Hugging Face's logo

satwikapaul
/
painting_movement_classifier

Image Classification
Transformers
PyTorch
vit
Model card Files
xet
Community
2

Upload folder using huggingface_hub

#1
by satwikapaul - opened
base: refs/heads/main
←
from: refs/pr/1
Discussion Files changed
+5905
-0
Files changed (23) hide show
  1. all_results.json +8 -0
  2. checkpoint-2920/config.json +122 -0
  3. checkpoint-2920/optimizer.pt +3 -0
  4. checkpoint-2920/preprocessor_config.json +22 -0
  5. checkpoint-2920/pytorch_model.bin +3 -0
  6. checkpoint-2920/rng_state.pth +3 -0
  7. checkpoint-2920/scheduler.pt +3 -0
  8. checkpoint-2920/trainer_state.json +2425 -0
  9. checkpoint-2920/training_args.bin +3 -0
  10. checkpoint-640/config.json +122 -0
  11. checkpoint-640/optimizer.pt +3 -0
  12. checkpoint-640/preprocessor_config.json +22 -0
  13. checkpoint-640/pytorch_model.bin +3 -0
  14. checkpoint-640/rng_state.pth +3 -0
  15. checkpoint-640/scheduler.pt +3 -0
  16. checkpoint-640/trainer_state.json +544 -0
  17. checkpoint-640/training_args.bin +3 -0
  18. config.json +122 -0
  19. preprocessor_config.json +22 -0
  20. pytorch_model.bin +3 -0
  21. train_results.json +8 -0
  22. trainer_state.json +2452 -0
  23. training_args.bin +3 -0
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 15.0,
3
+ "total_flos": 3.64901908299393e+18,
4
+ "train_loss": 0.5233747142333972,
5
+ "train_runtime": 4896.0515,
6
+ "train_samples_per_second": 9.614,
7
+ "train_steps_per_second": 0.604
8
+ }
checkpoint-2920/config.json ADDED
@@ -0,0 +1,122 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/vit-base-patch16-224-in21k",
3
+ "architectures": [
4
+ "ViTForImageClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "encoder_stride": 16,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.0,
10
+ "hidden_size": 768,
11
+ "id2label": {
12
+ "0": "Abstract art",
13
+ "1": "African Art",
14
+ "10": "Constructivism art",
15
+ "11": "Constructivsm art",
16
+ "12": "Contemporary art",
17
+ "13": "Cubism art",
18
+ "14": "Ethiopian Art",
19
+ "15": "Expressionism paintings",
20
+ "16": "Fauvism paintings",
21
+ "17": "Gothic art",
22
+ "18": "Graffiti",
23
+ "19": "Impressionism and Post-Impressionism",
24
+ "2": "Amazonian Art",
25
+ "20": "Indochristian art",
26
+ "21": "Islamic Art",
27
+ "22": "Japanese Edo Period",
28
+ "23": "Japanese Heian Period",
29
+ "24": "Japanese momoyama",
30
+ "25": "Kamakura or Muromachi Art",
31
+ "26": "Korean Goryeo Dynasty Art",
32
+ "27": "Late Baroque (Rococo) art",
33
+ "28": "Latin American Modernism Art",
34
+ "29": "Latin American muralism",
35
+ "3": "Ancient Egyptian and Nubian Art",
36
+ "30": "Mannerism paintings",
37
+ "31": "Medieval art",
38
+ "32": "Minimalism paintings",
39
+ "33": "Modernism artwork",
40
+ "34": "Naive art",
41
+ "35": "Neo-expressionism art",
42
+ "36": "Neoclassical art",
43
+ "37": "Ottoman Art",
44
+ "38": "Polynesian Art",
45
+ "39": "Pop art",
46
+ "4": "Ancient and Traditional Chinese Art",
47
+ "40": "Realism paintings",
48
+ "41": "Renaissance paintings",
49
+ "42": "Romanticism paintings",
50
+ "43": "Surrealist paintings",
51
+ "44": "cubism",
52
+ "45": "expressionism",
53
+ "46": "romanticism",
54
+ "5": "Anime art",
55
+ "6": "Art Nouveau paintings",
56
+ "7": "Art deco paintings",
57
+ "8": "Baroque paintings",
58
+ "9": "Cave Art"
59
+ },
60
+ "image_size": 224,
61
+ "initializer_range": 0.02,
62
+ "intermediate_size": 3072,
63
+ "label2id": {
64
+ "Abstract art": "0",
65
+ "African Art": "1",
66
+ "Amazonian Art": "2",
67
+ "Ancient Egyptian and Nubian Art": "3",
68
+ "Ancient and Traditional Chinese Art": "4",
69
+ "Anime art": "5",
70
+ "Art Nouveau paintings": "6",
71
+ "Art deco paintings": "7",
72
+ "Baroque paintings": "8",
73
+ "Cave Art": "9",
74
+ "Constructivism art": "10",
75
+ "Constructivsm art": "11",
76
+ "Contemporary art": "12",
77
+ "Cubism art": "13",
78
+ "Ethiopian Art": "14",
79
+ "Expressionism paintings": "15",
80
+ "Fauvism paintings": "16",
81
+ "Gothic art": "17",
82
+ "Graffiti": "18",
83
+ "Impressionism and Post-Impressionism": "19",
84
+ "Indochristian art": "20",
85
+ "Islamic Art": "21",
86
+ "Japanese Edo Period": "22",
87
+ "Japanese Heian Period": "23",
88
+ "Japanese momoyama": "24",
89
+ "Kamakura or Muromachi Art": "25",
90
+ "Korean Goryeo Dynasty Art": "26",
91
+ "Late Baroque (Rococo) art": "27",
92
+ "Latin American Modernism Art": "28",
93
+ "Latin American muralism": "29",
94
+ "Mannerism paintings": "30",
95
+ "Medieval art": "31",
96
+ "Minimalism paintings": "32",
97
+ "Modernism artwork": "33",
98
+ "Naive art": "34",
99
+ "Neo-expressionism art": "35",
100
+ "Neoclassical art": "36",
101
+ "Ottoman Art": "37",
102
+ "Polynesian Art": "38",
103
+ "Pop art": "39",
104
+ "Realism paintings": "40",
105
+ "Renaissance paintings": "41",
106
+ "Romanticism paintings": "42",
107
+ "Surrealist paintings": "43",
108
+ "cubism": "44",
109
+ "expressionism": "45",
110
+ "romanticism": "46"
111
+ },
112
+ "layer_norm_eps": 1e-12,
113
+ "model_type": "vit",
114
+ "num_attention_heads": 12,
115
+ "num_channels": 3,
116
+ "num_hidden_layers": 12,
117
+ "patch_size": 16,
118
+ "problem_type": "single_label_classification",
119
+ "qkv_bias": true,
120
+ "torch_dtype": "float32",
121
+ "transformers_version": "4.31.0"
122
+ }
checkpoint-2920/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d2ffcdec4b4d6735cf640503f044248ea48e1e1927732ae2d31ada3c0f1a057
3
+ size 686795653
