false,
50
+ "mm_vision_select_feature": "patch",
51
+ "mm_vision_select_layer": -2,
52
+ "mm_vision_tower": "openai/clip-vit-large-patch14-336",
53
+ "mm_vision_tower_lr": 2e-06,
54
+ "model_type": "llava_llama",
55
+ "num_attention_heads": 40,
56
+ "num_hidden_layers": 40,
57
+ "num_key_value_heads": 40,
58
+ "pad_token_id": 0,
59
+ "pretraining_tp": 1,
60
+ "rms_norm_eps": 1e-05,
61
+ "rope_scaling": null,
62
+ "rope_theta": 10000.0,
63
+ "tie_word_embeddings": false,
64
+ "tokenizer_model_max_length": 2048,
65
+ "tokenizer_padding_side": "right",
66
+ "torch_dtype": "bfloat16",
67
+ "transformers_version": "4.37.2",
68
+ "tune_mm_mlp_adapter": false,
69
+ "tune_mm_vision_resampler": false,
70
+ "unfreeze_mm_vision_tower": true,
71
+ "use_cache": false,
72
+ "use_mm_proj": true,
73
+ "vocab_size": 32000
74
+ }
llava-v1.6-13b-unk-vqa-v1.1/checkpoint-600/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad8a35afd8967cbb748405387e44426e43ad127028e826eddc9b67d2ca873c85
3
+ size 15984
llava-v1.6-13b-unk-vqa-v1.1/checkpoint-600/trainer_state.json ADDED
@@ -0,0 +1,3621 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.1538461538461537,
5
+ "eval_steps": 500,
6
+ "global_step": 600,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 6.25e-07,
14
+ "loss": 3.1205,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.0,
19
+ "learning_rate": 1.25e-06,
20
+ "loss": 3.2041,
21
+ "step": 2
22
+ },
23
+ {
24
+ "epoch": 0.01,
25
+ "learning_rate": 1.8750000000000003e-06,
26
+ "loss": 3.0927,
27
+ "step": 3
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 2.5e-06,
32
+ "loss": 3.1433,
33
+ "step": 4
34
+ },
35
+ {
36
+ "epoch": 0.01,
37
+ "learning_rate": 3.125e-06,
38
+ "loss": 3.036,
39
+ "step": 5
40
+ },
41
+ {
42
+ "epoch": 0.01,
43
+ "learning_rate": 3.7500000000000005e-06,
44
+ "loss": 2.4568,
45
+ "step": 6
46
+ },
47
+ {
48
+ "epoch": 0.01,
49
+ "learning_rate": 4.3750000000000005e-06,
50
+ "loss": 1.2898,
51
+ "step": 7
52
+ },
53
+ {
54
+ "epoch": 0.02,
55
+ "learning_rate": 5e-06,
56
+ "loss": 0.97,
57
+ "step": 8
58
+ },
59
+ {
60
+ "epoch": 0.02,
61
+ "learning_rate": 5.625e-06,
62
+ "loss": 0.6386,
63
+ "step": 9
64
+ },
65
+ {
66
+ "epoch": 0.02,
67
+ "learning_rate": 6.25e-06,
68
+ "loss": 0.4549,
69
+ "step": 10
70
+ },
71
+ {
72
+ "epoch": 0.02,
73
+ "learning_rate": 6.875e-06,
74
+ "loss": 0.4718,
75
+ "step": 11
76
+ },
77
+ {
78
+ "epoch": 0.02,
79
+ "learning_rate": 7.500000000000001e-06,
80
+ "loss": 0.4106,
81
+ "step": 12
82
+ },
83
+ {
84
+ "epoch": 0.03,
85
+ "learning_rate": 8.125000000000001e-06,
86
+ "loss": 0.3442,
87
+ "step": 13
88
+ },
89
+ {
90
+ "epoch": 0.03,
91
+ "learning_rate": 8.750000000000001e-06,
92
+ "loss": 0.3178,
93
+ "step": 14
94
+ },
95
+ {
96
+ "epoch": 0.03,
97
+ "learning_rate": 9.375000000000001e-06,
98
+ "loss": 0.2579,
99
+ "step": 15
100
+ },
101
+ {
102
+ "epoch": 0.03,
103
+ "learning_rate": 1e-05,
104
+ "loss": 0.3623,
105
+ "step": 16
106
+ },
107
+ {
108
+ "epoch": 0.03,
109
+ "learning_rate": 1.0625e-05,
110
+ "loss": 0.3338,
111
+ "step": 17
112
+ },
113
+ {
114
+ "epoch": 0.03,
115
+ "learning_rate": 1.125e-05,
116
+ "loss": 0.3265,
117
+ "step": 18
118
+ },
119
+ {
120
+ "epoch": 0.04,
121
+ "learning_rate": 1.1875e-05,
122
+ "loss": 0.3804,
123
+ "step": 19
124
+ },
125
+ {
126
+ "epoch": 0.04,
127
+ "learning_rate": 1.25e-05,
128
+ "loss": 0.2543,
129
+ "step": 20
130
+ },
131
+ {
132
+ "epoch": 0.04,
133
+ "learning_rate": 1.3125e-05,
134
+ "loss": 0.3619,
135
+ "step": 21
136
+ },
137
+ {
138
+ "epoch": 0.04,
139
+ "learning_rate": 1.375e-05,
140
+ "loss": 0.3095,
141
+ "step": 22
142
+ },
143
+ {
144
+ "epoch": 0.04,
145
+ "learning_rate": 1.4375e-05,
146
+ "loss": 0.2835,
147
+ "step": 23
148
+ },
149
+ {
150
+ "epoch": 0.05,
151
+ "learning_rate": 1.5000000000000002e-05,
152
+ "loss": 0.2666,
153
+ "step": 24
154
+ },
155
+ {
156
+ "epoch": 0.05,
157
+ "learning_rate": 1.5625e-05,
158
+ "loss": 0.3076,
159
+ "step": 25
160
+ },
161
+ {
162
+ "epoch": 0.05,
163
+ "learning_rate": 1.6250000000000002e-05,
164
+ "loss": 0.2789,
165
+ "step": 26
166
+ },
167
+ {
168
+ "epoch": 0.05,
169
+ "learning_rate": 1.6875e-05,
170
+ "loss": 0.3008,
171
+ "step": 27
172
+ },
173
+ {
174
+ "epoch": 0.05,
175
+ "learning_rate": 1.7500000000000002e-05,
176
+ "loss": 0.2998,
177
+ "step": 28
178
+ },
179
+ {
180
+ "epoch": 0.06,
181
+ "learning_rate": 1.8125e-05,
182
+ "loss": 0.272,
183
+ "step": 29
184
+ },
185
+ {
186
+ "epoch": 0.06,
187
+ "learning_rate": 1.8750000000000002e-05,
188
+ "loss": 0.371,
189
+ "step": 30
190
+ },
191
+ {
192
+ "epoch": 0.06,
193
+ "learning_rate": 1.9375e-05,
194
+ "loss": 0.3543,
195
+ "step": 31
196
+ },
197
+ {
198
+ "epoch": 0.06,
199
+ "learning_rate": 2e-05,
200
+ "loss": 0.2793,
201
+ "step": 32
202
+ },
203
+ {
204
+ "epoch": 0.06,
205
+ "learning_rate": 1.9999951432210905e-05,
206
+ "loss": 0.278,
207
+ "step": 33
208
+ },
209
+ {
210
+ "epoch": 0.07,
211
+ "learning_rate": 1.9999805729315383e-05,
212
+ "loss": 0.3078,
213
+ "step": 34
214
+ },
215
+ {
216
+ "epoch": 0.07,
217
+ "learning_rate": 1.999956289272873e-05,
218
+ "loss": 0.3137,
219
+ "step": 35
220
+ },
221
+ {
222
+ "epoch": 0.07,
223
+ "learning_rate": 1.999922292480975e-05,
224
+ "loss": 0.2804,
225
+ "step": 36
226
+ },
227
+ {
228
+ "epoch": 0.07,
229
+ "learning_rate": 1.9998785828860744e-05,
230
+ "loss": 0.2322,
231
+ "step": 37
232
+ },
233
+ {
234
+ "epoch": 0.07,
235
+ "learning_rate": 1.9998251609127465e-05,
236
+ "loss": 0.2754,
237
+ "step": 38
238
+ },
239
+ {
240
+ "epoch": 0.07,
241
+ "learning_rate": 1.999762027079909e-05,
242
+ "loss": 0.3334,
243
+ "step": 39
244
+ },
245
+ {
246
+ "epoch": 0.08,
247
+ "learning_rate": 1.9996891820008165e-05,
248
+ "loss": 0.3721,
249
+ "step": 40
250
+ },
251
+ {
252
+ "epoch": 0.08,
253
+ "learning_rate": 1.9996066263830533e-05,
254
+ "loss": 0.3324,
255
+ "step": 41
256
+ },
257
+ {
258
+ "epoch": 0.08,
259
+ "learning_rate": 1.9995143610285275e-05,
260
+ "loss": 0.2972,
261
+ "step": 42
262
+ },
263
+ {
264
+ "epoch": 0.08,
265
+ "learning_rate": 1.9994123868334655e-05,
266
+ "loss": 0.3039,
267
+ "step": 43
268
+ },
269
+ {
270
+ "epoch": 0.08,
271
+ "learning_rate": 1.9993007047883988e-05,
272
+ "loss": 0.2436,
273
+ "step": 44
274
+ },
275
+ {
276
+ "epoch": 0.09,
277
+ "learning_rate": 1.999179315978157e-05,
278
+ "loss": 0.3353,
279
+ "step": 45
280
+ },
281
+ {
282
+ "epoch": 0.09,
283
+ "learning_rate": 1.999048221581858e-05,
284
+ "loss": 0.2819,
285
+ "step": 46
286
+ },
287
+ {
288
+ "epoch": 0.09,
289
+ "learning_rate": 1.9989074228728942e-05,
290
+ "loss": 0.2769,
291
+ "step": 47
292
+ },
293
+ {
294
+ "epoch": 0.09,
295
+ "learning_rate": 1.9987569212189224e-05,
296
+ "loss": 0.253,
297
+ "step": 48
298
+ },
299
+ {
300
+ "epoch": 0.09,
301
+ "learning_rate": 1.9985967180818493e-05,
302
+ "loss": 0.2614,
303
+ "step": 49
304
+ },
305
+ {
306
+ "epoch": 0.1,
307
+ "learning_rate": 1.998426815017817e-05,
308
+ "loss": 0.3297,
309
+ "step": 50
310
+ },
311
+ {
312
+ "epoch": 0.1,
313
+ "learning_rate": 1.998247213677188e-05,
314
+ "loss": 0.3112,
315
+ "step": 51
316
+ },
317
+ {
318
+ "epoch": 0.1,
319
+ "learning_rate": 1.9980579158045322e-05,
320
+ "loss": 0.2611,
321
+ "step": 52
322
+ },
323
+ {
324
+ "epoch": 0.1,
325
+ "learning_rate": 1.9978589232386036e-05,
326
+ "loss": 0.3002,
327
+ "step": 53
328
+ },
329
+ {
330
+ "epoch": 0.1,
331
+ "learning_rate": 1.997650237912329e-05,
332
+ "loss": 0.2485,
333
+ "step": 54
334
+ },
335
+ {
336
+ "epoch": 0.11,
337
+ "learning_rate": 1.997431861852785e-05,
338
+ "loss": 0.287,
339
+ "step": 55
340
+ },
341
+ {
342
+ "epoch": 0.11,
343
+ "learning_rate": 1.9972037971811802e-05,
344
+ "loss": 0.331,
345
+ "step": 56
346
+ },
347
+ {
348
+ "epoch": 0.11,
349
+ "learning_rate": 1.996966046112834e-05,
350
+ "loss": 0.2785,
351
+ "step": 57
352
+ },
353
+ {
354
+ "epoch": 0.11,
355
+ "learning_rate": 1.996718610957155e-05,
356
+ "loss": 0.3088,
357
+ "step": 58
358
+ },
359
+ {
360
+ "epoch": 0.11,
361
+ "learning_rate": 1.9964614941176194e-05,
362
+ "loss": 0.2558,
363
+ "step": 59
364
+ },
365
+ {
366
+ "epoch": 0.12,
367
+ "learning_rate": 1.9961946980917457e-05,
368
+ "loss": 0.3732,
369
+ "step": 60
370
+ },
371
+ {
372
+ "epoch": 0.12,
373
+ "learning_rate": 1.995918225471073e-05,
374
+ "loss": 0.2935,
375
+ "step": 61
376
+ },
377
+ {
378
+ "epoch": 0.12,
379
+ "learning_rate": 1.9956320789411338e-05,
380
+ "loss": 0.2946,
381
+ "step": 62
382
+ },
383
+ {
384
+ "epoch": 0.12,
385
+ "learning_rate": 1.9953362612814294e-05,
386
+ "loss": 0.2806,
387
+ "step": 63
388
+ },
389
+ {
390
+ "epoch": 0.12,
391
+ "learning_rate": 1.9950307753654016e-05,
392
+ "loss": 0.2934,
393
+ "step": 64
394
+ },
395
+ {
396
+ "epoch": 0.12,
397
+ "learning_rate": 1.994715624160405e-05,
398
+ "loss": 0.3366,
399
+ "step": 65
400
+ },
401
+ {
402
+ "epoch": 0.13,
403
+ "learning_rate": 1.99439081072768e-05,
404
+ "loss": 0.2754,
405
+ "step": 66
406
+ },
407
+ {
408
+ "epoch": 0.13,
409
+ "learning_rate": 1.9940563382223196e-05,
410
+ "loss": 0.3244,
411
+ "step": 67
412
+ },
413
+ {
414
+ "epoch": 0.13,
415
+ "learning_rate": 1.9937122098932428e-05,
416
+ "loss": 0.3061,
417
+ "step": 68
418
+ },
419
+ {
420
+ "epoch": 0.13,
421
+ "learning_rate": 1.9933584290831593e-05,
422
+ "loss": 0.4533,
423
+ "step": 69
424
+ },
425
+ {
426
+ "epoch": 0.13,
427
+ "learning_rate": 1.9929949992285397e-05,
428
+ "loss": 1.2644,
429
+ "step": 70
430
+ },
431
+ {
432
+ "epoch": 0.14,
433
+ "learning_rate": 1.992621923859581e-05,
434
+ "loss": 0.3184,
435
+ "step": 71
436
+ },
437
+ {
438
+ "epoch": 0.14,
439
+ "learning_rate": 1.9922392066001724e-05,
440
+ "loss": 0.2933,
441
+ "step": 72
442
+ },
443
+ {
444
+ "epoch": 0.14,
445
+ "learning_rate": 1.99184685116786e-05,
446