checkpoint-2920/preprocessor_config.json ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "do_rescale": true,
4
+ "do_resize": true,
5
+ "image_mean": [
6
+ 0.5,
7
+ 0.5,
8
+ 0.5
9
+ ],
10
+ "image_processor_type": "ViTFeatureExtractor",
11
+ "image_std": [
12
+ 0.5,
13
+ 0.5,
14
+ 0.5
15
+ ],
16
+ "resample": 2,
17
+ "rescale_factor": 0.00392156862745098,
18
+ "size": {
19
+ "height": 224,
20
+ "width": 224
21
+ }
22
+ }
checkpoint-2920/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1545c27e8c73d20ca1d5b8dea2610dcef74f72a43680440e8b3c4e33421c4632
3
+ size 343407149
checkpoint-2920/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dcd2c3745cefd1335df4f45fe2337cefdda35aab08263cd6c271474c6117e66c
3
+ size 14575
checkpoint-2920/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62bc9a91a4d0a3541524e00ac6e69471cc02f3c884f75b3e1904fb84865b1bab
3
+ size 627
checkpoint-2920/trainer_state.json ADDED
@@ -0,0 +1,2425 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 1.8420542478561401,
3
+ "best_model_checkpoint": "./vit-base-beans/checkpoint-640",
4
+ "epoch": 14.82233502538071,
5
+ "global_step": 2920,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.05,
12
+ "learning_rate": 0.00019932318104906937,
13
+ "loss": 3.8192,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.1,
18
+ "learning_rate": 0.00019864636209813877,
19
+ "loss": 3.7847,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.15,
24
+ "learning_rate": 0.00019796954314720813,
25
+ "loss": 3.6849,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.2,
30
+ "learning_rate": 0.00019729272419627753,
31
+ "loss": 3.6122,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.2,
36
+ "eval_accuracy": 0.10967741935483871,
37
+ "eval_loss": 3.561206340789795,
38
+ "eval_runtime": 31.2002,
39
+ "eval_samples_per_second": 54.647,
40
+ "eval_steps_per_second": 6.859,
41
+ "step": 40
42
+ },
43
+ {
44
+ "epoch": 0.25,
45
+ "learning_rate": 0.00019661590524534686,
46
+ "loss": 3.5748,
47
+ "step": 50
48
+ },
49
+ {
50
+ "epoch": 0.3,
51
+ "learning_rate": 0.00019593908629441626,
52
+ "loss": 3.491,
53
+ "step": 60
54
+ },
55
+ {
56
+ "epoch": 0.36,
57
+ "learning_rate": 0.00019526226734348562,
58
+ "loss": 3.3556,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.41,
63
+ "learning_rate": 0.000194585448392555,
64
+ "loss": 3.4008,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.41,
69
+ "eval_accuracy": 0.2475073313782991,
70
+ "eval_loss": 3.2414700984954834,
71
+ "eval_runtime": 30.544,
72
+ "eval_samples_per_second": 55.821,
73
+ "eval_steps_per_second": 7.006,
74
+ "step": 80
75
+ },
76
+ {
77
+ "epoch": 0.46,
78
+ "learning_rate": 0.00019390862944162438,
79
+ "loss": 3.2255,
80
+ "step": 90
81
+ },
82
+ {
83
+ "epoch": 0.51,
84
+ "learning_rate": 0.00019323181049069374,
85
+ "loss": 3.1171,
86
+ "step": 100
87
+ },
88
+ {
89
+ "epoch": 0.56,
90
+ "learning_rate": 0.00019255499153976314,
91
+ "loss": 3.1394,
92
+ "step": 110
93
+ },
94
+ {
95
+ "epoch": 0.61,
96
+ "learning_rate": 0.0001918781725888325,
97
+ "loss": 3.0872,
98
+ "step": 120
99
+ },
100
+ {
101
+ "epoch": 0.61,
102
+ "eval_accuracy": 0.2809384164222874,
103
+ "eval_loss": 2.9989874362945557,
104
+ "eval_runtime": 30.8814,
105
+ "eval_samples_per_second": 55.211,
106
+ "eval_steps_per_second": 6.93,
107
+ "step": 120
108
+ },
109
+ {
110
+ "epoch": 0.66,
111
+ "learning_rate": 0.00019120135363790187,
112
+ "loss": 2.9773,
113
+ "step": 130
114
+ },
115
+ {
116
+ "epoch": 0.71,
117
+ "learning_rate": 0.00019052453468697123,
118
+ "loss": 2.7993,
119
+ "step": 140
120
+ },
121
+ {
122
+ "epoch": 0.76,
123
+ "learning_rate": 0.0001898477157360406,
124
+ "loss": 2.8672,
125
+ "step": 150
126
+ },
127
+ {
128
+ "epoch": 0.81,
129
+ "learning_rate": 0.00018917089678511,
130
+ "loss": 2.8733,
131
+ "step": 160
132
+ },
133
+ {
134
+ "epoch": 0.81,
135
+ "eval_accuracy": 0.34252199413489737,
136
+ "eval_loss": 2.7833735942840576,
137
+ "eval_runtime": 30.3566,
138
+ "eval_samples_per_second": 56.166,
139
+ "eval_steps_per_second": 7.05,
140
+ "step": 160
141
+ },
142
+ {
143
+ "epoch": 0.86,
144
+ "learning_rate": 0.00018849407783417936,
145
+ "loss": 2.7912,
146
+ "step": 170
147
+ },
148
+ {
149
+ "epoch": 0.91,
150
+ "learning_rate": 0.00018781725888324875,
151
+ "loss": 2.7577,
152
+ "step": 180
153
+ },
154
+ {
155
+ "epoch": 0.96,
156
+ "learning_rate": 0.00018714043993231811,
157
+ "loss": 2.7865,
158
+ "step": 190
159
+ },
160
+ {
161
+ "epoch": 1.02,
162
+ "learning_rate": 0.00018646362098138748,
163
+ "loss": 2.591,
164
+ "step": 200
165
+ },
166
+ {
167
+ "epoch": 1.02,
168
+ "eval_accuracy": 0.37360703812316715,
169
+ "eval_loss": 2.6250743865966797,
170
+ "eval_runtime": 30.4336,
171
+ "eval_samples_per_second": 56.024,
172
+ "eval_steps_per_second": 7.032,
173
+ "step": 200
174
+ },
175
+ {
176
+ "epoch": 1.07,
177
+ "learning_rate": 0.00018578680203045687,
178
+ "loss": 2.3103,
179
+ "step": 210
180
+ },
181
+ {
182
+ "epoch": 1.12,
183
+ "learning_rate": 0.00018510998307952624,
184
+ "loss": 2.2942,
185
+ "step": 220
186
+ },
187
+ {
188
+ "epoch": 1.17,
189
+ "learning_rate": 0.0001844331641285956,
190
+ "loss": 2.3883,
191
+ "step": 230
192
+ },
193
+ {
194
+ "epoch": 1.22,
195
+ "learning_rate": 0.00018375634517766497,
196
+ "loss": 2.1643,
197
+ "step": 240
198
+ },
199
+ {
200
+ "epoch": 1.22,
201
+ "eval_accuracy": 0.40527859237536656,
202
+ "eval_loss": 2.423466920852661,
203
+ "eval_runtime": 30.3039,
204
+ "eval_samples_per_second": 56.263,
205
+ "eval_steps_per_second": 7.062,
206
+ "step": 240
207
+ },
208
+ {
209
+ "epoch": 1.27,
210
+ "learning_rate": 0.00018307952622673436,
211
+ "loss": 2.2077,
212
+ "step": 250
213
+ },
214
+ {
215
+ "epoch": 1.32,
216
+ "learning_rate": 0.00018240270727580373,
217
+ "loss": 2.1227,
218
+ "step": 260
219
+ },
220
+ {
221
+ "epoch": 1.37,
222
+ "learning_rate": 0.0001817258883248731,
223
+ "loss": 2.0657,
224
+ "step": 270
225
+ },
226
+ {
227
+ "epoch": 1.42,
228
+ "learning_rate": 0.00018104906937394248,
229
+ "loss": 2.0669,
230
+ "step": 280
231
+ },
232
+ {
233
+ "epoch": 1.42,
234
+ "eval_accuracy": 0.3882697947214076,
235
+ "eval_loss": 2.3226494789123535,
236
+ "eval_runtime": 30.2544,
237
+ "eval_samples_per_second": 56.355,
238
+ "eval_steps_per_second": 7.073,
239
+ "step": 280
240
+ },
241
+ {
242
+ "epoch": 1.47,
243
+ "learning_rate": 0.00018037225042301185,
244
+ "loss": 2.1189,
245
+ "step": 290
246
+ },
247
+ {
248
+ "epoch": 1.52,
249
+ "learning_rate": 0.00017969543147208124,
250
+ "loss": 2.1857,
251
+ "step": 300
252
+ },
253
+ {
254
+ "epoch": 1.57,
255
+ "learning_rate": 0.0001790186125211506,
256
+ "loss": 2.272,
257
+ "step": 310
258
+ },
259
+ {
260
+ "epoch": 1.62,
261
+ "learning_rate": 0.00017834179357021997,
262
+ "loss": 1.9717,
263
+ "step": 320
264
+ },
265
+ {
266
+ "epoch": 1.62,
267
+ "eval_accuracy": 0.40821114369501466,
268
+ "eval_loss": 2.2509706020355225,
269
+ "eval_runtime": 30.2595,
270
+ "eval_samples_per_second": 56.346,
271
+ "eval_steps_per_second": 7.072,
272
+ "step": 320
273
+ },
274
+ {
275
+ "epoch": 1.68,
276
+ "learning_rate": 0.00017766497461928934,
277
+ "loss": 2.0974,
278
+ "step": 330
279
+ },
280
+ {
281
+ "epoch": 1.73,
282
+ "learning_rate": 0.0001769881556683587,
283
+ "loss": 2.1316,
284
+ "step": 340
285
+ },
286
+ {
287
+ "epoch": 1.78,
288
+ "learning_rate": 0.0001763113367174281,
289
+ "loss": 2.0845,
290
+ "step": 350
291
+ },
292
+ {
293
+ "epoch": 1.83,
294
+ "learning_rate": 0.00017563451776649746,
295
+ "loss": 2.1714,
296
+ "step": 360
297
+ },
298
+ {
299
+ "epoch": 1.83,
300
+ "eval_accuracy": 0.4351906158357771,
301
+ "eval_loss": 2.203674077987671,
302
+ "eval_runtime": 30.4099,
303
+ "eval_samples_per_second": 56.067,
304
+ "eval_steps_per_second": 7.037,
305
+ "step": 360
306
+ },
307
+ {
308
+ "epoch": 1.88,
309
+ "learning_rate": 0.00017495769881556685,
310
+ "loss": 2.0881,
311
+ "step": 370
312
+ },
313
+ {
314
+ "epoch": 1.93,
315
+ "learning_rate": 0.00017428087986463622,
316
+ "loss": 1.9428,
317
+ "step": 380
318
+ },
319
+ {
320
+ "epoch": 1.98,
321
+ "learning_rate": 0.0001736040609137056,
322
+ "loss": 1.8203,
323
+ "step": 390
324
+ },
325
+ {
326
+ "epoch": 2.03,
327
+ "learning_rate": 0.00017292724196277498,
328
+ "loss": 1.6183,
329
+ "step": 400
330
+ },
331
+ {
332
+ "epoch": 2.03,
333
+ "eval_accuracy": 0.4533724340175953,
334
+ "eval_loss": 2.0731186866760254,
335
+ "eval_runtime": 30.3675,
336
+ "eval_samples_per_second": 56.146,
337
+ "eval_steps_per_second": 7.047,
338
+ "step": 400
339
+ },
340
+ {
341
+ "epoch": 2.08,
342
+ "learning_rate": 0.00017225042301184434,
343
+ "loss": 1.3395,
344
+ "step": 410
345
+ },
346
+ {
347
+ "epoch": 2.13,
348
+ "learning_rate": 0.0001715736040609137,
349
+ "loss": 1.4052,
350
+ "step": 420
351
+ },
352
+ {
353
+ "epoch": 2.18,
354
+ "learning_rate": 0.00017089678510998307,
355
+ "loss": 1.3804,
356
+ "step": 430
357
+ },
358
+ {
359
+ "epoch": 2.23,
360
+ "learning_rate": 0.00017021996615905247,
361
+ "loss": 1.2771,
362
+ "step": 440
363
+ },
364
+ {
365
+ "epoch": 2.23,
366
+ "eval_accuracy": 0.42052785923753666,
367
+ "eval_loss": 2.0928714275360107,
368
+ "eval_runtime": 30.2484,
369
+ "eval_samples_per_second": 56.367,
370
+ "eval_steps_per_second": 7.075,
371
+ "step": 440
372
+ },
373
+ {
374
+ "epoch": 2.28,
375
+ "learning_rate": 0.00016954314720812183,
376
+ "loss": 1.431,
377
+ "step": 450
378
+ },
379
+ {
380
+ "epoch": 2.34,
381
+ "learning_rate": 0.00016886632825719122,
382
+ "loss": 1.353,
383
+ "step": 460
384
+ },
385
+ {
386
+ "epoch": 2.39,
387
+ "learning_rate": 0.0001681895093062606,
388
+ "loss": 1.307,
389
+ "step": 470
390
+ },
391
+ {
392
+ "epoch": 2.44,
393
+ "learning_rate": 0.00016751269035532995,
394
+ "loss": 1.2116,
395
+ "step": 480
396
+ },
397
+ {
398
+ "epoch": 2.44,
399
+ "eval_accuracy": 0.4504398826979472,
400
+ "eval_loss": 2.0171947479248047,
401
+ "eval_runtime": 30.2989,
402
+ "eval_samples_per_second": 56.273,
403
+ "eval_steps_per_second": 7.063,
404
+ "step": 480
405
+ },
406
+ {
407
+ "epoch": 2.49,
408
+ "learning_rate": 0.00016683587140439935,
409
+ "loss": 1.2762,
410
+ "step": 490
411
+ },
412
+ {
413
+ "epoch": 2.54,
414
+ "learning_rate": 0.0001661590524534687,
415
+ "loss": 1.2253,
416
+ "step": 500
417
+ },
418
+ {
419
+ "epoch": 2.59,
420
+ "learning_rate": 0.00016548223350253808,
421
+ "loss": 1.238,
422
+ "step": 510
423
+ },
424
+ {
425
+ "epoch": 2.64,
426
+ "learning_rate": 0.00016480541455160744,
427
+ "loss": 1.158,
428
+ "step": 520
429
+ },
430
+ {
431
+ "epoch": 2.64,
432
+ "eval_accuracy": 0.45689149560117304,
433
+ "eval_loss": 1.9670987129211426,
434
+ "eval_runtime": 30.2767,
435
+ "eval_samples_per_second": 56.314,
436
+ "eval_steps_per_second": 7.068,
437
+ "step": 520
438
+ },
439
+ {
440
+ "epoch": 2.69,
441
+ "learning_rate": 0.00016412859560067684,
442
+ "loss": 1.1488,
443
+ "step": 530
444
+ },
445
+ {
446
+ "epoch": 2.74,
447