+ "loss": 0.3077,
447
+ "step": 73
448
+ },
449
+ {
450
+ "epoch": 0.14,
451
+ "learning_rate": 1.9914448613738107e-05,
452
+ "loss": 0.3038,
453
+ "step": 74
454
+ },
455
+ {
456
+ "epoch": 0.14,
457
+ "learning_rate": 1.991033241122776e-05,
458
+ "loss": 0.3063,
459
+ "step": 75
460
+ },
461
+ {
462
+ "epoch": 0.15,
463
+ "learning_rate": 1.9906119944130527e-05,
464
+ "loss": 0.311,
465
+ "step": 76
466
+ },
467
+ {
468
+ "epoch": 0.15,
469
+ "learning_rate": 1.9901811253364458e-05,
470
+ "loss": 0.3081,
471
+ "step": 77
472
+ },
473
+ {
474
+ "epoch": 0.15,
475
+ "learning_rate": 1.9897406380782262e-05,
476
+ "loss": 0.261,
477
+ "step": 78
478
+ },
479
+ {
480
+ "epoch": 0.15,
481
+ "learning_rate": 1.989290536917093e-05,
482
+ "loss": 0.3248,
483
+ "step": 79
484
+ },
485
+ {
486
+ "epoch": 0.15,
487
+ "learning_rate": 1.9888308262251286e-05,
488
+ "loss": 0.2814,
489
+ "step": 80
490
+ },
491
+ {
492
+ "epoch": 0.16,
493
+ "learning_rate": 1.988361510467761e-05,
494
+ "loss": 0.2683,
495
+ "step": 81
496
+ },
497
+ {
498
+ "epoch": 0.16,
499
+ "learning_rate": 1.9878825942037147e-05,
500
+ "loss": 0.2755,
501
+ "step": 82
502
+ },
503
+ {
504
+ "epoch": 0.16,
505
+ "learning_rate": 1.9873940820849714e-05,
506
+ "loss": 0.3489,
507
+ "step": 83
508
+ },
509
+ {
510
+ "epoch": 0.16,
511
+ "learning_rate": 1.9868959788567213e-05,
512
+ "loss": 0.3311,
513
+ "step": 84
514
+ },
515
+ {
516
+ "epoch": 0.16,
517
+ "learning_rate": 1.9863882893573188e-05,
518
+ "loss": 0.3237,
519
+ "step": 85
520
+ },
521
+ {
522
+ "epoch": 0.17,
523
+ "learning_rate": 1.985871018518236e-05,
524
+ "loss": 0.3344,
525
+ "step": 86
526
+ },
527
+ {
528
+ "epoch": 0.17,
529
+ "learning_rate": 1.9853441713640123e-05,
530
+ "loss": 0.3195,
531
+ "step": 87
532
+ },
533
+ {
534
+ "epoch": 0.17,
535
+ "learning_rate": 1.9848077530122083e-05,
536
+ "loss": 0.3099,
537
+ "step": 88
538
+ },
539
+ {
540
+ "epoch": 0.17,
541
+ "learning_rate": 1.9842617686733546e-05,
542
+ "loss": 0.271,
543
+ "step": 89
544
+ },
545
+ {
546
+ "epoch": 0.17,
547
+ "learning_rate": 1.9837062236509013e-05,
548
+ "loss": 0.2612,
549
+ "step": 90
550
+ },
551
+ {
552
+ "epoch": 0.17,
553
+ "learning_rate": 1.983141123341168e-05,
554
+ "loss": 0.3401,
555
+ "step": 91
556
+ },
557
+ {
558
+ "epoch": 0.18,
559
+ "learning_rate": 1.9825664732332886e-05,
560
+ "loss": 0.2479,
561
+ "step": 92
562
+ },
563
+ {
564
+ "epoch": 0.18,
565
+ "learning_rate": 1.9819822789091597e-05,
566
+ "loss": 0.2462,
567
+ "step": 93
568
+ },
569
+ {
570
+ "epoch": 0.18,
571
+ "learning_rate": 1.981388546043388e-05,
572
+ "loss": 0.3064,
573
+ "step": 94
574
+ },
575
+ {
576
+ "epoch": 0.18,
577
+ "learning_rate": 1.9807852804032306e-05,
578
+ "loss": 0.3066,
579
+ "step": 95
580
+ },
581
+ {
582
+ "epoch": 0.18,
583
+ "learning_rate": 1.9801724878485438e-05,
584
+ "loss": 0.3402,
585
+ "step": 96
586
+ },
587
+ {
588
+ "epoch": 0.19,
589
+ "learning_rate": 1.979550174331724e-05,
590
+ "loss": 0.2497,
591
+ "step": 97
592
+ },
593
+ {
594
+ "epoch": 0.19,
595
+ "learning_rate": 1.9789183458976485e-05,
596
+ "loss": 0.2513,
597
+ "step": 98
598
+ },
599
+ {
600
+ "epoch": 0.19,
601
+ "learning_rate": 1.97827700868362e-05,
602
+ "loss": 0.3157,
603
+ "step": 99
604
+ },
605
+ {
606
+ "epoch": 0.19,
607
+ "learning_rate": 1.977626168919305e-05,
608
+ "loss": 0.2776,
609
+ "step": 100
610
+ },
611
+ {
612
+ "epoch": 0.19,
613
+ "learning_rate": 1.9769658329266718e-05,
614
+ "loss": 0.3394,
615
+ "step": 101
616
+ },
617
+ {
618
+ "epoch": 0.2,
619
+ "learning_rate": 1.9762960071199334e-05,
620
+ "loss": 0.2657,
621
+ "step": 102
622
+ },
623
+ {
624
+ "epoch": 0.2,
625
+ "learning_rate": 1.9756166980054812e-05,
626
+ "loss": 0.2507,
627
+ "step": 103
628
+ },
629
+ {
630
+ "epoch": 0.2,
631
+ "learning_rate": 1.9749279121818235e-05,
632
+ "loss": 0.3201,
633
+ "step": 104
634
+ },
635
+ {
636
+ "epoch": 0.2,
637
+ "learning_rate": 1.9742296563395218e-05,
638
+ "loss": 0.2509,
639
+ "step": 105
640
+ },
641
+ {
642
+ "epoch": 0.2,
643
+ "learning_rate": 1.9735219372611232e-05,
644
+ "loss": 0.3257,
645
+ "step": 106
646
+ },
647
+ {
648
+ "epoch": 0.21,
649
+ "learning_rate": 1.9728047618210995e-05,
650
+ "loss": 0.3234,
651
+ "step": 107
652
+ },
653
+ {
654
+ "epoch": 0.21,
655
+ "learning_rate": 1.9720781369857747e-05,
656
+ "loss": 0.3496,
657
+ "step": 108
658
+ },
659
+ {
660
+ "epoch": 0.21,
661
+ "learning_rate": 1.9713420698132614e-05,
662
+ "loss": 0.2802,
663
+ "step": 109
664
+ },
665
+ {
666
+ "epoch": 0.21,
667
+ "learning_rate": 1.970596567453391e-05,
668
+ "loss": 0.2662,
669
+ "step": 110
670
+ },
671
+ {
672
+ "epoch": 0.21,
673
+ "learning_rate": 1.9698416371476434e-05,
674
+ "loss": 0.2971,
675
+ "step": 111
676
+ },
677
+ {
678
+ "epoch": 0.22,
679
+ "learning_rate": 1.969077286229078e-05,
680
+ "loss": 0.271,
681
+ "step": 112
682
+ },
683
+ {
684
+ "epoch": 0.22,
685
+ "learning_rate": 1.9683035221222617e-05,
686
+ "loss": 0.3354,
687
+ "step": 113
688
+ },
689
+ {
690
+ "epoch": 0.22,
691
+ "learning_rate": 1.9675203523431964e-05,
692
+ "loss": 0.3134,
693
+ "step": 114
694
+ },
695
+ {
696
+ "epoch": 0.22,
697
+ "learning_rate": 1.9667277844992476e-05,
698
+ "loss": 0.2734,
699
+ "step": 115
700
+ },
701
+ {
702
+ "epoch": 0.22,
703
+ "learning_rate": 1.9659258262890683e-05,
704
+ "loss": 0.273,
705
+ "step": 116
706
+ },
707
+ {
708
+ "epoch": 0.23,
709
+ "learning_rate": 1.9651144855025265e-05,
710
+ "loss": 0.2422,
711
+ "step": 117
712
+ },
713
+ {
714
+ "epoch": 0.23,
715
+ "learning_rate": 1.964293770020628e-05,
716
+ "loss": 0.3048,
717
+ "step": 118
718
+ },
719
+ {
720
+ "epoch": 0.23,
721
+ "learning_rate": 1.9634636878154393e-05,
722
+ "loss": 0.3284,
723
+ "step": 119
724
+ },
725
+ {
726
+ "epoch": 0.23,
727
+ "learning_rate": 1.962624246950012e-05,
728
+ "loss": 0.289,
729
+ "step": 120
730
+ },
731
+ {
732
+ "epoch": 0.23,
733
+ "learning_rate": 1.9617754555783045e-05,
734
+ "loss": 0.2867,
735
+ "step": 121
736
+ },
737
+ {
738
+ "epoch": 0.23,
739
+ "learning_rate": 1.9609173219450998e-05,
740
+ "loss": 0.2646,
741
+ "step": 122
742
+ },
743
+ {
744
+ "epoch": 0.24,
745
+ "learning_rate": 1.960049854385929e-05,
746
+ "loss": 0.3302,
747
+ "step": 123
748
+ },
749
+ {
750
+ "epoch": 0.24,
751
+ "learning_rate": 1.9591730613269878e-05,
752
+ "loss": 0.3099,
753
+ "step": 124
754
+ },
755
+ {
756
+ "epoch": 0.24,
757
+ "learning_rate": 1.9582869512850576e-05,
758
+ "loss": 0.2953,
759
+ "step": 125
760
+ },
761
+ {
762
+ "epoch": 0.24,
763
+ "learning_rate": 1.957391532867418e-05,
764
+ "loss": 0.3058,
765
+ "step": 126
766
+ },
767
+ {
768
+ "epoch": 0.24,
769
+ "learning_rate": 1.956486814771769e-05,
770
+ "loss": 0.2787,
771
+ "step": 127
772
+ },
773
+ {
774
+ "epoch": 0.25,
775
+ "learning_rate": 1.955572805786141e-05,
776
+ "loss": 0.2357,
777
+ "step": 128
778
+ },
779
+ {
780
+ "epoch": 0.25,
781
+ "learning_rate": 1.9546495147888134e-05,
782
+ "loss": 0.3695,
783
+ "step": 129
784
+ },
785
+ {
786
+ "epoch": 0.25,
787
+ "learning_rate": 1.953716950748227e-05,
788
+ "loss": 0.2825,
789
+ "step": 130
790
+ },
791
+ {
792
+ "epoch": 0.25,
793
+ "learning_rate": 1.9527751227228964e-05,
794
+ "loss": 0.279,
795
+ "step": 131
796
+ },
797
+ {
798
+ "epoch": 0.25,
799
+ "learning_rate": 1.9518240398613226e-05,
800
+ "loss": 0.2977,
801
+ "step": 132
802
+ },
803
+ {
804
+ "epoch": 0.26,
805
+ "learning_rate": 1.9508637114019037e-05,
806
+ "loss": 0.3032,
807
+ "step": 133
808
+ },
809
+ {
810
+ "epoch": 0.26,
811
+ "learning_rate": 1.9498941466728462e-05,
812
+ "loss": 0.2568,
813
+ "step": 134
814
+ },
815
+ {
816
+ "epoch": 0.26,
817
+ "learning_rate": 1.9489153550920726e-05,
818
+ "loss": 0.3435,
819
+ "step": 135
820
+ },
821
+ {
822
+ "epoch": 0.26,
823
+ "learning_rate": 1.947927346167132e-05,
824
+ "loss": 0.2819,
825
+ "step": 136
826
+ },
827
+ {
828
+ "epoch": 0.26,
829
+ "learning_rate": 1.946930129495106e-05,
830
+ "loss": 0.2415,
831
+ "step": 137
832
+ },
833
+ {
834
+ "epoch": 0.27,
835
+ "learning_rate": 1.945923714762516e-05,
836
+ "loss": 0.3235,
837
+ "step": 138
838
+ },
839
+ {
840
+ "epoch": 0.27,
841
+ "learning_rate": 1.9449081117452304e-05,
842
+ "loss": 0.2474,
843
+ "step": 139
844
+ },
845
+ {
846
+ "epoch": 0.27,
847
+ "learning_rate": 1.9438833303083677e-05,
848
+ "loss": 0.2894,
849
+ "step": 140
850
+ },
851
+ {
852
+ "epoch": 0.27,
853
+ "learning_rate": 1.9428493804062013e-05,
854
+ "loss": 0.3165,
855
+ "step": 141
856
+ },
857
+ {
858
+ "epoch": 0.27,
859
+ "learning_rate": 1.9418062720820636e-05,
860
+ "loss": 0.2488,
861
+ "step": 142
862
+ },
863
+ {
864
+ "epoch": 0.28,
865
+ "learning_rate": 1.9407540154682473e-05,
866
+ "loss": 0.2624,
867
+ "step": 143
868
+ },
869
+ {
870
+ "epoch": 0.28,
871
+ "learning_rate": 1.9396926207859085e-05,
872
+ "loss": 0.3221,
873
+ "step": 144
874
+ },
875
+ {
876
+ "epoch": 0.28,
877
+ "learning_rate": 1.9386220983449652e-05,
878
+ "loss": 0.2695,
879
+ "step": 145
880
+ },
881
+ {
882
+ "epoch": 0.28,
883
+ "learning_rate": 1.9375424585439994e-05,
884
+ "loss": 0.2303,
885
+ "step": 146
886
+ },
887
+ {
888
+ "epoch": 0.28,
889
+ "learning_rate": 1.9364537118701542e-05,
890
+ "loss": 0.2722,
891
+ "step": 147
892
+ },
893
+ {
894
+ "epoch": 0.28,
895
+ "learning_rate": 1.935355868899034e-05,
896
+ "loss": 0.3543,
897
+ "step": 148
898
+ },
899
+ {
900
+ "epoch": 0.29,
901
+ "learning_rate": 1.9342489402945997e-05,
902
+ "loss": 0.2498,
903
+ "step": 149
904
+ },
905
+ {
906
+ "epoch": 0.29,
907
+ "learning_rate": 1.9331329368090664e-05,
908
+ "loss": 0.2593,
909
+ "step": 150
910
+ },
911
+ {
912
+ "epoch": 0.29,
913
+ "learning_rate": 1.932007869282799e-05,
914
+ "loss": 0.3057,
915
+ "step": 151
916
+ },
917
+ {
918
+ "epoch": 0.29,
919
+ "learning_rate": 1.9308737486442045e-05,
920