+ "learning_rate": 0.0001634517766497462,
448
+ "loss": 1.294,
449
+ "step": 540
450
+ },
451
+ {
452
+ "epoch": 2.79,
453
+ "learning_rate": 0.00016277495769881557,
454
+ "loss": 1.0431,
455
+ "step": 550
456
+ },
457
+ {
458
+ "epoch": 2.84,
459
+ "learning_rate": 0.00016209813874788496,
460
+ "loss": 1.1864,
461
+ "step": 560
462
+ },
463
+ {
464
+ "epoch": 2.84,
465
+ "eval_accuracy": 0.4563049853372434,
466
+ "eval_loss": 1.9570399522781372,
467
+ "eval_runtime": 30.2455,
468
+ "eval_samples_per_second": 56.372,
469
+ "eval_steps_per_second": 7.075,
470
+ "step": 560
471
+ },
472
+ {
473
+ "epoch": 2.89,
474
+ "learning_rate": 0.00016142131979695432,
475
+ "loss": 1.1271,
476
+ "step": 570
477
+ },
478
+ {
479
+ "epoch": 2.94,
480
+ "learning_rate": 0.00016074450084602372,
481
+ "loss": 1.1006,
482
+ "step": 580
483
+ },
484
+ {
485
+ "epoch": 2.99,
486
+ "learning_rate": 0.00016006768189509308,
487
+ "loss": 1.1867,
488
+ "step": 590
489
+ },
490
+ {
491
+ "epoch": 3.05,
492
+ "learning_rate": 0.00015939086294416242,
493
+ "loss": 0.7778,
494
+ "step": 600
495
+ },
496
+ {
497
+ "epoch": 3.05,
498
+ "eval_accuracy": 0.49560117302052786,
499
+ "eval_loss": 1.8706681728363037,
500
+ "eval_runtime": 30.3606,
501
+ "eval_samples_per_second": 56.158,
502
+ "eval_steps_per_second": 7.049,
503
+ "step": 600
504
+ },
505
+ {
506
+ "epoch": 3.1,
507
+ "learning_rate": 0.0001587140439932318,
508
+ "loss": 0.6935,
509
+ "step": 610
510
+ },
511
+ {
512
+ "epoch": 3.15,
513
+ "learning_rate": 0.00015803722504230118,
514
+ "loss": 0.5692,
515
+ "step": 620
516
+ },
517
+ {
518
+ "epoch": 3.2,
519
+ "learning_rate": 0.00015736040609137057,
520
+ "loss": 0.5985,
521
+ "step": 630
522
+ },
523
+ {
524
+ "epoch": 3.25,
525
+ "learning_rate": 0.00015668358714043994,
526
+ "loss": 0.6021,
527
+ "step": 640
528
+ },
529
+ {
530
+ "epoch": 3.25,
531
+ "eval_accuracy": 0.4950146627565982,
532
+ "eval_loss": 1.8420542478561401,
533
+ "eval_runtime": 30.3002,
534
+ "eval_samples_per_second": 56.27,
535
+ "eval_steps_per_second": 7.063,
536
+ "step": 640
537
+ },
538
+ {
539
+ "epoch": 3.3,
540
+ "learning_rate": 0.00015600676818950933,
541
+ "loss": 0.5766,
542
+ "step": 650
543
+ },
544
+ {
545
+ "epoch": 3.35,
546
+ "learning_rate": 0.0001553299492385787,
547
+ "loss": 0.7232,
548
+ "step": 660
549
+ },
550
+ {
551
+ "epoch": 3.4,
552
+ "learning_rate": 0.00015465313028764806,
553
+ "loss": 0.5083,
554
+ "step": 670
555
+ },
556
+ {
557
+ "epoch": 3.45,
558
+ "learning_rate": 0.00015397631133671742,
559
+ "loss": 0.5362,
560
+ "step": 680
561
+ },
562
+ {
563
+ "epoch": 3.45,
564
+ "eval_accuracy": 0.4756598240469208,
565
+ "eval_loss": 1.9472501277923584,
566
+ "eval_runtime": 30.3609,
567
+ "eval_samples_per_second": 56.158,
568
+ "eval_steps_per_second": 7.049,
569
+ "step": 680
570
+ },
571
+ {
572
+ "epoch": 3.5,
573
+ "learning_rate": 0.0001532994923857868,
574
+ "loss": 0.6784,
575
+ "step": 690
576
+ },
577
+ {
578
+ "epoch": 3.55,
579
+ "learning_rate": 0.00015262267343485618,
580
+ "loss": 0.5623,
581
+ "step": 700
582
+ },
583
+ {
584
+ "epoch": 3.6,
585
+ "learning_rate": 0.00015194585448392555,
586
+ "loss": 0.5619,
587
+ "step": 710
588
+ },
589
+ {
590
+ "epoch": 3.65,
591
+ "learning_rate": 0.00015126903553299494,
592
+ "loss": 0.5469,
593
+ "step": 720
594
+ },
595
+ {
596
+ "epoch": 3.65,
597
+ "eval_accuracy": 0.4621700879765396,
598
+ "eval_loss": 1.9724935293197632,
599
+ "eval_runtime": 30.833,
600
+ "eval_samples_per_second": 55.298,
601
+ "eval_steps_per_second": 6.941,
602
+ "step": 720
603
+ },
604
+ {
605
+ "epoch": 3.71,
606
+ "learning_rate": 0.0001505922165820643,
607
+ "loss": 0.6058,
608
+ "step": 730
609
+ },
610
+ {
611
+ "epoch": 3.76,
612
+ "learning_rate": 0.00014991539763113367,
613
+ "loss": 0.5332,
614
+ "step": 740
615
+ },
616
+ {
617
+ "epoch": 3.81,
618
+ "learning_rate": 0.00014923857868020306,
619
+ "loss": 0.5891,
620
+ "step": 750
621
+ },
622
+ {
623
+ "epoch": 3.86,
624
+ "learning_rate": 0.00014856175972927243,
625
+ "loss": 0.5486,
626
+ "step": 760
627
+ },
628
+ {
629
+ "epoch": 3.86,
630
+ "eval_accuracy": 0.4744868035190616,
631
+ "eval_loss": 1.9740947484970093,
632
+ "eval_runtime": 30.1738,
633
+ "eval_samples_per_second": 56.506,
634
+ "eval_steps_per_second": 7.092,
635
+ "step": 760
636
+ },
637
+ {
638
+ "epoch": 3.91,
639
+ "learning_rate": 0.0001478849407783418,
640
+ "loss": 0.624,
641
+ "step": 770
642
+ },
643
+ {
644
+ "epoch": 3.96,
645
+ "learning_rate": 0.00014720812182741116,
646
+ "loss": 0.5719,
647
+ "step": 780
648
+ },
649
+ {
650
+ "epoch": 4.01,
651
+ "learning_rate": 0.00014653130287648055,
652
+ "loss": 0.6461,
653
+ "step": 790
654
+ },
655
+ {
656
+ "epoch": 4.06,
657
+ "learning_rate": 0.00014585448392554992,
658
+ "loss": 0.339,
659
+ "step": 800
660
+ },
661
+ {
662
+ "epoch": 4.06,
663
+ "eval_accuracy": 0.47038123167155427,
664
+ "eval_loss": 1.9945420026779175,
665
+ "eval_runtime": 30.3298,
666
+ "eval_samples_per_second": 56.215,
667
+ "eval_steps_per_second": 7.056,
668
+ "step": 800
669
+ },
670
+ {
671
+ "epoch": 4.11,
672
+ "learning_rate": 0.00014517766497461928,
673
+ "loss": 0.2895,
674
+ "step": 810
675
+ },
676
+ {
677
+ "epoch": 4.16,
678
+ "learning_rate": 0.00014450084602368868,
679
+ "loss": 0.3324,
680
+ "step": 820
681
+ },
682
+ {
683
+ "epoch": 4.21,
684
+ "learning_rate": 0.00014382402707275804,
685
+ "loss": 0.3296,
686
+ "step": 830
687
+ },
688
+ {
689
+ "epoch": 4.26,
690
+ "learning_rate": 0.00014314720812182743,
691
+ "loss": 0.2798,
692
+ "step": 840
693
+ },
694
+ {
695
+ "epoch": 4.26,
696
+ "eval_accuracy": 0.46686217008797654,
697
+ "eval_loss": 2.03950572013855,
698
+ "eval_runtime": 30.2799,
699
+ "eval_samples_per_second": 56.308,
700
+ "eval_steps_per_second": 7.067,
701
+ "step": 840
702
+ },
703
+ {
704
+ "epoch": 4.31,
705
+ "learning_rate": 0.0001424703891708968,
706
+ "loss": 0.254,
707
+ "step": 850
708
+ },
709
+ {
710
+ "epoch": 4.37,
711
+ "learning_rate": 0.00014179357021996616,
712
+ "loss": 0.2533,
713
+ "step": 860
714
+ },
715
+ {
716
+ "epoch": 4.42,
717
+ "learning_rate": 0.00014111675126903553,
718
+ "loss": 0.1897,
719
+ "step": 870
720
+ },
721
+ {
722
+ "epoch": 4.47,
723
+ "learning_rate": 0.0001404399323181049,
724
+ "loss": 0.2813,
725
+ "step": 880
726
+ },
727
+ {
728
+ "epoch": 4.47,
729
+ "eval_accuracy": 0.47390029325513194,
730
+ "eval_loss": 2.0258262157440186,
731
+ "eval_runtime": 30.5513,
732
+ "eval_samples_per_second": 55.808,
733
+ "eval_steps_per_second": 7.005,
734
+ "step": 880
735
+ },
736
+ {
737
+ "epoch": 4.52,
738
+ "learning_rate": 0.0001397631133671743,
739
+ "loss": 0.3286,
740
+ "step": 890
741
+ },
742
+ {
743
+ "epoch": 4.57,
744
+ "learning_rate": 0.00013908629441624365,
745
+ "loss": 0.2575,
746
+ "step": 900
747
+ },
748
+ {
749
+ "epoch": 4.62,
750
+ "learning_rate": 0.00013840947546531305,
751
+ "loss": 0.201,
752
+ "step": 910
753
+ },
754
+ {
755
+ "epoch": 4.67,
756
+ "learning_rate": 0.0001377326565143824,
757
+ "loss": 0.209,
758
+ "step": 920
759
+ },
760
+ {
761
+ "epoch": 4.67,
762
+ "eval_accuracy": 0.47624633431085045,
763
+ "eval_loss": 2.0693771839141846,
764
+ "eval_runtime": 30.3252,
765
+ "eval_samples_per_second": 56.224,
766
+ "eval_steps_per_second": 7.057,
767
+ "step": 920
768
+ },
769
+ {
770
+ "epoch": 4.72,
771
+ "learning_rate": 0.00013705583756345178,
772
+ "loss": 0.2873,
773
+ "step": 930
774
+ },
775
+ {
776
+ "epoch": 4.77,
777
+ "learning_rate": 0.00013637901861252117,
778
+ "loss": 0.33,
779
+ "step": 940
780
+ },
781
+ {
782
+ "epoch": 4.82,
783
+ "learning_rate": 0.00013570219966159053,
784
+ "loss": 0.2578,
785
+ "step": 950
786
+ },
787
+ {
788
+ "epoch": 4.87,
789
+ "learning_rate": 0.0001350253807106599,
790
+ "loss": 0.3733,
791
+ "step": 960
792
+ },
793
+ {
794
+ "epoch": 4.87,
795
+ "eval_accuracy": 0.4797653958944281,
796
+ "eval_loss": 2.0537850856781006,
797
+ "eval_runtime": 30.0679,
798
+ "eval_samples_per_second": 56.705,
799
+ "eval_steps_per_second": 7.117,
800
+ "step": 960
801
+ },
802
+ {
803
+ "epoch": 4.92,
804
+ "learning_rate": 0.00013434856175972926,
805
+ "loss": 0.2922,
806
+ "step": 970
807
+ },
808
+ {
809
+ "epoch": 4.97,
810
+ "learning_rate": 0.00013367174280879866,
811
+ "loss": 0.2236,
812
+ "step": 980
813
+ },
814
+ {
815
+ "epoch": 5.03,
816
+ "learning_rate": 0.00013299492385786802,
817
+ "loss": 0.2532,
818
+ "step": 990
819
+ },
820
+ {
821
+ "epoch": 5.08,
822
+ "learning_rate": 0.0001323181049069374,
823
+ "loss": 0.1729,
824
+ "step": 1000
825
+ },
826
+ {
827
+ "epoch": 5.08,
828
+ "eval_accuracy": 0.4563049853372434,
829
+ "eval_loss": 2.179201126098633,
830
+ "eval_runtime": 30.3122,
831
+ "eval_samples_per_second": 56.248,
832
+ "eval_steps_per_second": 7.06,
833
+ "step": 1000
834
+ },
835
+ {
836
+ "epoch": 5.13,
837
+ "learning_rate": 0.00013164128595600678,
838
+ "loss": 0.1751,
839
+ "step": 1010
840
+ },
841
+ {
842
+ "epoch": 5.18,
843
+ "learning_rate": 0.00013096446700507615,
844
+ "loss": 0.1609,
845
+ "step": 1020
846
+ },
847
+ {
848
+ "epoch": 5.23,
849
+ "learning_rate": 0.00013028764805414554,
850
+ "loss": 0.1254,
851
+ "step": 1030
852
+ },
853
+ {
854
+ "epoch": 5.28,
855
+ "learning_rate": 0.0001296108291032149,
856
+ "loss": 0.0891,
857
+ "step": 1040
858
+ },
859
+ {
860
+ "epoch": 5.28,
861
+ "eval_accuracy": 0.4715542521994135,
862
+ "eval_loss": 2.140515089035034,
863
+ "eval_runtime": 30.4734,
864
+ "eval_samples_per_second": 55.95,
865
+ "eval_steps_per_second": 7.023,
866
+ "step": 1040
867
+ },
868
+ {
869
+ "epoch": 5.33,
870
+ "learning_rate": 0.00012893401015228427,
871
+ "loss": 0.0719,
872
+ "step": 1050
873
+ },
874
+ {
875
+ "epoch": 5.38,
876
+ "learning_rate": 0.00012825719120135363,
877
+ "loss": 0.1418,
878
+ "step": 1060
879
+ },
880
+ {
881
+ "epoch": 5.43,
882
+ "learning_rate": 0.000127580372250423,
883
+ "loss": 0.1633,
884
+ "step": 1070
885
+ },
886
+ {
887
+ "epoch": 5.48,
888
+ "learning_rate": 0.0001269035532994924,
889
+ "loss": 0.114,
890
+ "step": 1080
891
+ },
892
+ {
893
+ "epoch": 5.48,
894
+ "eval_accuracy": 0.4774193548387097,
895
+ "eval_loss": 2.172985315322876,
896
+ "eval_runtime": 30.271,
897
+ "eval_samples_per_second": 56.325,
898
+ "eval_steps_per_second": 7.069,
899
+ "step": 1080
900
+ },
901
+ {
902
+ "epoch": 5.53,
903
+ "learning_rate": 0.00012622673434856176,
904
+ "loss": 0.1615,
905
+ "step": 1090
906
+ },
907
+ {
908
+ "epoch": 5.58,
909
+ "learning_rate": 0.00012554991539763115,
910
+ "loss": 0.1589,
911
+ "step": 1100
912
+ },
913
+ {
914
+ "epoch": 5.63,
915
+ "learning_rate": 0.00012487309644670052,
916
+ "loss": 0.1249,
917
+ "step": 1110
918
+ },
919
+ {
920
+ "epoch": 5.69,
921
+ "learning_rate": 0.0001241962774957699,
922
+ "loss": 0.207,
923
+ "step": 1120
924
+ },
925
+ {
926
+ "epoch": 5.69,
927
+ "eval_accuracy": 0.4475073313782991,
928
+ "eval_loss": 2.32855486869812,
929
+ "eval_runtime": 30.3218,
930