+ "loss": 0.2861,
921
+ "step": 152
922
+ },
923
+ {
924
+ "epoch": 0.29,
925
+ "learning_rate": 1.9297305859096305e-05,
926
+ "loss": 0.2738,
927
+ "step": 153
928
+ },
929
+ {
930
+ "epoch": 0.3,
931
+ "learning_rate": 1.9285783921832537e-05,
932
+ "loss": 0.2727,
933
+ "step": 154
934
+ },
935
+ {
936
+ "epoch": 0.3,
937
+ "learning_rate": 1.927417178656975e-05,
938
+ "loss": 0.2787,
939
+ "step": 155
940
+ },
941
+ {
942
+ "epoch": 0.3,
943
+ "learning_rate": 1.926246956610309e-05,
944
+ "loss": 0.3542,
945
+ "step": 156
946
+ },
947
+ {
948
+ "epoch": 0.3,
949
+ "learning_rate": 1.9250677374102752e-05,
950
+ "loss": 0.2699,
951
+ "step": 157
952
+ },
953
+ {
954
+ "epoch": 0.3,
955
+ "learning_rate": 1.9238795325112867e-05,
956
+ "loss": 0.3191,
957
+ "step": 158
958
+ },
959
+ {
960
+ "epoch": 0.31,
961
+ "learning_rate": 1.9226823534550418e-05,
962
+ "loss": 0.3212,
963
+ "step": 159
964
+ },
965
+ {
966
+ "epoch": 0.31,
967
+ "learning_rate": 1.921476211870408e-05,
968
+ "loss": 0.3175,
969
+ "step": 160
970
+ },
971
+ {
972
+ "epoch": 0.31,
973
+ "learning_rate": 1.9202611194733107e-05,
974
+ "loss": 0.3035,
975
+ "step": 161
976
+ },
977
+ {
978
+ "epoch": 0.31,
979
+ "learning_rate": 1.9190370880666206e-05,
980
+ "loss": 0.2824,
981
+ "step": 162
982
+ },
983
+ {
984
+ "epoch": 0.31,
985
+ "learning_rate": 1.9178041295400383e-05,
986
+ "loss": 0.3072,
987
+ "step": 163
988
+ },
989
+ {
990
+ "epoch": 0.32,
991
+ "learning_rate": 1.9165622558699763e-05,
992
+ "loss": 0.222,
993
+ "step": 164
994
+ },
995
+ {
996
+ "epoch": 0.32,
997
+ "learning_rate": 1.9153114791194475e-05,
998
+ "loss": 0.2342,
999
+ "step": 165
1000
+ },
1001
+ {
1002
+ "epoch": 0.32,
1003
+ "learning_rate": 1.9140518114379433e-05,
1004
+ "loss": 0.2743,
1005
+ "step": 166
1006
+ },
1007
+ {
1008
+ "epoch": 0.32,
1009
+ "learning_rate": 1.912783265061319e-05,
1010
+ "loss": 0.3544,
1011
+ "step": 167
1012
+ },
1013
+ {
1014
+ "epoch": 0.32,
1015
+ "learning_rate": 1.9115058523116734e-05,
1016
+ "loss": 0.2701,
1017
+ "step": 168
1018
+ },
1019
+ {
1020
+ "epoch": 0.33,
1021
+ "learning_rate": 1.9102195855972287e-05,
1022
+ "loss": 0.2766,
1023
+ "step": 169
1024
+ },
1025
+ {
1026
+ "epoch": 0.33,
1027
+ "learning_rate": 1.908924477412211e-05,
1028
+ "loss": 0.2867,
1029
+ "step": 170
1030
+ },
1031
+ {
1032
+ "epoch": 0.33,
1033
+ "learning_rate": 1.9076205403367287e-05,
1034
+ "loss": 0.2272,
1035
+ "step": 171
1036
+ },
1037
+ {
1038
+ "epoch": 0.33,
1039
+ "learning_rate": 1.9063077870366504e-05,
1040
+ "loss": 0.2459,
1041
+ "step": 172
1042
+ },
1043
+ {
1044
+ "epoch": 0.33,
1045
+ "learning_rate": 1.90498623026348e-05,
1046
+ "loss": 0.319,
1047
+ "step": 173
1048
+ },
1049
+ {
1050
+ "epoch": 0.33,
1051
+ "learning_rate": 1.903655882854237e-05,
1052
+ "loss": 0.258,
1053
+ "step": 174
1054
+ },
1055
+ {
1056
+ "epoch": 0.34,
1057
+ "learning_rate": 1.9023167577313267e-05,
1058
+ "loss": 0.307,
1059
+ "step": 175
1060
+ },
1061
+ {
1062
+ "epoch": 0.34,
1063
+ "learning_rate": 1.900968867902419e-05,
1064
+ "loss": 0.3004,
1065
+ "step": 176
1066
+ },
1067
+ {
1068
+ "epoch": 0.34,
1069
+ "learning_rate": 1.8996122264603202e-05,
1070
+ "loss": 0.2587,
1071
+ "step": 177
1072
+ },
1073
+ {
1074
+ "epoch": 0.34,
1075
+ "learning_rate": 1.898246846582844e-05,
1076
+ "loss": 0.2956,
1077
+ "step": 178
1078
+ },
1079
+ {
1080
+ "epoch": 0.34,
1081
+ "learning_rate": 1.8968727415326885e-05,
1082
+ "loss": 0.3749,
1083
+ "step": 179
1084
+ },
1085
+ {
1086
+ "epoch": 0.35,
1087
+ "learning_rate": 1.895489924657301e-05,
1088
+ "loss": 0.3463,
1089
+ "step": 180
1090
+ },
1091
+ {
1092
+ "epoch": 0.35,
1093
+ "learning_rate": 1.894098409388754e-05,
1094
+ "loss": 0.3514,
1095
+ "step": 181
1096
+ },
1097
+ {
1098
+ "epoch": 0.35,
1099
+ "learning_rate": 1.8926982092436117e-05,
1100
+ "loss": 0.2839,
1101
+ "step": 182
1102
+ },
1103
+ {
1104
+ "epoch": 0.35,
1105
+ "learning_rate": 1.8912893378227984e-05,
1106
+ "loss": 0.2168,
1107
+ "step": 183
1108
+ },
1109
+ {
1110
+ "epoch": 0.35,
1111
+ "learning_rate": 1.8898718088114688e-05,
1112
+ "loss": 0.2748,
1113
+ "step": 184
1114
+ },
1115
+ {
1116
+ "epoch": 0.36,
1117
+ "learning_rate": 1.8884456359788725e-05,
1118
+ "loss": 0.2885,
1119
+ "step": 185
1120
+ },
1121
+ {
1122
+ "epoch": 0.36,
1123
+ "learning_rate": 1.887010833178222e-05,
1124
+ "loss": 0.2699,
1125
+ "step": 186
1126
+ },
1127
+ {
1128
+ "epoch": 0.36,
1129
+ "learning_rate": 1.8855674143465567e-05,
1130
+ "loss": 0.3133,
1131
+ "step": 187
1132
+ },
1133
+ {
1134
+ "epoch": 0.36,
1135
+ "learning_rate": 1.8841153935046098e-05,
1136
+ "loss": 0.2773,
1137
+ "step": 188
1138
+ },
1139
+ {
1140
+ "epoch": 0.36,
1141
+ "learning_rate": 1.8826547847566692e-05,
1142
+ "loss": 0.3472,
1143
+ "step": 189
1144
+ },
1145
+ {
1146
+ "epoch": 0.37,
1147
+ "learning_rate": 1.8811856022904423e-05,
1148
+ "loss": 0.3342,
1149
+ "step": 190
1150
+ },
1151
+ {
1152
+ "epoch": 0.37,
1153
+ "learning_rate": 1.8797078603769184e-05,
1154
+ "loss": 0.2864,
1155
+ "step": 191
1156
+ },
1157
+ {
1158
+ "epoch": 0.37,
1159
+ "learning_rate": 1.8782215733702286e-05,
1160
+ "loss": 0.3115,
1161
+ "step": 192
1162
+ },
1163
+ {
1164
+ "epoch": 0.37,
1165
+ "learning_rate": 1.876726755707508e-05,
1166
+ "loss": 0.2726,
1167
+ "step": 193
1168
+ },
1169
+ {
1170
+ "epoch": 0.37,
1171
+ "learning_rate": 1.8752234219087538e-05,
1172
+ "loss": 0.3033,
1173
+ "step": 194
1174
+ },
1175
+ {
1176
+ "epoch": 0.38,
1177
+ "learning_rate": 1.8737115865766865e-05,
1178
+ "loss": 0.2785,
1179
+ "step": 195
1180
+ },
1181
+ {
1182
+ "epoch": 0.38,
1183
+ "learning_rate": 1.8721912643966055e-05,
1184
+ "loss": 0.3222,
1185
+ "step": 196
1186
+ },
1187
+ {
1188
+ "epoch": 0.38,
1189
+ "learning_rate": 1.8706624701362485e-05,
1190
+ "loss": 0.2268,
1191
+ "step": 197
1192
+ },
1193
+ {
1194
+ "epoch": 0.38,
1195
+ "learning_rate": 1.8691252186456465e-05,
1196
+ "loss": 0.2259,
1197
+ "step": 198
1198
+ },
1199
+ {
1200
+ "epoch": 0.38,
1201
+ "learning_rate": 1.8675795248569816e-05,
1202
+ "loss": 0.2777,
1203
+ "step": 199
1204
+ },
1205
+ {
1206
+ "epoch": 0.38,
1207
+ "learning_rate": 1.866025403784439e-05,
1208
+ "loss": 0.2765,
1209
+ "step": 200
1210
+ },
1211
+ {
1212
+ "epoch": 0.39,
1213
+ "learning_rate": 1.8644628705240636e-05,
1214
+ "loss": 0.3088,
1215
+ "step": 201
1216
+ },
1217
+ {
1218
+ "epoch": 0.39,
1219
+ "learning_rate": 1.862891940253613e-05,
1220
+ "loss": 0.2766,
1221
+ "step": 202
1222
+ },
1223
+ {
1224
+ "epoch": 0.39,
1225
+ "learning_rate": 1.8613126282324092e-05,
1226
+ "loss": 0.2699,
1227
+ "step": 203
1228
+ },
1229
+ {
1230
+ "epoch": 0.39,
1231
+ "learning_rate": 1.8597249498011906e-05,
1232
+ "loss": 0.2973,
1233
+ "step": 204
1234
+ },
1235
+ {
1236
+ "epoch": 0.39,
1237
+ "learning_rate": 1.858128920381963e-05,
1238
+ "loss": 0.3202,
1239
+ "step": 205
1240
+ },
1241
+ {
1242
+ "epoch": 0.4,
1243
+ "learning_rate": 1.8565245554778516e-05,
1244
+ "loss": 0.2645,
1245
+ "step": 206
1246
+ },
1247
+ {
1248
+ "epoch": 0.4,
1249
+ "learning_rate": 1.854911870672947e-05,
1250
+ "loss": 0.2514,
1251
+ "step": 207
1252
+ },
1253
+ {
1254
+ "epoch": 0.4,
1255
+ "learning_rate": 1.8532908816321557e-05,
1256
+ "loss": 0.2983,
1257
+ "step": 208
1258
+ },
1259
+ {
1260
+ "epoch": 0.4,
1261
+ "learning_rate": 1.8516616041010495e-05,
1262
+ "loss": 0.2516,
1263
+ "step": 209
1264
+ },
1265
+ {
1266
+ "epoch": 0.4,
1267
+ "learning_rate": 1.8500240539057093e-05,
1268
+ "loss": 0.2891,
1269
+ "step": 210
1270
+ },
1271
+ {
1272
+ "epoch": 0.41,
1273
+ "learning_rate": 1.848378246952574e-05,
1274
+ "loss": 0.2445,
1275
+ "step": 211
1276
+ },
1277
+ {
1278
+ "epoch": 0.41,
1279
+ "learning_rate": 1.8467241992282842e-05,
1280
+ "loss": 0.2506,
1281
+ "step": 212
1282
+ },
1283
+ {
1284
+ "epoch": 0.41,
1285
+ "learning_rate": 1.8450619267995283e-05,
1286
+ "loss": 0.317,
1287
+ "step": 213
1288
+ },
1289
+ {
1290
+ "epoch": 0.41,
1291
+ "learning_rate": 1.843391445812886e-05,
1292
+ "loss": 0.2946,
1293
+ "step": 214
1294
+ },
1295
+ {
1296
+ "epoch": 0.41,
1297
+ "learning_rate": 1.84171277249467e-05,
1298
+ "loss": 0.3034,
1299
+ "step": 215
1300
+ },
1301
+ {
1302
+ "epoch": 0.42,
1303
+ "learning_rate": 1.8400259231507716e-05,
1304
+ "loss": 0.3061,
1305
+ "step": 216
1306
+ },
1307
+ {
1308
+ "epoch": 0.42,
1309
+ "learning_rate": 1.8383309141664992e-05,
1310
+ "loss": 0.2333,
1311
+ "step": 217
1312
+ },
1313
+ {
1314
+ "epoch": 0.42,
1315
+ "learning_rate": 1.83662776200642e-05,
1316
+ "loss": 0.318,
1317
+ "step": 218
1318
+ },
1319
+ {
1320
+ "epoch": 0.42,
1321
+ "learning_rate": 1.8349164832142015e-05,
1322
+ "loss": 0.2176,
1323
+ "step": 219
1324
+ },
1325
+ {
1326
+ "epoch": 0.42,
1327
+ "learning_rate": 1.833197094412449e-05,
1328
+ "loss": 0.2698,
1329
+ "step": 220
1330
+ },
1331
+ {
1332
+ "epoch": 0.42,
1333
+ "learning_rate": 1.8314696123025456e-05,
1334
+ "loss": 0.2831,
1335
+ "step": 221
1336
+ },
1337
+ {
1338
+ "epoch": 0.43,
1339
+ "learning_rate": 1.8297340536644877e-05,
1340
+ "loss": 0.2773,
1341
+ "step": 222
1342
+ },
1343
+ {
1344
+ "epoch": 0.43,
1345
+ "learning_rate": 1.827990435356725e-05,
1346
+ "loss": 0.2892,
1347
+ "step": 223
1348
+ },
1349
+ {
1350
+ "epoch": 0.43,
1351
+ "learning_rate": 1.826238774315995e-05,
1352
+ "loss": 0.3196,
1353
+ "step": 224
1354
+ },
1355
+ {
1356
+ "epoch": 0.43,
1357
+ "learning_rate": 1.8244790875571582e-05,
1358
+ "loss": 0.2989,
1359
+ "step": 225
1360
+ },
1361
+ {
1362
+ "epoch": 0.43,
1363
+ "learning_rate": 1.8227113921730336e-05,
1364
+ "loss": 0.2624,
1365
+ "step": 226
1366
+ },
1367
+ {
1368
+ "epoch": 0.44,
1369
+ "learning_rate": 1.8209357053342325e-05,
1370
+ "loss": 0.2856,
1371
+ "step": 227
1372
+ },
1373
+ {
1374
+ "epoch": 0.44,
1375
+ "learning_rate": 1.819152044288992e-05,