+ "eval_samples_per_second": 56.23,
931
+ "eval_steps_per_second": 7.058,
932
+ "step": 1120
933
+ },
934
+ {
935
+ "epoch": 5.74,
936
+ "learning_rate": 0.00012351945854483927,
937
+ "loss": 0.1255,
938
+ "step": 1130
939
+ },
940
+ {
941
+ "epoch": 5.79,
942
+ "learning_rate": 0.00012284263959390864,
943
+ "loss": 0.0847,
944
+ "step": 1140
945
+ },
946
+ {
947
+ "epoch": 5.84,
948
+ "learning_rate": 0.000122165820642978,
949
+ "loss": 0.2418,
950
+ "step": 1150
951
+ },
952
+ {
953
+ "epoch": 5.89,
954
+ "learning_rate": 0.00012148900169204738,
955
+ "loss": 0.0984,
956
+ "step": 1160
957
+ },
958
+ {
959
+ "epoch": 5.89,
960
+ "eval_accuracy": 0.4826979472140763,
961
+ "eval_loss": 2.280658006668091,
962
+ "eval_runtime": 30.3832,
963
+ "eval_samples_per_second": 56.116,
964
+ "eval_steps_per_second": 7.043,
965
+ "step": 1160
966
+ },
967
+ {
968
+ "epoch": 5.94,
969
+ "learning_rate": 0.00012081218274111676,
970
+ "loss": 0.2079,
971
+ "step": 1170
972
+ },
973
+ {
974
+ "epoch": 5.99,
975
+ "learning_rate": 0.00012013536379018613,
976
+ "loss": 0.1239,
977
+ "step": 1180
978
+ },
979
+ {
980
+ "epoch": 6.04,
981
+ "learning_rate": 0.00011945854483925552,
982
+ "loss": 0.1165,
983
+ "step": 1190
984
+ },
985
+ {
986
+ "epoch": 6.09,
987
+ "learning_rate": 0.00011878172588832489,
988
+ "loss": 0.0544,
989
+ "step": 1200
990
+ },
991
+ {
992
+ "epoch": 6.09,
993
+ "eval_accuracy": 0.44868035190615835,
994
+ "eval_loss": 2.3984222412109375,
995
+ "eval_runtime": 30.0364,
996
+ "eval_samples_per_second": 56.764,
997
+ "eval_steps_per_second": 7.125,
998
+ "step": 1200
999
+ },
1000
+ {
1001
+ "epoch": 6.14,
1002
+ "learning_rate": 0.00011810490693739425,
1003
+ "loss": 0.1029,
1004
+ "step": 1210
1005
+ },
1006
+ {
1007
+ "epoch": 6.19,
1008
+ "learning_rate": 0.00011742808798646363,
1009
+ "loss": 0.1151,
1010
+ "step": 1220
1011
+ },
1012
+ {
1013
+ "epoch": 6.24,
1014
+ "learning_rate": 0.000116751269035533,
1015
+ "loss": 0.0611,
1016
+ "step": 1230
1017
+ },
1018
+ {
1019
+ "epoch": 6.29,
1020
+ "learning_rate": 0.00011607445008460239,
1021
+ "loss": 0.1109,
1022
+ "step": 1240
1023
+ },
1024
+ {
1025
+ "epoch": 6.29,
1026
+ "eval_accuracy": 0.4780058651026393,
1027
+ "eval_loss": 2.314086437225342,
1028
+ "eval_runtime": 30.0544,
1029
+ "eval_samples_per_second": 56.73,
1030
+ "eval_steps_per_second": 7.12,
1031
+ "step": 1240
1032
+ },
1033
+ {
1034
+ "epoch": 6.35,
1035
+ "learning_rate": 0.00011539763113367175,
1036
+ "loss": 0.064,
1037
+ "step": 1250
1038
+ },
1039
+ {
1040
+ "epoch": 6.4,
1041
+ "learning_rate": 0.00011472081218274113,
1042
+ "loss": 0.0555,
1043
+ "step": 1260
1044
+ },
1045
+ {
1046
+ "epoch": 6.45,
1047
+ "learning_rate": 0.0001140439932318105,
1048
+ "loss": 0.0774,
1049
+ "step": 1270
1050
+ },
1051
+ {
1052
+ "epoch": 6.5,
1053
+ "learning_rate": 0.00011336717428087986,
1054
+ "loss": 0.121,
1055
+ "step": 1280
1056
+ },
1057
+ {
1058
+ "epoch": 6.5,
1059
+ "eval_accuracy": 0.49032258064516127,
1060
+ "eval_loss": 2.225583791732788,
1061
+ "eval_runtime": 30.2835,
1062
+ "eval_samples_per_second": 56.301,
1063
+ "eval_steps_per_second": 7.067,
1064
+ "step": 1280
1065
+ },
1066
+ {
1067
+ "epoch": 6.55,
1068
+ "learning_rate": 0.00011269035532994925,
1069
+ "loss": 0.0801,
1070
+ "step": 1290
1071
+ },
1072
+ {
1073
+ "epoch": 6.6,
1074
+ "learning_rate": 0.0001120135363790186,
1075
+ "loss": 0.045,
1076
+ "step": 1300
1077
+ },
1078
+ {
1079
+ "epoch": 6.65,
1080
+ "learning_rate": 0.000111336717428088,
1081
+ "loss": 0.0725,
1082
+ "step": 1310
1083
+ },
1084
+ {
1085
+ "epoch": 6.7,
1086
+ "learning_rate": 0.00011065989847715736,
1087
+ "loss": 0.0387,
1088
+ "step": 1320
1089
+ },
1090
+ {
1091
+ "epoch": 6.7,
1092
+ "eval_accuracy": 0.5032258064516129,
1093
+ "eval_loss": 2.191972494125366,
1094
+ "eval_runtime": 30.6293,
1095
+ "eval_samples_per_second": 55.666,
1096
+ "eval_steps_per_second": 6.987,
1097
+ "step": 1320
1098
+ },
1099
+ {
1100
+ "epoch": 6.75,
1101
+ "learning_rate": 0.00010998307952622673,
1102
+ "loss": 0.0376,
1103
+ "step": 1330
1104
+ },
1105
+ {
1106
+ "epoch": 6.8,
1107
+ "learning_rate": 0.00010930626057529611,
1108
+ "loss": 0.082,
1109
+ "step": 1340
1110
+ },
1111
+ {
1112
+ "epoch": 6.85,
1113
+ "learning_rate": 0.00010862944162436547,
1114
+ "loss": 0.026,
1115
+ "step": 1350
1116
+ },
1117
+ {
1118
+ "epoch": 6.9,
1119
+ "learning_rate": 0.00010795262267343487,
1120
+ "loss": 0.0469,
1121
+ "step": 1360
1122
+ },
1123
+ {
1124
+ "epoch": 6.9,
1125
+ "eval_accuracy": 0.481524926686217,
1126
+ "eval_loss": 2.3258745670318604,
1127
+ "eval_runtime": 30.3501,
1128
+ "eval_samples_per_second": 56.178,
1129
+ "eval_steps_per_second": 7.051,
1130
+ "step": 1360
1131
+ },
1132
+ {
1133
+ "epoch": 6.95,
1134
+ "learning_rate": 0.00010727580372250423,
1135
+ "loss": 0.0221,
1136
+ "step": 1370
1137
+ },
1138
+ {
1139
+ "epoch": 7.01,
1140
+ "learning_rate": 0.00010659898477157362,
1141
+ "loss": 0.0234,
1142
+ "step": 1380
1143
+ },
1144
+ {
1145
+ "epoch": 7.06,
1146
+ "learning_rate": 0.00010592216582064298,
1147
+ "loss": 0.0536,
1148
+ "step": 1390
1149
+ },
1150
+ {
1151
+ "epoch": 7.11,
1152
+ "learning_rate": 0.00010524534686971234,
1153
+ "loss": 0.0269,
1154
+ "step": 1400
1155
+ },
1156
+ {
1157
+ "epoch": 7.11,
1158
+ "eval_accuracy": 0.49325513196480936,
1159
+ "eval_loss": 2.3082282543182373,
1160
+ "eval_runtime": 30.3956,
1161
+ "eval_samples_per_second": 56.094,
1162
+ "eval_steps_per_second": 7.04,
1163