1376
+ "loss": 0.3125,
1377
+ "step": 228
1378
+ },
1379
+ {
1380
+ "epoch": 0.44,
1381
+ "learning_rate": 1.8173604263630066e-05,
1382
+ "loss": 0.2491,
1383
+ "step": 229
1384
+ },
1385
+ {
1386
+ "epoch": 0.44,
1387
+ "learning_rate": 1.8155608689592604e-05,
1388
+ "loss": 0.2719,
1389
+ "step": 230
1390
+ },
1391
+ {
1392
+ "epoch": 0.44,
1393
+ "learning_rate": 1.8137533895578585e-05,
1394
+ "loss": 0.2805,
1395
+ "step": 231
1396
+ },
1397
+ {
1398
+ "epoch": 0.45,
1399
+ "learning_rate": 1.811938005715857e-05,
1400
+ "loss": 0.2564,
1401
+ "step": 232
1402
+ },
1403
+ {
1404
+ "epoch": 0.45,
1405
+ "learning_rate": 1.8101147350670905e-05,
1406
+ "loss": 0.2422,
1407
+ "step": 233
1408
+ },
1409
+ {
1410
+ "epoch": 0.45,
1411
+ "learning_rate": 1.8082835953220055e-05,
1412
+ "loss": 0.259,
1413
+ "step": 234
1414
+ },
1415
+ {
1416
+ "epoch": 0.45,
1417
+ "learning_rate": 1.806444604267483e-05,
1418
+ "loss": 0.2272,
1419
+ "step": 235
1420
+ },
1421
+ {
1422
+ "epoch": 0.45,
1423
+ "learning_rate": 1.8045977797666685e-05,
1424
+ "loss": 0.266,
1425
+ "step": 236
1426
+ },
1427
+ {
1428
+ "epoch": 0.46,
1429
+ "learning_rate": 1.8027431397587993e-05,
1430
+ "loss": 0.2909,
1431
+ "step": 237
1432
+ },
1433
+ {
1434
+ "epoch": 0.46,
1435
+ "learning_rate": 1.8008807022590283e-05,
1436
+ "loss": 0.2731,
1437
+ "step": 238
1438
+ },
1439
+ {
1440
+ "epoch": 0.46,
1441
+ "learning_rate": 1.7990104853582494e-05,
1442
+ "loss": 0.2517,
1443
+ "step": 239
1444
+ },
1445
+ {
1446
+ "epoch": 0.46,
1447
+ "learning_rate": 1.7971325072229227e-05,
1448
+ "loss": 0.2572,
1449
+ "step": 240
1450
+ },
1451
+ {
1452
+ "epoch": 0.46,
1453
+ "learning_rate": 1.7952467860948975e-05,
1454
+ "loss": 0.2328,
1455
+ "step": 241
1456
+ },
1457
+ {
1458
+ "epoch": 0.47,
1459
+ "learning_rate": 1.7933533402912354e-05,
1460
+ "loss": 0.241,
1461
+ "step": 242
1462
+ },
1463
+ {
1464
+ "epoch": 0.47,
1465
+ "learning_rate": 1.791452188204031e-05,
1466
+ "loss": 0.3629,
1467
+ "step": 243
1468
+ },
1469
+ {
1470
+ "epoch": 0.47,
1471
+ "learning_rate": 1.7895433483002356e-05,
1472
+ "loss": 0.3116,
1473
+ "step": 244
1474
+ },
1475
+ {
1476
+ "epoch": 0.47,
1477
+ "learning_rate": 1.7876268391214756e-05,
1478
+ "loss": 0.2936,
1479
+ "step": 245
1480
+ },
1481
+ {
1482
+ "epoch": 0.47,
1483
+ "learning_rate": 1.785702679283874e-05,
1484
+ "loss": 0.2804,
1485
+ "step": 246
1486
+ },
1487
+ {
1488
+ "epoch": 0.47,
1489
+ "learning_rate": 1.7837708874778683e-05,
1490
+ "loss": 0.2711,
1491
+ "step": 247
1492
+ },
1493
+ {
1494
+ "epoch": 0.48,
1495
+ "learning_rate": 1.78183148246803e-05,
1496
+ "loss": 0.3023,
1497
+ "step": 248
1498
+ },
1499
+ {
1500
+ "epoch": 0.48,
1501
+ "learning_rate": 1.7798844830928818e-05,
1502
+ "loss": 0.251,
1503
+ "step": 249
1504
+ },
1505
+ {
1506
+ "epoch": 0.48,
1507
+ "learning_rate": 1.777929908264715e-05,
1508
+ "loss": 0.2831,
1509
+ "step": 250
1510
+ },
1511
+ {
1512
+ "epoch": 0.48,
1513
+ "learning_rate": 1.775967776969405e-05,
1514
+ "loss": 0.2369,
1515
+ "step": 251
1516
+ },
1517
+ {
1518
+ "epoch": 0.48,
1519
+ "learning_rate": 1.7739981082662275e-05,
1520
+ "loss": 0.2613,
1521
+ "step": 252
1522
+ },
1523
+ {
1524
+ "epoch": 0.49,
1525
+ "learning_rate": 1.772020921287674e-05,
1526
+ "loss": 0.2944,
1527
+ "step": 253
1528
+ },
1529
+ {
1530
+ "epoch": 0.49,
1531
+ "learning_rate": 1.7700362352392632e-05,
1532
+ "loss": 0.2577,
1533
+ "step": 254
1534
+ },
1535
+ {
1536
+ "epoch": 0.49,
1537
+ "learning_rate": 1.7680440693993586e-05,
1538
+ "loss": 0.277,
1539
+ "step": 255
1540
+ },
1541
+ {
1542
+ "epoch": 0.49,
1543
+ "learning_rate": 1.766044443118978e-05,
1544
+ "loss": 0.256,
1545
+ "step": 256
1546
+ },
1547
+ {
1548
+ "epoch": 0.49,
1549
+ "learning_rate": 1.7640373758216075e-05,
1550
+ "loss": 0.3175,
1551
+ "step": 257
1552
+ },
1553
+ {
1554
+ "epoch": 0.5,
1555
+ "learning_rate": 1.762022887003011e-05,
1556
+ "loss": 0.3031,
1557
+ "step": 258
1558
+ },
1559
+ {
1560
+ "epoch": 0.5,
1561
+ "learning_rate": 1.7600009962310417e-05,
1562
+ "loss": 0.262,
1563
+ "step": 259
1564
+ },
1565
+ {
1566
+ "epoch": 0.5,
1567
+ "learning_rate": 1.757971723145453e-05,
1568
+ "loss": 0.3123,
1569
+ "step": 260
1570
+ },
1571
+ {
1572
+ "epoch": 0.5,
1573
+ "learning_rate": 1.7559350874577066e-05,
1574
+ "loss": 0.2985,
1575
+ "step": 261
1576
+ },
1577
+ {
1578
+ "epoch": 0.5,
1579
+ "learning_rate": 1.75389110895078e-05,
1580
+ "loss": 0.2605,
1581
+ "step": 262
1582
+ },
1583
+ {
1584
+ "epoch": 0.51,
1585
+ "learning_rate": 1.7518398074789776e-05,
1586
+ "loss": 0.3238,
1587
+ "step": 263
1588
+ },
1589
+ {
1590
+ "epoch": 0.51,
1591
+ "learning_rate": 1.7497812029677344e-05,
1592
+ "loss": 0.3072,
1593
+ "step": 264
1594
+ },
1595
+ {
1596
+ "epoch": 0.51,
1597
+ "learning_rate": 1.7477153154134244e-05,
1598
+ "loss": 0.2944,
1599
+ "step": 265
1600
+ },
1601
+ {
1602
+ "epoch": 0.51,
1603
+ "learning_rate": 1.7456421648831658e-05,
1604
+ "loss": 0.2557,
1605
+ "step": 266
1606
+ },
1607
+ {
1608
+ "epoch": 0.51,
1609
+ "learning_rate": 1.743561771514626e-05,
1610
+ "loss": 0.2991,
1611
+ "step": 267
1612
+ },
1613
+ {
1614
+ "epoch": 0.52,
1615
+ "learning_rate": 1.741474155515827e-05,
1616
+ "loss": 0.3116,
1617
+ "step": 268
1618
+ },
1619
+ {
1620
+ "epoch": 0.52,
1621
+ "learning_rate": 1.739379337164946e-05,
1622
+ "loss": 0.2911,
1623
+ "step": 269
1624
+ },
1625
+ {
1626
+ "epoch": 0.52,
1627
+ "learning_rate": 1.737277336810124e-05,
1628
+ "loss": 0.3144,
1629
+ "step": 270
1630
+ },
1631
+ {
1632
+ "epoch": 0.52,
1633
+ "learning_rate": 1.7351681748692622e-05,
1634
+ "loss": 0.3243,
1635
+ "step": 271
1636
+ },
1637
+ {
1638
+ "epoch": 0.52,
1639
+ "learning_rate": 1.7330518718298263e-05,
1640
+ "loss": 0.2743,
1641
+ "step": 272
1642
+ },
1643
+ {
1644
+ "epoch": 0.53,
1645
+ "learning_rate": 1.7309284482486494e-05,
1646
+ "loss": 0.2532,
1647
+ "step": 273
1648
+ },
1649
+ {
1650
+ "epoch": 0.53,
1651
+ "learning_rate": 1.7287979247517285e-05,
1652
+ "loss": 0.2777,
1653
+ "step": 274
1654
+ },
1655
+ {
1656
+ "epoch": 0.53,
1657
+ "learning_rate": 1.7266603220340273e-05,
1658
+ "loss": 0.3015,
1659
+ "step": 275
1660
+ },
1661
+ {
1662
+ "epoch": 0.53,
1663
+ "learning_rate": 1.7245156608592727e-05,
1664
+ "loss": 0.3001,
1665
+ "step": 276
1666
+ },
1667
+ {
1668
+ "epoch": 0.53,
1669
+ "learning_rate": 1.7223639620597556e-05,
1670
+ "loss": 0.3408,
1671
+ "step": 277
1672
+ },
1673
+ {
1674
+ "epoch": 0.53,
1675
+ "learning_rate": 1.7202052465361268e-05,
1676
+ "loss": 0.2568,
1677
+ "step": 278
1678
+ },
1679
+ {
1680
+ "epoch": 0.54,
1681
+ "learning_rate": 1.718039535257194e-05,
1682
+ "loss": 0.3001,
1683
+ "step": 279
1684
+ },
1685
+ {
1686
+ "epoch": 0.54,
1687
+ "learning_rate": 1.7158668492597186e-05,
1688
+ "loss": 0.231,
1689
+ "step": 280
1690
+ },
1691
+ {
1692
+ "epoch": 0.54,
1693
+ "learning_rate": 1.7136872096482123e-05,
1694
+ "loss": 0.2371,
1695
+ "step": 281
1696
+ },
1697
+ {
1698
+ "epoch": 0.54,
1699
+ "learning_rate": 1.7115006375947304e-05,
1700
+ "loss": 0.3373,
1701
+ "step": 282
1702
+ },
1703
+ {
1704
+ "epoch": 0.54,
1705
+ "learning_rate": 1.7093071543386667e-05,
1706
+ "loss": 0.3651,
1707
+ "step": 283
1708
+ },
1709
+ {
1710
+ "epoch": 0.55,
1711
+ "learning_rate": 1.7071067811865477e-05,
1712
+ "loss": 0.307,
1713
+ "step": 284
1714
+ },
1715
+ {
1716
+ "epoch": 0.55,
1717
+ "learning_rate": 1.7048995395118253e-05,
1718
+ "loss": 0.3449,
1719
+ "step": 285
1720
+ },
1721
+ {
1722
+ "epoch": 0.55,
1723
+ "learning_rate": 1.7026854507546694e-05,
1724
+ "loss": 0.3196,
1725
+ "step": 286
1726
+ },
1727
+ {
1728
+ "epoch": 0.55,
1729
+ "learning_rate": 1.7004645364217584e-05,
1730
+ "loss": 0.272,
1731
+ "step": 287
1732
+ },
1733
+ {
1734
+ "epoch": 0.55,
1735
+ "learning_rate": 1.698236818086073e-05,
1736
+ "loss": 0.2899,
1737
+ "step": 288
1738
+ },
1739
+ {
1740
+ "epoch": 0.56,
1741
+ "learning_rate": 1.6960023173866834e-05,
1742
+ "loss": 0.2488,
1743
+ "step": 289
1744
+ },
1745
+ {
1746
+ "epoch": 0.56,
1747
+ "learning_rate": 1.693761056028542e-05,
1748
+ "loss": 0.2224,
1749
+ "step": 290
1750
+ },
1751
+ {
1752
+ "epoch": 0.56,
1753
+ "learning_rate": 1.6915130557822698e-05,
1754
+ "loss": 0.3136,
1755
+ "step": 291
1756
+ },
1757
+ {
1758
+ "epoch": 0.56,
1759
+ "learning_rate": 1.689258338483947e-05,
1760
+ "loss": 0.2671,
1761
+ "step": 292
1762
+ },
1763
+ {
1764
+ "epoch": 0.56,
1765
+ "learning_rate": 1.686996926034902e-05,
1766
+ "loss": 0.2923,
1767
+ "step": 293
1768
+ },
1769
+ {
1770
+ "epoch": 0.57,
1771
+ "learning_rate": 1.6847288404014937e-05,
1772
+ "loss": 0.3313,
1773
+ "step": 294
1774
+ },
1775
+ {
1776
+ "epoch": 0.57,
1777
+ "learning_rate": 1.682454103614904e-05,
1778
+ "loss": 0.2638,
1779
+ "step": 295
1780
+ },
1781
+ {
1782
+ "epoch": 0.57,
1783
+ "learning_rate": 1.6801727377709195e-05,
1784
+ "loss": 0.3061,
1785
+ "step": 296
1786
+ },
1787
+ {
1788
+ "epoch": 0.57,
1789
+ "learning_rate": 1.67788476502972e-05,
1790
+ "loss": 0.2966,
1791
+ "step": 297
1792
+ },
1793
+ {
1794
+ "epoch": 0.57,
1795
+ "learning_rate": 1.6755902076156606e-05,
1796
+ "loss": 0.251,
1797
+ "step": 298
1798
+ },
1799
+ {
1800
+ "epoch": 0.57,
1801
+ "learning_rate": 1.6732890878170573e-05,
1802
+ "loss": 0.2769,
1803
+ "step": 299
1804
+ },
1805
+ {
1806
+ "epoch": 0.58,
1807
+ "learning_rate": 1.67098142798597e-05,
1808
+ "loss": 0.2796,
1809
+ "step": 300
1810
+ },
1811
+ {
1812
+ "epoch": 0.58,
1813
+ "learning_rate": 1.668667250537987e-05,
1814
+ "loss": 0.3202,
1815
+ "step": 301
1816
+ },
1817
+ {
1818
+ "epoch": 0.58,
1819
+ "learning_rate": 1.6663465779520042e-05,
1820
+ "loss": 0.2775,
1821
+ "step": 302
1822
+ },
1823
+ {
1824
+ "epoch": 0.58,
1825
+ "learning_rate": 1.6640194327700087e-05,
1826
+ "loss": 0.3842,