+ "step": 1400
1164
+ },
1165
+ {
1166
+ "epoch": 7.16,
1167
+ "learning_rate": 0.00010456852791878173,
1168
+ "loss": 0.0626,
1169
+ "step": 1410
1170
+ },
1171
+ {
1172
+ "epoch": 7.21,
1173
+ "learning_rate": 0.0001038917089678511,
1174
+ "loss": 0.0338,
1175
+ "step": 1420
1176
+ },
1177
+ {
1178
+ "epoch": 7.26,
1179
+ "learning_rate": 0.00010321489001692048,
1180
+ "loss": 0.0534,
1181
+ "step": 1430
1182
+ },
1183
+ {
1184
+ "epoch": 7.31,
1185
+ "learning_rate": 0.00010253807106598984,
1186
+ "loss": 0.0482,
1187
+ "step": 1440
1188
+ },
1189
+ {
1190
+ "epoch": 7.31,
1191
+ "eval_accuracy": 0.47624633431085045,
1192
+ "eval_loss": 2.3973476886749268,
1193
+ "eval_runtime": 30.8314,
1194
+ "eval_samples_per_second": 55.301,
1195
+ "eval_steps_per_second": 6.941,
1196
+ "step": 1440
1197
+ },
1198
+ {
1199
+ "epoch": 7.36,
1200
+ "learning_rate": 0.00010186125211505924,
1201
+ "loss": 0.043,
1202
+ "step": 1450
1203
+ },
1204
+ {
1205
+ "epoch": 7.41,
1206
+ "learning_rate": 0.0001011844331641286,
1207
+ "loss": 0.0217,
1208
+ "step": 1460
1209
+ },
1210
+ {
1211
+ "epoch": 7.46,
1212
+ "learning_rate": 0.00010050761421319797,
1213
+ "loss": 0.0253,
1214
+ "step": 1470
1215
+ },
1216
+ {
1217
+ "epoch": 7.51,
1218
+ "learning_rate": 9.983079526226735e-05,
1219
+ "loss": 0.052,
1220
+ "step": 1480
1221
+ },
1222
+ {
1223
+ "epoch": 7.51,
1224
+ "eval_accuracy": 0.4750733137829912,
1225
+ "eval_loss": 2.3328230381011963,
1226
+ "eval_runtime": 30.3845,
1227
+ "eval_samples_per_second": 56.114,
1228
+ "eval_steps_per_second": 7.043,
1229
+ "step": 1480
1230
+ },
1231
+ {
1232
+ "epoch": 7.56,
1233
+ "learning_rate": 9.915397631133673e-05,
1234
+ "loss": 0.0445,
1235
+ "step": 1490
1236
+ },
1237
+ {
1238
+ "epoch": 7.61,
1239
+ "learning_rate": 9.847715736040609e-05,
1240
+ "loss": 0.0927,
1241
+ "step": 1500
1242
+ },
1243
+ {
1244
+ "epoch": 7.66,
1245
+ "learning_rate": 9.780033840947547e-05,
1246
+ "loss": 0.0418,
1247
+ "step": 1510
1248
+ },
1249
+ {
1250
+ "epoch": 7.72,
1251
+ "learning_rate": 9.712351945854485e-05,
1252
+ "loss": 0.0342,
1253
+ "step": 1520
1254
+ },
1255
+ {
1256
+ "epoch": 7.72,
1257
+ "eval_accuracy": 0.5020527859237537,
1258
+ "eval_loss": 2.301551580429077,
1259
+ "eval_runtime": 31.0377,
1260
+ "eval_samples_per_second": 54.933,
1261
+ "eval_steps_per_second": 6.895,
1262
+ "step": 1520
1263
+ },
1264
+ {
1265
+ "epoch": 7.77,
1266
+ "learning_rate": 9.644670050761421e-05,
1267
+ "loss": 0.0925,
1268
+ "step": 1530
1269
+ },
1270
+ {
1271
+ "epoch": 7.82,
1272
+ "learning_rate": 9.576988155668359e-05,
1273
+ "loss": 0.0187,
1274
+ "step": 1540
1275
+ },
1276
+ {
1277
+ "epoch": 7.87,
1278
+ "learning_rate": 9.509306260575297e-05,
1279
+ "loss": 0.0816,
1280
+ "step": 1550
1281
+ },
1282
+ {
1283
+ "epoch": 7.92,
1284
+ "learning_rate": 9.441624365482235e-05,
1285
+ "loss": 0.0374,
1286
+ "step": 1560
1287
+ },
1288
+ {
1289
+ "epoch": 7.92,
1290
+ "eval_accuracy": 0.49736070381231673,
1291
+ "eval_loss": 2.250404119491577,
1292
+ "eval_runtime": 31.1263,
1293
+ "eval_samples_per_second": 54.777,
1294
+ "eval_steps_per_second": 6.875,
1295
+ "step": 1560
1296
+ },
1297
+ {
1298
+ "epoch": 7.97,
1299
+ "learning_rate": 9.373942470389172e-05,
1300
+ "loss": 0.0521,
1301
+ "step": 1570
1302
+ },
1303
+ {
1304
+ "epoch": 8.02,
1305
+ "learning_rate": 9.306260575296108e-05,
1306
+ "loss": 0.0443,
1307
+ "step": 1580
1308
+ },
1309
+ {
1310
+ "epoch": 8.07,
1311
+ "learning_rate": 9.238578680203046e-05,
1312
+ "loss": 0.019,
1313
+ "step": 1590
1314
+ },
1315
+ {
1316
+ "epoch": 8.12,
1317
+ "learning_rate": 9.170896785109984e-05,
1318
+ "loss": 0.0169,
1319
+ "step": 1600
1320
+ },
1321
+ {
1322
+ "epoch": 8.12,
1323
+ "eval_accuracy": 0.48328445747800586,
1324
+ "eval_loss": 2.3862178325653076,
1325
+ "eval_runtime": 30.7127,
1326
+ "eval_samples_per_second": 55.514,
1327
+ "eval_steps_per_second": 6.968,
1328
+ "step": 1600
1329
+ },
1330
+ {
1331
+ "epoch": 8.17,
1332
+ "learning_rate": 9.103214890016922e-05,
1333
+ "loss": 0.0239,
1334
+ "step": 1610
1335
+ },
1336
+ {
1337
+ "epoch": 8.22,
1338
+ "learning_rate": 9.035532994923858e-05,
1339
+ "loss": 0.0145,
1340
+ "step": 1620
1341
+ },
1342
+ {
1343
+ "epoch": 8.27,
1344
+ "learning_rate": 8.967851099830795e-05,
1345
+ "loss": 0.0398,
1346
+ "step": 1630
1347
+ },
1348
+ {
1349
+ "epoch": 8.32,
1350
+ "learning_rate": 8.900169204737733e-05,
1351
+ "loss": 0.084,
1352
+ "step": 1640
1353
+ },
1354
+ {
1355
+ "epoch": 8.32,
1356
+ "eval_accuracy": 0.49442815249266864,
1357
+ "eval_loss": 2.4026331901550293,
1358
+ "eval_runtime": 30.8185,
1359
+ "eval_samples_per_second": 55.324,
1360
+ "eval_steps_per_second": 6.944,
1361
+ "step": 1640
1362
+ },
1363
+ {
1364
+ "epoch": 8.38,
1365
+ "learning_rate": 8.83248730964467e-05,
1366
+ "loss": 0.0215,
1367
+ "step": 1650
1368
+ },
1369
+ {
1370
+ "epoch": 8.43,
1371
+ "learning_rate": 8.764805414551607e-05,
1372
+ "loss": 0.0115,
1373
+ "step": 1660
1374
+ },
1375
+ {
1376
+ "epoch": 8.48,
1377
+ "learning_rate": 8.697123519458545e-05,
1378
+ "loss": 0.0331,
1379
+ "step": 1670
1380
+ },
1381
+ {
1382
+ "epoch": 8.53,
1383
+ "learning_rate": 8.629441624365483e-05,
1384
+ "loss": 0.0115,
1385
+ "step": 1680
1386
+ },
1387
+ {
1388
+ "epoch": 8.53,
1389
+ "eval_accuracy": 0.49149560117302055,
1390
+ "eval_loss": 2.382765293121338,
1391
+ "eval_runtime": 30.6052,
1392
+ "eval_samples_per_second": 55.71,
1393
+ "eval_steps_per_second": 6.992,