1827
+ "step": 303
1828
+ },
1829
+ {
1830
+ "epoch": 0.58,
1831
+ "learning_rate": 1.6616858375968596e-05,
1832
+ "loss": 0.2959,
1833
+ "step": 304
1834
+ },
1835
+ {
1836
+ "epoch": 0.59,
1837
+ "learning_rate": 1.659345815100069e-05,
1838
+ "loss": 0.2942,
1839
+ "step": 305
1840
+ },
1841
+ {
1842
+ "epoch": 0.59,
1843
+ "learning_rate": 1.6569993880095807e-05,
1844
+ "loss": 0.2933,
1845
+ "step": 306
1846
+ },
1847
+ {
1848
+ "epoch": 0.59,
1849
+ "learning_rate": 1.6546465791175498e-05,
1850
+ "loss": 0.2575,
1851
+ "step": 307
1852
+ },
1853
+ {
1854
+ "epoch": 0.59,
1855
+ "learning_rate": 1.6522874112781213e-05,
1856
+ "loss": 0.2374,
1857
+ "step": 308
1858
+ },
1859
+ {
1860
+ "epoch": 0.59,
1861
+ "learning_rate": 1.6499219074072087e-05,
1862
+ "loss": 0.297,
1863
+ "step": 309
1864
+ },
1865
+ {
1866
+ "epoch": 0.6,
1867
+ "learning_rate": 1.6475500904822707e-05,
1868
+ "loss": 0.2665,
1869
+ "step": 310
1870
+ },
1871
+ {
1872
+ "epoch": 0.6,
1873
+ "learning_rate": 1.645171983542088e-05,
1874
+ "loss": 0.3204,
1875
+ "step": 311
1876
+ },
1877
+ {
1878
+ "epoch": 0.6,
1879
+ "learning_rate": 1.6427876096865394e-05,
1880
+ "loss": 0.3059,
1881
+ "step": 312
1882
+ },
1883
+ {
1884
+ "epoch": 0.6,
1885
+ "learning_rate": 1.640396992076379e-05,
1886
+ "loss": 0.2196,
1887
+ "step": 313
1888
+ },
1889
+ {
1890
+ "epoch": 0.6,
1891
+ "learning_rate": 1.6380001539330088e-05,
1892
+ "loss": 0.2718,
1893
+ "step": 314
1894
+ },
1895
+ {
1896
+ "epoch": 0.61,
1897
+ "learning_rate": 1.6355971185382547e-05,
1898
+ "loss": 0.2303,
1899
+ "step": 315
1900
+ },
1901
+ {
1902
+ "epoch": 0.61,
1903
+ "learning_rate": 1.6331879092341402e-05,
1904
+ "loss": 0.3033,
1905
+ "step": 316
1906
+ },
1907
+ {
1908
+ "epoch": 0.61,
1909
+ "learning_rate": 1.6307725494226586e-05,
1910
+ "loss": 0.2913,
1911
+ "step": 317
1912
+ },
1913
+ {
1914
+ "epoch": 0.61,
1915
+ "learning_rate": 1.6283510625655474e-05,
1916
+ "loss": 0.2674,
1917
+ "step": 318
1918
+ },
1919
+ {
1920
+ "epoch": 0.61,
1921
+ "learning_rate": 1.6259234721840595e-05,
1922
+ "loss": 0.2809,
1923
+ "step": 319
1924
+ },
1925
+ {
1926
+ "epoch": 0.62,
1927
+ "learning_rate": 1.6234898018587336e-05,
1928
+ "loss": 0.2553,
1929
+ "step": 320
1930
+ },
1931
+ {
1932
+ "epoch": 0.62,
1933
+ "learning_rate": 1.6210500752291682e-05,
1934
+ "loss": 0.3065,
1935
+ "step": 321
1936
+ },
1937
+ {
1938
+ "epoch": 0.62,
1939
+ "learning_rate": 1.6186043159937884e-05,
1940
+ "loss": 0.3345,
1941
+ "step": 322
1942
+ },
1943
+ {
1944
+ "epoch": 0.62,
1945
+ "learning_rate": 1.616152547909618e-05,
1946
+ "loss": 0.3098,
1947
+ "step": 323
1948
+ },
1949
+ {
1950
+ "epoch": 0.62,
1951
+ "learning_rate": 1.6136947947920477e-05,
1952
+ "loss": 0.2784,
1953
+ "step": 324
1954
+ },
1955
+ {
1956
+ "epoch": 0.62,
1957
+ "learning_rate": 1.611231080514605e-05,
1958
+ "loss": 0.2735,
1959
+ "step": 325
1960
+ },
1961
+ {
1962
+ "epoch": 0.63,
1963
+ "learning_rate": 1.608761429008721e-05,
1964
+ "loss": 0.2737,
1965
+ "step": 326
1966
+ },
1967
+ {
1968
+ "epoch": 0.63,
1969
+ "learning_rate": 1.606285864263498e-05,
1970
+ "loss": 0.2493,
1971
+ "step": 327
1972
+ },
1973
+ {
1974
+ "epoch": 0.63,
1975
+ "learning_rate": 1.6038044103254775e-05,
1976
+ "loss": 0.3252,
1977
+ "step": 328
1978
+ },
1979
+ {
1980
+ "epoch": 0.63,
1981
+ "learning_rate": 1.601317091298406e-05,
1982
+ "loss": 0.3153,
1983
+ "step": 329
1984
+ },
1985
+ {
1986
+ "epoch": 0.63,
1987
+ "learning_rate": 1.5988239313430004e-05,
1988
+ "loss": 0.3649,
1989
+ "step": 330
1990
+ },
1991
+ {
1992
+ "epoch": 0.64,
1993
+ "learning_rate": 1.5963249546767144e-05,
1994
+ "loss": 0.2791,
1995
+ "step": 331
1996
+ },
1997
+ {
1998
+ "epoch": 0.64,
1999
+ "learning_rate": 1.5938201855735017e-05,
2000
+ "loss": 0.2858,
2001
+ "step": 332
2002
+ },
2003
+ {
2004
+ "epoch": 0.64,
2005
+ "learning_rate": 1.5913096483635827e-05,
2006
+ "loss": 0.2663,
2007
+ "step": 333
2008
+ },
2009
+ {
2010
+ "epoch": 0.64,
2011
+ "learning_rate": 1.5887933674332048e-05,
2012
+ "loss": 0.3026,
2013
+ "step": 334
2014
+ },
2015
+ {
2016
+ "epoch": 0.64,
2017
+ "learning_rate": 1.5862713672244092e-05,
2018
+ "loss": 0.2936,
2019
+ "step": 335
2020
+ },
2021
+ {
2022
+ "epoch": 0.65,
2023
+ "learning_rate": 1.5837436722347902e-05,
2024
+ "loss": 0.3158,
2025
+ "step": 336
2026
+ },
2027
+ {
2028
+ "epoch": 0.65,
2029
+ "learning_rate": 1.5812103070172592e-05,
2030
+ "loss": 0.2877,
2031
+ "step": 337
2032
+ },
2033
+ {
2034
+ "epoch": 0.65,
2035
+ "learning_rate": 1.578671296179806e-05,
2036
+ "loss": 0.322,
2037
+ "step": 338
2038
+ },
2039
+ {
2040
+ "epoch": 0.65,
2041
+ "learning_rate": 1.5761266643852587e-05,
2042
+ "loss": 0.2558,
2043
+ "step": 339
2044
+ },
2045
+ {
2046
+ "epoch": 0.65,
2047
+ "learning_rate": 1.573576436351046e-05,
2048
+ "loss": 0.2491,
2049
+ "step": 340
2050
+ },
2051
+ {
2052
+ "epoch": 0.66,
2053
+ "learning_rate": 1.5710206368489555e-05,
2054
+ "loss": 0.2728,
2055
+ "step": 341
2056
+ },
2057
+ {
2058
+ "epoch": 0.66,
2059
+ "learning_rate": 1.5684592907048925e-05,
2060
+ "loss": 0.3044,
2061
+ "step": 342
2062
+ },
2063
+ {
2064
+ "epoch": 0.66,
2065
+ "learning_rate": 1.5658924227986415e-05,
2066
+ "loss": 0.2578,
2067
+ "step": 343
2068
+ },
2069
+ {
2070
+ "epoch": 0.66,
2071
+ "learning_rate": 1.563320058063622e-05,
2072
+ "loss": 0.2324,
2073
+ "step": 344
2074
+ },
2075
+ {
2076
+ "epoch": 0.66,
2077
+ "learning_rate": 1.560742221486648e-05,
2078
+ "loss": 0.2622,
2079
+ "step": 345
2080
+ },
2081
+ {
2082
+ "epoch": 0.67,
2083
+ "learning_rate": 1.5581589381076843e-05,
2084
+ "loss": 0.2837,
2085
+ "step": 346
2086
+ },
2087
+ {
2088
+ "epoch": 0.67,
2089
+ "learning_rate": 1.5555702330196024e-05,
2090
+ "loss": 0.2865,
2091
+ "step": 347
2092
+ },
2093
+ {
2094
+ "epoch": 0.67,
2095
+ "learning_rate": 1.5529761313679396e-05,
2096
+ "loss": 0.242,
2097
+ "step": 348
2098
+ },
2099
+ {
2100
+ "epoch": 0.67,
2101
+ "learning_rate": 1.5503766583506522e-05,
2102
+ "loss": 0.3048,
2103
+ "step": 349
2104
+ },
2105
+ {
2106
+ "epoch": 0.67,
2107
+ "learning_rate": 1.5477718392178716e-05,
2108
+ "loss": 0.2945,
2109
+ "step": 350
2110
+ },
2111
+ {
2112
+ "epoch": 0.68,
2113
+ "learning_rate": 1.545161699271659e-05,
2114
+ "loss": 0.2965,
2115
+ "step": 351
2116
+ },
2117
+ {
2118
+ "epoch": 0.68,
2119
+ "learning_rate": 1.5425462638657597e-05,
2120
+ "loss": 0.2797,
2121
+ "step": 352
2122
+ },
2123
+ {
2124
+ "epoch": 0.68,
2125
+ "learning_rate": 1.5399255584053568e-05,
2126
+ "loss": 0.2599,
2127
+ "step": 353
2128
+ },
2129
+ {
2130
+ "epoch": 0.68,
2131
+ "learning_rate": 1.5372996083468242e-05,
2132
+ "loss": 0.2638,
2133
+ "step": 354
2134
+ },
2135
+ {
2136
+ "epoch": 0.68,
2137
+ "learning_rate": 1.5346684391974792e-05,
2138
+ "loss": 0.2961,
2139
+ "step": 355
2140
+ },
2141
+ {
2142
+ "epoch": 0.68,
2143
+ "learning_rate": 1.5320320765153367e-05,
2144
+ "loss": 0.3614,
2145
+ "step": 356
2146
+ },
2147
+ {
2148
+ "epoch": 0.69,
2149
+ "learning_rate": 1.529390545908857e-05,
2150
+ "loss": 0.2709,
2151
+ "step": 357
2152
+ },
2153
+ {
2154
+ "epoch": 0.69,
2155
+ "learning_rate": 1.526743873036701e-05,
2156
+ "loss": 0.3059,
2157
+ "step": 358
2158
+ },
2159
+ {
2160
+ "epoch": 0.69,
2161
+ "learning_rate": 1.5240920836074777e-05,
2162
+ "loss": 0.2747,
2163
+ "step": 359
2164
+ },
2165
+ {
2166
+ "epoch": 0.69,
2167
+ "learning_rate": 1.5214352033794981e-05,
2168
+ "loss": 0.2479,
2169
+ "step": 360
2170
+ },
2171
+ {
2172
+ "epoch": 0.69,
2173
+ "learning_rate": 1.5187732581605217e-05,
2174
+ "loss": 0.2541,
2175
+ "step": 361
2176
+ },
2177
+ {
2178
+ "epoch": 0.7,
2179
+ "learning_rate": 1.5161062738075068e-05,
2180
+ "loss": 0.2887,
2181
+ "step": 362
2182
+ },
2183
+ {
2184
+ "epoch": 0.7,
2185
+ "learning_rate": 1.5134342762263606e-05,
2186
+ "loss": 0.32,
2187
+ "step": 363
2188
+ },
2189
+ {
2190
+ "epoch": 0.7,
2191
+ "learning_rate": 1.5107572913716859e-05,
2192
+ "loss": 0.2669,
2193
+ "step": 364
2194
+ },
2195
+ {
2196
+ "epoch": 0.7,
2197
+ "learning_rate": 1.5080753452465296e-05,
2198
+ "loss": 0.2927,
2199
+ "step": 365
2200
+ },
2201
+ {
2202
+ "epoch": 0.7,
2203
+ "learning_rate": 1.505388463902131e-05,
2204
+ "loss": 0.352,
2205
+ "step": 366
2206
+ },
2207
+ {
2208
+ "epoch": 0.71,
2209
+ "learning_rate": 1.502696673437667e-05,
2210
+ "loss": 0.243,
2211
+ "step": 367
2212
+ },
2213
+ {
2214
+ "epoch": 0.71,
2215
+ "learning_rate": 1.5000000000000002e-05,
2216
+ "loss": 0.3295,
2217
+ "step": 368
2218
+ },
2219
+ {
2220
+ "epoch": 0.71,
2221
+ "learning_rate": 1.4972984697834238e-05,
2222
+ "loss": 0.2579,
2223
+ "step": 369
2224
+ },
2225
+ {
2226
+ "epoch": 0.71,
2227
+ "learning_rate": 1.4945921090294076e-05,
2228
+ "loss": 0.2393,
2229
+ "step": 370
2230
+ },
2231
+ {
2232
+ "epoch": 0.71,
2233
+ "learning_rate": 1.4918809440263435e-05,
2234
+ "loss": 0.2985,
2235
+ "step": 371
2236
+ },
2237
+ {
2238
+ "epoch": 0.72,
2239
+ "learning_rate": 1.4891650011092896e-05,
2240
+ "loss": 0.3277,
2241
+ "step": 372
2242
+ },
2243
+ {
2244
+ "epoch": 0.72,
2245
+ "learning_rate": 1.486444306659714e-05,
2246
+ "loss": 0.2591,
2247
+ "step": 373
2248
+ },
2249
+ {
2250
+ "epoch": 0.72,
2251
+ "learning_rate": 1.4837188871052399e-05,
2252
+ "loss": 0.2722,
2253
+ "step": 374
2254
+ },
2255
+ {
2256
+ "epoch": 0.72,
2257
+ "learning_rate": 1.4809887689193878e-05,
2258
+ "loss": 0.2754,
2259
+ "step": 375
2260
+ },
2261
+ {
2262
+ "epoch": 0.72,
2263
+ "learning_rate": 1.4782539786213184e-05,
2264
+ "loss": 0.2422,
2265
+ "step": 376
2266
+ },
2267
+ {
2268
+ "epoch": 0.72,
2269
+ "learning_rate": 1.4755145427755755e-05,
2270
+ "loss": 0.2508,
2271
+ "step": 377
2272
+ },
2273
+ {
2274
+ "epoch": 0.73,
2275
+ "learning_rate": 1.4727704879918272e-05,
2276
+ "loss": 0.2739,
2277
+ "step": 378
2278
+ },
2279
+ {
2280
+ "epoch": 0.73,
2281
+ "learning_rate": 1.4700218409246087e-05,
2282
+ "loss": 0.3071,
2283
+ "step": 379
2284
+ },
2285
+ {
2286
+ "epoch": 0.73,
2287
+ "learning_rate": 1.4672686282730622e-05,
2288
+ "loss": 0.2606,
2289
+ "step": 380
2290
+ },
2291
+ {
2292
+ "epoch": 0.73,
2293
+ "learning_rate": 1.4645108767806778e-05,
2294
+ "loss": 0.3498,
2295
+ "step": 381
2296
+ },
2297
+ {
2298
+ "epoch": 0.73,
2299
+ "learning_rate": 1.4617486132350343e-05,
2300
+ "loss": 0.3105,
2301
+ "step": 382
2302
+ },
2303
+ {
2304
+ "epoch": 0.74,
2305
+ "learning_rate": 1.4589818644675378e-05,
2306
+ "loss": 0.2651,
2307
+ "step": 383
2308
+ },
2309
+ {
2310
+ "epoch": 0.74,
2311
+ "learning_rate": 1.4562106573531632e-05,
2312
+ "loss": 0.2546,
2313
+ "step": 384
2314
+ },
2315
+ {
2316
+ "epoch": 0.74,
2317
+ "learning_rate": 1.4534350188101905e-05,
2318
+ "loss": 0.3083,
2319
+ "step": 385
2320
+ },
2321
+ {
2322
+ "epoch": 0.74,
2323
+ "learning_rate": 1.4506549757999456e-05,
2324
+ "loss": 0.2503,
2325
+ "step": 386
2326
+ },
2327
+ {
2328
+ "epoch": 0.74,
2329
+ "learning_rate": 1.4478705553265363e-05,
2330
+ "loss": 0.2802,
2331
+ "step": 387
2332
+ },
2333
+ {
2334
+ "epoch": 0.75,
2335
+ "learning_rate": 1.4450817844365924e-05,
2336
+ "loss": 0.2667,
2337
+ "step": 388
2338
+ },
2339
+ {
2340
+ "epoch": 0.75,
2341
+ "learning_rate": 1.4422886902190014e-05,
2342
+ "loss": 0.2697,
2343
+ "step": 389
2344
+ },
2345
+ {
2346
+ "epoch": 0.75,
2347
+ "learning_rate": 1.4394912998046451e-05,
2348
+ "loss": 0.2197,
2349
+ "step": 390
2350
+ },
2351
+ {
2352
+ "epoch": 0.75,
2353
+ "learning_rate": 1.436689640366137e-05,
2354
+ "loss": 0.2745,
2355
+ "step": 391
2356
+ },
2357
+ {
2358
+ "epoch": 0.75,
2359
+ "learning_rate": 1.4338837391175582e-05,
2360
+ "loss": 0.2911,
2361
+ "step": 392
2362
+ },
2363
+ {
2364
+ "epoch": 0.76,
2365
+ "learning_rate": 1.4310736233141926e-05,
2366
+ "loss": 0.3044,
2367
+ "step": 393
2368
+ },
2369
+ {
2370
+ "epoch": 0.76,
2371
+ "learning_rate": 1.4282593202522627e-05,
2372
+ "loss": 0.2614,
2373
+ "step": 394
2374
+ },
2375
+ {
2376
+ "epoch": 0.76,
2377
+ "learning_rate": 1.4254408572686642e-05,
2378
+ "loss": 0.2115,
2379
+ "step": 395
2380
+ },
2381
+ {
2382
+ "epoch": 0.76,
2383
+ "learning_rate": 1.4226182617406996e-05,
2384
+ "loss": 0.2747,
2385
+ "step": 396
2386
+ },
2387
+ {
2388
+ "epoch": 0.76,
2389
+ "learning_rate": 1.4197915610858143e-05,
2390
+ "loss": 0.2757,
2391
+ "step": 397
2392
+ },
2393
+ {
2394
+ "epoch": 0.77,
2395
+ "learning_rate": 1.4169607827613284e-05,
2396
+ "loss": 0.319,
2397
+ "step": 398
2398
+ },
2399
+ {
2400
+ "epoch": 0.77,
2401
+ "learning_rate": 1.4141259542641706e-05,
2402
+ "loss": 0.2993,
2403
+ "step": 399
2404
+ },
2405
+ {
2406
+ "epoch": 0.77,
2407
+ "learning_rate": 1.4112871031306118e-05,
2408
+ "loss": 0.3396,
2409
+ "step": 400
2410
+ },
2411
+ {
2412
+ "epoch": 0.77,
2413
+ "learning_rate": 1.4084442569359964e-05,
2414
+ "loss": 0.3013,
2415
+ "step": 401
2416
+ },
2417
+ {
2418
+ "epoch": 0.77,
2419
+ "learning_rate": 1.4055974432944753e-05,
2420
+ "loss": 0.3251,
2421
+ "step": 402
2422
+ },
2423
+ {
2424
+ "epoch": 0.78,
2425
+ "learning_rate": 1.4027466898587375e-05,
2426
+ "loss": 0.2337,
2427
+ "step": 403
2428
+ },
2429
+ {
2430
+ "epoch": 0.78,
2431
+ "learning_rate": 1.3998920243197408e-05,
2432
+ "loss": 0.265,
2433
+ "step": 404
2434
+ },
2435
+ {
2436
+ "epoch": 0.78,
2437
+ "learning_rate": 1.3970334744064451e-05,
2438
+ "loss": 0.2818,
2439
+ "step": 405
2440
+ },
2441
+ {
2442
+ "epoch": 0.78,
2443
+ "learning_rate": 1.3941710678855396e-05,
2444
+ "loss": 0.2539,
2445
+ "step": 406
2446
+ },
2447
+ {
2448
+ "epoch": 0.78,
2449
+ "learning_rate": 1.391304832561175e-05,
2450
+ "loss": 0.2297,
2451
+ "step": 407
2452
+ },
2453
+ {
2454
+ "epoch": 0.78,
2455
+ "learning_rate": 1.3884347962746949e-05,
2456
+ "loss": 0.4147,
2457
+ "step": 408
2458
+ },
2459
+ {
2460
+ "epoch": 0.79,
2461
+ "learning_rate": 1.3855609869043618e-05,
2462
+ "loss": 0.2954,
2463
+ "step": 409
2464
+ },
2465
+ {
2466
+ "epoch": 0.79,
2467
+ "learning_rate": 1.3826834323650899e-05,
2468
+ "loss": 0.3311,
2469
+ "step": 410
2470
+ },
2471
+ {
2472
+ "epoch": 0.79,
2473
+ "learning_rate": 1.3798021606081713e-05,
2474
+ "loss": 0.2711,
2475
+ "step": 411
2476
+ },
2477
+ {
2478
+ "epoch": 0.79,
2479
+ "learning_rate": 1.3769171996210053e-05,
2480
+ "loss": 0.3318,
2481
+ "step": 412
2482
+ },
2483
+ {
2484
+ "epoch": 0.79,
2485
+ "learning_rate": 1.3740285774268282e-05,
2486
+ "loss": 0.2981,
2487
+ "step": 413
2488
+ },
2489
+ {
2490
+ "epoch": 0.8,
2491
+ "learning_rate": 1.371136322084438e-05,
2492
+ "loss": 0.232,
2493
+ "step": 414
2494
+ },
2495
+ {
2496
+ "epoch": 0.8,
2497
+ "learning_rate": 1.3682404616879246e-05,
2498
+ "loss": 0.2655,
2499
+ "step": 415
2500
+ },
2501
+ {
2502
+ "epoch": 0.8,
2503
+ "learning_rate": 1.3653410243663953e-05,
2504
+ "loss": 0.2999,
2505
+ "step": 416
2506
+ },
2507
+ {
2508
+ "epoch": 0.8,
2509
+ "learning_rate": 1.3624380382837017e-05,
2510
+ "loss": 0.3296,
2511
+ "step": 417
2512
+ },
2513
+ {
2514
+ "epoch": 0.8,
2515
+ "learning_rate": 1.3595315316381676e-05,
2516
+ "loss": 0.2932,
2517
+ "step": 418
2518
+ },
2519
+ {
2520
+ "epoch": 0.81,
2521
+ "learning_rate": 1.3566215326623131e-05,
2522
+ "loss": 0.3093,
2523
+ "step": 419
2524
+ },
2525
+ {
2526
+ "epoch": 0.81,
2527
+ "learning_rate": 1.3537080696225815e-05,
2528
+ "loss": 0.2761,
2529
+ "step": 420
2530
+ },
2531
+ {
2532
+ "epoch": 0.81,
2533
+ "learning_rate": 1.3507911708190646e-05,
2534
+ "loss": 0.2225,
2535
+ "step": 421
2536
+ },
2537
+ {
2538
+ "epoch": 0.81,
2539
+ "learning_rate": 1.3478708645852272e-05,
2540
+ "loss": 0.2524,
2541
+ "step": 422
2542
+ },
2543
+ {
2544
+ "epoch": 0.81,
2545
+ "learning_rate": 1.3449471792876333e-05,
2546
+ "loss": 0.329,
2547
+ "step": 423
2548
+ },
2549
+ {
2550
+ "epoch": 0.82,
2551
+ "learning_rate": 1.342020143325669e-05,
2552
+ "loss": 0.226,
2553
+ "step": 424
2554
+ },
2555
+ {
2556
+ "epoch": 0.82,
2557
+ "learning_rate": 1.3390897851312667e-05,
2558
+ "loss": 0.2471,
2559
+ "step": 425
2560
+ },
2561
+ {
2562
+ "epoch": 0.82,
2563
+ "learning_rate": 1.336156133168631e-05,
2564
+ "loss": 0.3506,
2565
+ "step": 426
2566
+ },
2567
+ {
2568
+ "epoch": 0.82,
2569
+ "learning_rate": 1.3332192159339595e-05,
2570
+ "loss": 0.3431,
2571
+ "step": 427
2572
+ },
2573
+ {
2574
+ "epoch": 0.82,
2575
+ "learning_rate": 1.3302790619551673e-05,
2576
+ "loss": 0.3355,
2577
+ "step": 428
2578
+ },
2579
+ {
2580
+ "epoch": 0.82,
2581
+ "learning_rate": 1.3273356997916106e-05,
2582
+ "loss": 0.2985,
2583
+ "step": 429
2584
+ },
2585
+ {
2586
+ "epoch": 0.83,
2587
+ "learning_rate": 1.3243891580338074e-05,
2588
+ "loss": 0.2887,
2589
+ "step": 430
2590
+ },
2591
+ {
2592
+ "epoch": 0.83,
2593
+ "learning_rate": 1.3214394653031616e-05,
2594
+ "loss": 0.2807,
2595
+ "step": 431
2596
+ },
2597
+ {
2598
+ "epoch": 0.83,
2599
+ "learning_rate": 1.3184866502516846e-05,
2600
+ "loss": 0.2298,
2601
+ "step": 432
2602
+ },
2603
+ {
2604
+ "epoch": 0.83,
2605
+ "learning_rate": 1.3155307415617156e-05,
2606
+ "loss": 0.2633,
2607
+ "step": 433
2608
+ },
2609
+ {
2610
+ "epoch": 0.83,
2611
+ "learning_rate": 1.3125717679456447e-05,
2612
+ "loss": 0.2659,
2613
+ "step": 434
2614
+ },
2615
+ {
2616
+ "epoch": 0.84,
2617
+ "learning_rate": 1.309609758145633e-05,
2618
+ "loss": 0.2968,
2619
+ "step": 435
2620
+ },
2621
+ {
2622
+ "epoch": 0.84,
2623
+ "learning_rate": 1.3066447409333345e-05,
2624
+ "loss": 0.2782,
2625
+ "step": 436
2626
+ },
2627
+ {
2628
+ "epoch": 0.84,
2629
+ "learning_rate": 1.3036767451096148e-05,
2630
+ "loss": 0.2766,
2631
+ "step": 437
2632
+ },
2633
+ {
2634
+ "epoch": 0.84,
2635
+ "learning_rate": 1.300705799504273e-05,
2636
+ "loss": 0.2436,
2637
+ "step": 438
2638
+ },
2639
+ {
2640
+ "epoch": 0.84,
2641
+ "learning_rate": 1.2977319329757616e-05,
2642
+ "loss": 0.3231,
2643
+ "step": 439
2644
+ },
2645
+ {
2646
+ "epoch": 0.85,
2647
+ "learning_rate": 1.2947551744109044e-05,
2648
+ "loss": 0.272,
2649
+ "step": 440
2650
+ },
2651
+ {
2652
+ "epoch": 0.85,
2653
+ "learning_rate": 1.2917755527246179e-05,
2654
+ "loss": 0.2204,
2655
+ "step": 441
2656
+ },
2657
+ {
2658
+ "epoch": 0.85,
2659
+ "learning_rate": 1.28879309685963e-05,
2660
+ "loss": 0.294,
2661
+ "step": 442
2662
+ },
2663
+ {
2664
+ "epoch": 0.85,
2665
+ "learning_rate": 1.2858078357861979e-05,
2666
+ "loss": 0.2625,
2667
+ "step": 443
2668
+ },
2669
+ {
2670
+ "epoch": 0.85,
2671
+ "learning_rate": 1.2828197985018276e-05,
2672
+ "loss": 0.2652,
2673
+ "step": 444
2674
+ },
2675
+ {
2676
+ "epoch": 0.86,
2677
+ "learning_rate": 1.2798290140309924e-05,
2678
+ "loss": 0.297,
2679
+ "step": 445
2680
+ },
2681
+ {
2682
+ "epoch": 0.86,
2683
+ "learning_rate": 1.2768355114248493e-05,
2684
+ "loss": 0.2919,
2685
+ "step": 446
2686
+ },
2687
+ {
2688
+ "epoch": 0.86,
2689
+ "learning_rate": 1.2738393197609602e-05,
2690
+ "loss": 0.3062,
2691
+ "step": 447
2692
+ },
2693
+ {
2694
+ "epoch": 0.86,
2695
+ "learning_rate": 1.2708404681430054e-05,
2696
+ "loss": 0.2893,
2697
+ "step": 448
2698
+ },
2699
+ {
2700
+ "epoch": 0.86,
2701
+ "learning_rate": 1.2678389857005033e-05,
2702
+ "loss": 0.2747,
2703
+ "step": 449
2704
+ },
2705
+ {
2706
+ "epoch": 0.87,
2707
+ "learning_rate": 1.2648349015885272e-05,
2708
+ "loss": 0.2863,
2709
+ "step": 450
2710
+ },
2711
+ {
2712
+ "epoch": 0.87,
2713
+ "learning_rate": 1.2618282449874221e-05,
2714
+ "loss": 0.3022,
2715
+ "step": 451
2716
+ },
2717
+ {
2718
+ "epoch": 0.87,
2719
+ "learning_rate": 1.2588190451025209e-05,
2720
+ "loss": 0.2665,
2721
+ "step": 452
2722
+ },
2723
+ {
2724
+ "epoch": 0.87,
2725
+ "learning_rate": 1.2558073311638604e-05,
2726
+ "loss": 0.2627,
2727
+ "step": 453
2728
+ },
2729
+ {
2730
+ "epoch": 0.87,
2731
+ "learning_rate": 1.2527931324258975e-05,
2732
+ "loss": 0.2748,
2733
+ "step": 454
2734
+ },
2735
+ {
2736
+ "epoch": 0.88,
2737
+ "learning_rate": 1.249776478167227e-05,
2738
+ "loss": 0.2935,
2739
+ "step": 455
2740
+ },
2741
+ {
2742
+ "epoch": 0.88,
2743
+ "learning_rate": 1.2467573976902936e-05,
2744
+ "loss": 0.3557,
2745
+ "step": 456
2746
+ },
2747
+ {
2748
+ "epoch": 0.88,
2749
+ "learning_rate": 1.2437359203211109e-05,
2750
+ "loss": 0.269,
2751
+ "step": 457
2752
+ },
2753
+ {
2754
+ "epoch": 0.88,
2755
+ "learning_rate": 1.2407120754089733e-05,
2756
+ "loss": 0.3202,
2757
+ "step": 458
2758
+ },
2759
+ {
2760
+ "epoch": 0.88,
2761
+ "learning_rate": 1.2376858923261732e-05,
2762
+ "loss": 0.284,
2763
+ "step": 459
2764
+ },
2765
+ {
2766
+ "epoch": 0.88,
2767
+ "learning_rate": 1.2346574004677154e-05,
2768
+ "loss": 0.2671,
2769
+ "step": 460
2770
+ },
2771
+ {
2772
+ "epoch": 0.89,
2773
+ "learning_rate": 1.2316266292510305e-05,
2774
+ "loss": 0.2731,
2775
+ "step": 461
2776
+ },
2777
+ {
2778
+ "epoch": 0.89,
2779
+ "learning_rate": 1.2285936081156897e-05,
2780
+ "loss": 0.3232,
2781
+ "step": 462
2782
+ },
2783
+ {
2784
+ "epoch": 0.89,
2785
+ "learning_rate": 1.2255583665231196e-05,
2786
+ "loss": 0.2675,
2787
+ "step": 463
2788
+ },
2789
+ {
2790
+ "epoch": 0.89,
2791
+ "learning_rate": 1.2225209339563144e-05,
2792
+ "loss": 0.3438,
2793
+ "step": 464
2794
+ },
2795
+ {
2796
+ "epoch": 0.89,
2797
+ "learning_rate": 1.2194813399195518e-05,
2798
+ "loss": 0.269,
2799
+ "step": 465
2800
+ },
2801
+ {
2802
+ "epoch": 0.9,
2803
+ "learning_rate": 1.2164396139381029e-05,
2804
+ "loss": 0.3115,
2805
+ "step": 466
2806
+ },
2807
+ {
2808
+ "epoch": 0.9,
2809
+ "learning_rate": 1.2133957855579501e-05,
2810
+ "loss": 0.2436,
2811
+ "step": 467
2812
+ },
2813
+ {
2814
+ "epoch": 0.9,
2815
+ "learning_rate": 1.210349884345496e-05,
2816
+ "loss": 0.3131,
2817
+ "step": 468
2818
+ },
2819
+ {
2820
+ "epoch": 0.9,
2821
+ "learning_rate": 1.2073019398872778e-05,
2822
+ "loss": 0.3019,
2823
+ "step": 469
2824
+ },
2825
+ {
2826
+ "epoch": 0.9,
2827
+ "learning_rate": 1.2042519817896805e-05,
2828
+ "loss": 0.2505,
2829
+ "step": 470
2830
+ },
2831
+ {
2832
+ "epoch": 0.91,
2833
+ "learning_rate": 1.2012000396786485e-05,
2834
+ "loss": 0.2825,
2835
+ "step": 471
2836
+ },
2837
+ {
2838
+ "epoch": 0.91,
2839
+ "learning_rate": 1.1981461431993978e-05,
2840
+ "loss": 0.2789,
2841
+ "step": 472
2842
+ },
2843
+ {
2844
+ "epoch": 0.91,
2845
+ "learning_rate": 1.1950903220161286e-05,
2846
+ "loss": 0.2276,
2847
+ "step": 473
2848
+ },
2849
+ {
2850
+ "epoch": 0.91,
2851
+ "learning_rate": 1.1920326058117364e-05,
2852
+ "loss": 0.2777,
2853
+ "step": 474
2854
+ },
2855
+ {
2856
+ "epoch": 0.91,
2857
+ "learning_rate": 1.1889730242875243e-05,
2858
+ "loss": 0.2774,
2859
+ "step": 475
2860
+ },
2861
+ {
2862
+ "epoch": 0.92,
2863
+ "learning_rate": 1.1859116071629148e-05,
2864
+ "loss": 0.2911,
2865
+ "step": 476
2866
+ },
2867
+ {
2868
+ "epoch": 0.92,
2869
+ "learning_rate": 1.1828483841751597e-05,
2870
+ "loss": 0.3153,
2871
+ "step": 477
2872
+ },
2873
+ {
2874
+ "epoch": 0.92,
2875
+ "learning_rate": 1.1797833850790527e-05,
2876
+ "loss": 0.3061,
2877
+ "step": 478
2878
+ },
2879
+ {
2880
+ "epoch": 0.92,
2881
+ "learning_rate": 1.1767166396466404e-05,
2882
+ "loss": 0.2931,
2883
+ "step": 479
2884
+ },
2885
+ {
2886
+ "epoch": 0.92,
2887
+ "learning_rate": 1.1736481776669307e-05,
2888
+ "loss": 0.2509,
2889
+ "step": 480
2890
+ },
2891
+ {
2892
+ "epoch": 0.93,
2893
+ "learning_rate": 1.1705780289456069e-05,
2894
+ "loss": 0.2861,
2895
+ "step": 481
2896
+ },
2897
+ {
2898
+ "epoch": 0.93,
2899
+ "learning_rate": 1.1675062233047365e-05,
2900
+ "loss": 0.2847,
2901
+ "step": 482
2902
+ },
2903
+ {
2904
+ "epoch": 0.93,
2905
+ "learning_rate": 1.1644327905824808e-05,
2906
+ "loss": 0.2579,
2907
+ "step": 483
2908
+ },
2909
+ {
2910
+ "epoch": 0.93,
2911
+ "learning_rate": 1.1613577606328068e-05,
2912
+ "loss": 0.3258,
2913
+ "step": 484
2914
+ },
2915
+ {
2916
+ "epoch": 0.93,
2917
+ "learning_rate": 1.1582811633251949e-05,
2918
+ "loss": 0.2762,
2919
+ "step": 485
2920
+ },
2921
+ {
2922
+ "epoch": 0.93,
2923
+ "learning_rate": 1.1552030285443516e-05,
2924
+ "loss": 0.2241,
2925
+ "step": 486
2926
+ },
2927
+ {
2928
+ "epoch": 0.94,
2929
+ "learning_rate": 1.1521233861899168e-05,
2930
+ "loss": 0.2401,
2931
+ "step": 487
2932
+ },
2933
+ {
2934
+ "epoch": 0.94,
2935
+ "learning_rate": 1.1490422661761744e-05,
2936
+ "loss": 0.2195,
2937
+ "step": 488
2938
+ },
2939
+ {
2940
+ "epoch": 0.94,
2941
+ "learning_rate": 1.1459596984317622e-05,
2942
+ "loss": 0.2947,
2943
+ "step": 489
2944
+ },
2945
+ {
2946
+ "epoch": 0.94,
2947
+ "learning_rate": 1.1428757128993801e-05,
2948
+ "loss": 0.257,
2949
+ "step": 490
2950
+ },
2951
+ {
2952
+ "epoch": 0.94,
2953
+ "learning_rate": 1.1397903395354996e-05,
2954
+ "loss": 0.2639,
2955
+ "step": 491
2956
+ },
2957
+ {
2958
+ "epoch": 0.95,
2959
+ "learning_rate": 1.1367036083100735e-05,
2960
+ "loss": 0.2851,
2961
+ "step": 492
2962
+ },
2963
+ {
2964
+ "epoch": 0.95,
2965
+ "learning_rate": 1.1336155492062439e-05,
2966
+ "loss": 0.2933,
2967
+ "step": 493
2968
+ },
2969
+ {
2970
+ "epoch": 0.95,
2971
+ "learning_rate": 1.130526192220052e-05,
2972
+ "loss": 0.2513,
2973
+ "step": 494
2974
+ },
2975
+ {
2976
+ "epoch": 0.95,
2977
+ "learning_rate": 1.1274355673601446e-05,
2978
+ "loss": 0.2302,
2979
+ "step": 495
2980
+ },
2981
+ {
2982
+ "epoch": 0.95,
2983
+ "learning_rate": 1.1243437046474854e-05,
2984
+ "loss": 0.2771,
2985
+ "step": 496
2986
+ },
2987
+ {
2988
+ "epoch": 0.96,
2989
+ "learning_rate": 1.1212506341150615e-05,
2990
+ "loss": 0.3137,
2991
+ "step": 497
2992
+ },
2993
+ {
2994
+ "epoch": 0.96,
2995
+ "learning_rate": 1.118156385807593e-05,
2996
+ "loss": 0.3084,
2997
+ "step": 498
2998
+ },
2999
+ {
3000
+ "epoch": 0.96,
3001
+ "learning_rate": 1.1150609897812387e-05,
3002
+ "loss": 0.2404,
3003
+ "step": 499
3004
+ },
3005
+ {
3006
+ "epoch": 0.96,
3007
+ "learning_rate": 1.1119644761033079e-05,
3008
+ "loss": 0.2192,
3009
+ "step": 500
3010
+ },
3011
+ {
3012
+ "epoch": 0.96,
3013
+ "learning_rate": 1.1088668748519646e-05,
3014
+ "loss": 0.2899,
3015
+ "step": 501
3016
+ },
3017
+ {
3018
+ "epoch": 0.97,
3019
+ "learning_rate": 1.105768216115938e-05,
3020
+ "loss": 0.2206,
3021
+ "step": 502
3022
+ },
3023
+ {
3024
+ "epoch": 0.97,
3025
+ "learning_rate": 1.1026685299942286e-05,
3026
+ "loss": 0.2641,
3027
+ "step": 503
3028
+ },
3029
+ {
3030
+ "epoch": 0.97,
3031
+ "learning_rate": 1.0995678465958168e-05,
3032
+ "loss": 0.2539,
3033
+ "step": 504
3034
+ },
3035
+ {
3036
+ "epoch": 0.97,
3037
+ "learning_rate": 1.0964661960393703e-05,
3038
+ "loss": 0.2721,
3039
+ "step": 505
3040
+ },
3041
+ {
3042
+ "epoch": 0.97,
3043
+ "learning_rate": 1.0933636084529507e-05,
3044
+ "loss": 0.2854,
3045
+ "step": 506
3046
+ },
3047
+ {
3048
+ "epoch": 0.97,
3049
+ "learning_rate": 1.0902601139737225e-05,
3050
+ "loss": 0.2651,
3051
+ "step": 507
3052
+ },
3053
+ {
3054
+ "epoch": 0.98,
3055
+ "learning_rate": 1.0871557427476585e-05,
3056
+ "loss": 0.2654,
3057
+ "step": 508
3058
+ },
3059
+ {
3060
+ "epoch": 0.98,
3061
+ "learning_rate": 1.0840505249292477e-05,
3062
+ "loss": 0.2759,
3063
+ "step": 509
3064
+ },
3065
+ {
3066
+ "epoch": 0.98,
3067
+ "learning_rate": 1.0809444906812034e-05,
3068
+ "loss": 0.246,
3069
+ "step": 510
3070
+ },
3071
+ {
3072
+ "epoch": 0.98,
3073
+ "learning_rate": 1.0778376701741688e-05,
3074
+ "loss": 0.2517,
3075
+ "step": 511
3076
+ },
3077
+ {
3078
+ "epoch": 0.98,
3079
+ "learning_rate": 1.0747300935864245e-05,
3080
+ "loss": 0.2957,
3081
+ "step": 512
3082
+ },
3083
+ {
3084
+ "epoch": 0.99,
3085
+ "learning_rate": 1.0716217911035952e-05,
3086
+ "loss": 0.2462,
3087
+ "step": 513
3088
+ },
3089
+ {
3090
+ "epoch": 0.99,
3091
+ "learning_rate": 1.0685127929183567e-05,
3092
+ "loss": 0.3094,
3093
+ "step": 514
3094
+ },
3095
+ {
3096
+ "epoch": 0.99,
3097
+ "learning_rate": 1.0654031292301432e-05,
3098
+ "loss": 0.3021,
3099
+ "step": 515
3100
+ },
3101
+ {
3102
+ "epoch": 0.99,
3103
+ "learning_rate": 1.0622928302448523e-05,
3104
+ "loss": 0.2938,
3105
+ "step": 516
3106
+ },
3107
+ {
3108
+ "epoch": 0.99,
3109
+ "learning_rate": 1.0591819261745528e-05,
3110
+ "loss": 0.3073,
3111
+ "step": 517
3112
+ },
3113
+ {
3114
+ "epoch": 1.0,
3115
+ "learning_rate": 1.0560704472371919e-05,
3116
+ "loss": 0.2348,
3117
+ "step": 518
3118
+ },
3119
+ {
3120
+ "epoch": 1.0,
3121
+ "learning_rate": 1.0529584236562995e-05,
3122
+ "loss": 0.2721,
3123
+ "step": 519
3124
+ },
3125
+ {
3126
+ "epoch": 1.0,
3127
+ "learning_rate": 1.0498458856606972e-05,
3128
+ "loss": 0.2808,
3129
+ "step": 520
3130
+ },
3131
+ {
3132
+ "epoch": 1.0,
3133
+ "learning_rate": 1.0467328634842024e-05,
3134
+ "loss": 0.2259,
3135
+ "step": 521
3136
+ },
3137
+ {
3138
+ "epoch": 1.0,
3139
+ "learning_rate": 1.0436193873653362e-05,
3140
+ "loss": 0.2134,
3141
+ "step": 522
3142
+ },
3143
+ {
3144
+ "epoch": 1.01,
3145
+ "learning_rate": 1.0405054875470287e-05,
3146
+ "loss": 0.2164,
3147
+ "step": 523
3148
+ },
3149
+ {
3150
+ "epoch": 1.01,
3151
+ "learning_rate": 1.037391194276326e-05,
3152
+ "loss": 0.2579,
3153
+ "step": 524
3154
+ },
3155
+ {
3156
+ "epoch": 1.01,
3157
+ "learning_rate": 1.0342765378040953e-05,
3158
+ "loss": 0.2045,
3159
+ "step": 525
3160
+ },
3161
+ {
3162
+ "epoch": 1.01,
3163
+ "learning_rate": 1.0311615483847333e-05,
3164
+ "loss": 0.1862,
3165
+ "step": 526
3166
+ },
3167
+ {
3168
+ "epoch": 1.01,
3169
+ "learning_rate": 1.028046256275869e-05,
3170
+ "loss": 0.2848,
3171
+ "step": 527
3172
+ },
3173
+ {
3174
+ "epoch": 1.02,
3175
+ "learning_rate": 1.0249306917380731e-05,
3176
+ "loss": 0.2049,
3177
+ "step": 528
3178
+ },
3179
+ {
3180
+ "epoch": 1.02,
3181
+ "learning_rate": 1.0218148850345613e-05,
3182
+ "loss": 0.2282,
3183
+ "step": 529
3184
+ },
3185
+ {
3186
+ "epoch": 1.02,
3187
+ "learning_rate": 1.0186988664309023e-05,
3188
+ "loss": 0.2015,
3189
+ "step": 530
3190
+ },
3191
+ {
3192
+ "epoch": 1.02,
3193
+ "learning_rate": 1.0155826661947232e-05,
3194
+ "loss": 0.1778,
3195
+ "step": 531
3196
+ },
3197
+ {
3198
+ "epoch": 1.02,
3199
+ "learning_rate": 1.0124663145954152e-05,
3200
+ "loss": 0.2184,
3201
+ "step": 532
3202
+ },
3203
+ {
3204
+ "epoch": 1.02,
3205
+ "learning_rate": 1.0093498419038394e-05,
3206
+ "loss": 0.2326,
3207
+ "step": 533
3208
+ },
3209
+ {
3210
+ "epoch": 1.03,
3211
+ "learning_rate": 1.0062332783920337e-05,
3212
+ "loss": 0.2122,
3213
+ "step": 534
3214
+ },
3215
+ {
3216
+ "epoch": 1.03,
3217
+ "learning_rate": 1.0031166543329179e-05,
3218
+ "loss": 0.2254,
3219
+ "step": 535
3220
+ },
3221
+ {
3222
+ "epoch": 1.03,
3223
+ "learning_rate": 1e-05,
3224
+ "loss": 0.2029,
3225
+ "step": 536
3226
+ },
3227
+ {
3228
+ "epoch": 1.03,
3229
+ "learning_rate": 9.968833456670824e-06,
3230
+ "loss": 0.2226,
3231
+ "step": 537
3232
+ },
3233
+ {
3234
+ "epoch": 1.03,
3235
+ "learning_rate": 9.937667216079665e-06,
3236
+ "loss": 0.2244,
3237
+ "step": 538
3238
+ },
3239
+ {
3240
+ "epoch": 1.04,
3241
+ "learning_rate": 9.90650158096161e-06,
3242
+ "loss": 0.1925,
3243
+ "step": 539
3244
+ },
3245
+ {
3246
+ "epoch": 1.04,
3247
+ "learning_rate": 9.87533685404585e-06,
3248
+ "loss": 0.177,
3249
+ "step": 540
3250
+ },
3251
+ {
3252
+ "epoch": 1.04,
3253
+ "learning_rate": 9.844173338052771e-06,
3254
+ "loss": 0.2026,
3255
+ "step": 541
3256
+ },
3257
+ {
3258
+ "epoch": 1.04,
3259
+ "learning_rate": 9.81301133569098e-06,
3260
+ "loss": 0.21,
3261
+ "step": 542
3262
+ },
3263
+ {
3264
+ "epoch": 1.04,
3265
+ "learning_rate": 9.78185114965439e-06,
3266
+ "loss": 0.2318,
3267
+ "step": 543
3268
+ },
3269
+ {
3270
+ "epoch": 1.05,
3271
+ "learning_rate": 9.750693082619274e-06,
3272
+ "loss": 0.1783,
3273
+ "step": 544
3274
+ },
3275
+ {
3276
+ "epoch": 1.05,
3277
+ "learning_rate": 9.719537437241311e-06,
3278
+ "loss": 0.1722,
3279
+ "step": 545
3280
+ },
3281
+ {
3282
+ "epoch": 1.05,
3283
+ "learning_rate": 9.68838451615267e-06,
3284
+ "loss": 0.1958,
3285
+ "step": 546
3286
+ },
3287
+ {
3288
+ "epoch": 1.05,
3289
+ "learning_rate": 9.65723462195905e-06,
3290
+ "loss": 0.209,
3291
+ "step": 547
3292
+ },
3293
+ {
3294
+ "epoch": 1.05,
3295
+ "learning_rate": 9.626088057236745e-06,
3296
+ "loss": 0.2057,
3297
+ "step": 548
3298
+ },
3299
+ {
3300
+ "epoch": 1.06,
3301
+ "learning_rate": 9.594945124529718e-06,
3302
+ "loss": 0.2427,
3303
+ "step": 549
3304
+ },
3305
+ {
3306
+ "epoch": 1.06,
3307
+ "learning_rate": 9.563806126346643e-06,
3308
+ "loss": 0.2325,
3309
+ "step": 550
3310
+ },
3311
+ {
3312
+ "epoch": 1.06,
3313
+ "learning_rate": 9.532671365157979e-06,
3314
+ "loss": 0.218,
3315
+ "step": 551
3316
+ },
3317
+ {
3318
+ "epoch": 1.06,
3319
+ "learning_rate": 9.501541143393028e-06,
3320
+ "loss": 0.1853,
3321
+ "step": 552
3322
+ },
3323
+ {
3324
+ "epoch": 1.06,
3325
+ "learning_rate": 9.470415763437003e-06,
3326
+ "loss": 0.2607,
3327
+ "step": 553
3328
+ },
3329
+ {
3330
+ "epoch": 1.07,
3331
+ "learning_rate": 9.439295527628083e-06,
3332
+ "loss": 0.2413,
3333
+ "step": 554
3334
+ },
3335
+ {
3336
+ "epoch": 1.07,
3337
+ "learning_rate": 9.408180738254472e-06,
3338
+ "loss": 0.1771,
3339
+ "step": 555
3340
+ },
3341
+ {
3342
+ "epoch": 1.07,
3343
+ "learning_rate": 9.377071697551479e-06,
3344
+ "loss": 0.1908,
3345
+ "step": 556
3346
+ },
3347
+ {
3348
+ "epoch": 1.07,
3349
+ "learning_rate": 9.34596870769857e-06,
3350
+ "loss": 0.1837,
3351
+ "step": 557
3352
+ },
3353
+ {
3354
+ "epoch": 1.07,
3355
+ "learning_rate": 9.314872070816435e-06,
3356
+ "loss": 0.1998,
3357
+ "step": 558
3358
+ },
3359
+ {
3360
+ "epoch": 1.07,
3361
+ "learning_rate": 9.28378208896405e-06,
3362
+ "loss": 0.1774,
3363
+ "step": 559
3364
+ },
3365
+ {
3366
+ "epoch": 1.08,
3367
+ "learning_rate": 9.252699064135759e-06,
3368
+ "loss": 0.2101,
3369
+ "step": 560
3370
+ },
3371
+ {
3372
+ "epoch": 1.08,
3373
+ "learning_rate": 9.221623298258315e-06,
3374
+ "loss": 0.2664,
3375
+ "step": 561
3376
+ },
3377
+ {
3378
+ "epoch": 1.08,
3379
+ "learning_rate": 9.190555093187968e-06,
3380
+ "loss": 0.2086,
3381
+ "step": 562
3382
+ },
3383
+ {
3384
+ "epoch": 1.08,
3385
+ "learning_rate": 9.159494750707527e-06,
3386
+ "loss": 0.2233,
3387
+ "step": 563
3388
+ },
3389
+ {
3390
+ "epoch": 1.08,
3391
+ "learning_rate": 9.128442572523418e-06,
3392
+ "loss": 0.183,
3393
+ "step": 564
3394
+ },
3395
+ {
3396
+ "epoch": 1.09,
3397
+ "learning_rate": 9.097398860262777e-06,
3398
+ "loss": 0.2087,
3399
+ "step": 565
3400
+ },
3401
+ {
3402
+ "epoch": 1.09,
3403
+ "learning_rate": 9.066363915470494e-06,
3404
+ "loss": 0.2189,
3405
+ "step": 566
3406
+ },
3407
+ {
3408
+ "epoch": 1.09,
3409
+ "learning_rate": 9.0353380396063e-06,
3410
+ "loss": 0.2335,
3411
+ "step": 567
3412
+ },
3413
+ {
3414
+ "epoch": 1.09,
3415
+ "learning_rate": 9.004321534041836e-06,
3416
+ "loss": 0.174,
3417
+ "step": 568
3418
+ },
3419
+ {
3420
+ "epoch": 1.09,
3421
+ "learning_rate": 8.973314700057717e-06,
3422
+ "loss": 0.1857,
3423
+ "step": 569
3424
+ },
3425
+ {
3426
+ "epoch": 1.1,
3427
+ "learning_rate": 8.942317838840625e-06,
3428
+ "loss": 0.1739,
3429
+ "step": 570
3430
+ },
3431
+ {
3432
+ "epoch": 1.1,
3433
+ "learning_rate": 8.911331251480357e-06,
3434
+ "loss": 0.2238,
3435
+ "step": 571
3436
+ },
3437
+ {
3438
+ "epoch": 1.1,
3439
+ "learning_rate": 8.880355238966923e-06,
3440
+ "loss": 0.181,
3441
+ "step": 572
3442
+ },
3443
+ {
3444
+ "epoch": 1.1,
3445
+ "learning_rate": 8.849390102187615e-06,
3446
+ "loss": 0.2084,
3447
+ "step": 573
3448
+ },
3449
+ {
3450
+ "epoch": 1.1,
3451
+ "learning_rate": 8.818436141924072e-06,
3452
+ "loss": 0.1913,
3453
+ "step": 574
3454
+ },
3455
+ {
3456
+ "epoch": 1.11,
3457
+ "learning_rate": 8.787493658849387e-06,
3458
+ "loss": 0.2048,
3459
+ "step": 575
3460
+ },
3461
+ {
3462
+ "epoch": 1.11,
3463
+ "learning_rate": 8.756562953525151e-06,
3464
+ "loss": 0.1866,
3465
+ "step": 576
3466
+ },
3467
+ {
3468
+ "epoch": 1.11,
3469
+ "learning_rate": 8.72564432639856e-06,
3470
+ "loss": 0.206,
3471
+ "step": 577
3472
+ },
3473
+ {
3474
+ "epoch": 1.11,
3475
+ "learning_rate": 8.694738077799487e-06,
3476
+ "loss": 0.2072,
3477
+ "step": 578
3478
+ },
3479
+ {
3480
+ "epoch": 1.11,
3481
+ "learning_rate": 8.663844507937563e-06,
3482
+ "loss": 0.2051,
3483
+ "step": 579
3484
+ },
3485
+ {
3486
+ "epoch": 1.12,
3487
+ "learning_rate": 8.632963916899268e-06,
3488
+ "loss": 0.1961,
3489
+ "step": 580
3490
+ },
3491
+ {
3492
+ "epoch": 1.12,
3493
+ "learning_rate": 8.602096604645009e-06,
3494
+ "loss": 0.2125,
3495
+ "step": 581
3496
+ },
3497
+ {
3498
+ "epoch": 1.12,
3499
+ "learning_rate": 8.571242871006202e-06,
3500
+ "loss": 0.2137,
3501
+ "step": 582
3502
+ },
3503
+ {
3504
+ "epoch": 1.12,
3505
+ "learning_rate": 8.540403015682382e-06,
3506
+ "loss": 0.2458,
3507
+ "step": 583
3508
+ },
3509
+ {
3510
+ "epoch": 1.12,
3511
+ "learning_rate": 8.509577338238255e-06,
3512
+ "loss": 0.1986,
3513
+ "step": 584
3514
+ },
3515
+ {
3516
+ "epoch": 1.12,
3517
+ "learning_rate": 8.478766138100834e-06,
3518
+ "loss": 0.243,
3519
+ "step": 585
3520
+ },
3521
+ {
3522
+ "epoch": 1.13,
3523
+ "learning_rate": 8.447969714556484e-06,
3524
+ "loss": 0.1845,
3525
+ "step": 586
3526
+ },
3527
+ {
3528
+ "epoch": 1.13,
3529
+ "learning_rate": 8.417188366748051e-06,
3530
+ "loss": 0.1862,
3531
+ "step": 587
3532
+ },
3533
+ {
3534
+ "epoch": 1.13,
3535
+ "learning_rate": 8.386422393671934e-06,
3536
+ "loss": 0.2013,
3537
+ "step": 588
3538
+ },
3539
+ {
3540
+ "epoch": 1.13,
3541
+ "learning_rate": 8.355672094175192e-06,
3542
+ "loss": 0.2136,
3543
+ "step": 589
3544
+ },
3545
+ {
3546
+ "epoch": 1.13,
3547
+ "learning_rate": 8.324937766952638e-06,
3548
+ "loss": 0.2283,
3549
+ "step": 590
3550
+ },
3551
+ {
3552
+ "epoch": 1.14,
3553
+ "learning_rate": 8.294219710543931e-06,
3554
+ "loss": 0.156,
3555
+ "step": 591
3556
+ },
3557
+ {
3558
+ "epoch": 1.14,
3559
+ "learning_rate": 8.263518223330698e-06,
3560
+ "loss": 0.2428,
3561
+ "step": 592
3562
+ },
3563
+ {
3564
+ "epoch": 1.14,
3565
+ "learning_rate": 8.232833603533601e-06,
3566
+ "loss": 0.1803,
3567
+ "step": 593
3568
+ },
3569
+ {
3570
+ "epoch": 1.14,
3571
+ "learning_rate": 8.202166149209475e-06,
3572
+ "loss": 0.1704,
3573
+ "step": 594
3574
+ },
3575
+ {
3576
+ "epoch": 1.14,
3577
+ "learning_rate": 8.171516158248406e-06,
3578
+ "loss": 0.204,
3579
+ "step": 595
3580
+ },
3581
+ {
3582
+ "epoch": 1.15,
3583
+ "learning_rate": 8.140883928370855e-06,
3584
+ "loss": 0.2087,
3585
+ "step": 596
3586
+ },
3587
+ {
3588
+ "epoch": 1.15,
3589
+ "learning_rate": 8.11026975712476e-06,
3590
+ "loss": 0.2008,
3591
+ "step": 597
3592
+ },
3593
+ {
3594
+ "epoch": 1.15,
3595
+ "learning_rate": 8.079673941882639e-06,
3596
+ "loss": 0.206,
3597
+ "step": 598
3598
+ },
3599
+ {
3600
+ "epoch": 1.15,
3601
+ "learning_rate": 8.04909677983872e-06,
3602
+ "loss": 0.1813,
3603
+ "step": 599
3604
+ },
3605
+ {
3606
+ "epoch": 1.15,
3607
+ "learning_rate": 8.018538568006027e-06,
3608
+ "loss": 0.1867,
3609
+ "step": 600
3610
+ }
3611
+ ],
3612
+ "logging_steps": 1.0,
3613
+ "max_steps": 1040,
3614
+ "num_input_tokens_seen": 0,
3615
+ "num_train_epochs": 2,
3616
+ "save_steps": 100,
3617
+ "total_flos": 31305375940608.0,
3618
+ "train_batch_size": 16,
3619
+ "trial_name": null,
3620
+ "trial_params": null
3621
+ }
llava-v1.6-13b-unk-vqa-v1.1/checkpoint-700/model-00001-of-00006.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:517d03ee15345bc59cee2cc7b7acc991b758d0f4d03698251405860f4f55c8f0
3
+ size 4978276128