1394
+ "step": 1680
1395
+ },
1396
+ {
1397
+ "epoch": 8.58,
1398
+ "learning_rate": 8.561759729272421e-05,
1399
+ "loss": 0.0304,
1400
+ "step": 1690
1401
+ },
1402
+ {
1403
+ "epoch": 8.63,
1404
+ "learning_rate": 8.494077834179357e-05,
1405
+ "loss": 0.0108,
1406
+ "step": 1700
1407
+ },
1408
+ {
1409
+ "epoch": 8.68,
1410
+ "learning_rate": 8.426395939086294e-05,
1411
+ "loss": 0.0252,
1412
+ "step": 1710
1413
+ },
1414
+ {
1415
+ "epoch": 8.73,
1416
+ "learning_rate": 8.358714043993232e-05,
1417
+ "loss": 0.0285,
1418
+ "step": 1720
1419
+ },
1420
+ {
1421
+ "epoch": 8.73,
1422
+ "eval_accuracy": 0.49560117302052786,
1423
+ "eval_loss": 2.352389097213745,
1424
+ "eval_runtime": 30.3007,
1425
+ "eval_samples_per_second": 56.269,
1426
+ "eval_steps_per_second": 7.063,
1427
+ "step": 1720
1428
+ },
1429
+ {
1430
+ "epoch": 8.78,
1431
+ "learning_rate": 8.29103214890017e-05,
1432
+ "loss": 0.0605,
1433
+ "step": 1730
1434
+ },
1435
+ {
1436
+ "epoch": 8.83,
1437
+ "learning_rate": 8.223350253807108e-05,
1438
+ "loss": 0.0314,
1439
+ "step": 1740
1440
+ },
1441
+ {
1442
+ "epoch": 8.88,
1443
+ "learning_rate": 8.155668358714044e-05,
1444
+ "loss": 0.0135,
1445
+ "step": 1750
1446
+ },
1447
+ {
1448
+ "epoch": 8.93,
1449
+ "learning_rate": 8.087986463620981e-05,
1450
+ "loss": 0.1062,
1451
+ "step": 1760
1452
+ },
1453
+ {
1454
+ "epoch": 8.93,
1455
+ "eval_accuracy": 0.49149560117302055,
1456
+ "eval_loss": 2.412898063659668,
1457
+ "eval_runtime": 30.498,
1458
+ "eval_samples_per_second": 55.905,
1459
+ "eval_steps_per_second": 7.017,
1460
+ "step": 1760
1461
+ },
1462
+ {
1463
+ "epoch": 8.98,
1464
+ "learning_rate": 8.020304568527919e-05,
1465
+ "loss": 0.025,
1466
+ "step": 1770
1467
+ },
1468
+ {
1469
+ "epoch": 9.04,
1470
+ "learning_rate": 7.952622673434857e-05,
1471
+ "loss": 0.0383,
1472
+ "step": 1780
1473
+ },
1474
+ {
1475
+ "epoch": 9.09,
1476
+ "learning_rate": 7.884940778341794e-05,
1477
+ "loss": 0.0116,
1478
+ "step": 1790
1479
+ },
1480
+ {
1481
+ "epoch": 9.14,
1482
+ "learning_rate": 7.817258883248731e-05,
1483
+ "loss": 0.0113,
1484
+ "step": 1800
1485
+ },
1486
+ {
1487
+ "epoch": 9.14,
1488
+ "eval_accuracy": 0.49208211143695013,
1489
+ "eval_loss": 2.3613696098327637,
1490
+ "eval_runtime": 30.6596,
1491
+ "eval_samples_per_second": 55.611,
1492
+ "eval_steps_per_second": 6.98,
1493
+ "step": 1800
1494
+ },
1495
+ {
1496
+ "epoch": 9.19,
1497
+ "learning_rate": 7.749576988155669e-05,
1498
+ "loss": 0.0226,
1499
+ "step": 1810
1500
+ },
1501
+ {
1502
+ "epoch": 9.24,
1503
+ "learning_rate": 7.681895093062607e-05,
1504
+ "loss": 0.0203,
1505
+ "step": 1820
1506
+ },
1507
+ {
1508
+ "epoch": 9.29,
1509
+ "learning_rate": 7.614213197969543e-05,
1510
+ "loss": 0.0354,
1511
+ "step": 1830
1512
+ },
1513
+ {
1514
+ "epoch": 9.34,
1515
+ "learning_rate": 7.546531302876481e-05,
1516
+ "loss": 0.0101,
1517
+ "step": 1840
1518
+ },
1519
+ {
1520
+ "epoch": 9.34,
1521
+ "eval_accuracy": 0.5002932551319648,
1522
+ "eval_loss": 2.3706891536712646,
1523
+ "eval_runtime": 30.567,
1524
+ "eval_samples_per_second": 55.779,
1525
+ "eval_steps_per_second": 7.001,
1526
+ "step": 1840
1527
+ },
1528
+ {
1529
+ "epoch": 9.39,
1530
+ "learning_rate": 7.478849407783418e-05,
1531
+ "loss": 0.0181,
1532
+ "step": 1850
1533
+ },
1534
+ {
1535
+ "epoch": 9.44,
1536
+ "learning_rate": 7.411167512690356e-05,
1537
+ "loss": 0.0368,
1538
+ "step": 1860
1539
+ },
1540
+ {
1541
+ "epoch": 9.49,
1542
+ "learning_rate": 7.343485617597293e-05,
1543
+ "loss": 0.0082,
1544
+ "step": 1870
1545
+ },
1546
+ {
1547
+ "epoch": 9.54,
1548
+ "learning_rate": 7.275803722504231e-05,
1549
+ "loss": 0.028,
1550
+ "step": 1880
1551
+ },
1552
+ {
1553
+ "epoch": 9.54,
1554
+ "eval_accuracy": 0.4967741935483871,
1555
+ "eval_loss": 2.414238452911377,
1556
+ "eval_runtime": 30.6609,
1557
+ "eval_samples_per_second": 55.608,
1558
+ "eval_steps_per_second": 6.98,
1559
+ "step": 1880
1560
+ },
1561
+ {
1562
+ "epoch": 9.59,
1563
+ "learning_rate": 7.208121827411168e-05,
1564
+ "loss": 0.0404,
1565
+ "step": 1890
1566
+ },
1567
+ {
1568
+ "epoch": 9.64,
1569
+ "learning_rate": 7.140439932318104e-05,
1570
+ "loss": 0.0486,
1571
+ "step": 1900
1572
+ },
1573
+ {
1574
+ "epoch": 9.7,
1575
+ "learning_rate": 7.072758037225042e-05,
1576
+ "loss": 0.0197,
1577
+ "step": 1910
1578
+ },
1579
+ {
1580
+ "epoch": 9.75,
1581
+ "learning_rate": 7.00507614213198e-05,
1582
+ "loss": 0.0131,
1583
+ "step": 1920
1584
+ },
1585
+ {
1586
+ "epoch": 9.75,
1587
+ "eval_accuracy": 0.501466275659824,
1588
+ "eval_loss": 2.4358253479003906,
1589
+ "eval_runtime": 30.4732,
1590
+ "eval_samples_per_second": 55.951,
1591
+ "eval_steps_per_second": 7.023,
1592
+ "step": 1920
1593
+ },
1594
+ {
1595
+ "epoch": 9.8,
1596
+ "learning_rate": 6.937394247038918e-05,
1597
+ "loss": 0.0394,
1598
+ "step": 1930
1599
+ },
1600
+ {
1601
+ "epoch": 9.85,
1602
+ "learning_rate": 6.869712351945855e-05,
1603
+ "loss": 0.0402,
1604
+ "step": 1940
1605
+ },
1606
+ {
1607
+ "epoch": 9.9,
1608
+ "learning_rate": 6.802030456852793e-05,
1609
+ "loss": 0.0215,
1610
+ "step": 1950
1611
+ },
1612
+ {
1613
+ "epoch": 9.95,
1614
+ "learning_rate": 6.734348561759729e-05,
1615
+ "loss": 0.0115,
1616
+ "step": 1960
1617
+ },
1618
+ {
1619
+ "epoch": 9.95,
1620
+ "eval_accuracy": 0.4997067448680352,
1621
+ "eval_loss": 2.3999907970428467,
1622
+ "eval_runtime": 30.5014,
1623
+ "eval_samples_per_second": 55.899,
1624
+ "eval_steps_per_second": 7.016,
1625
+ "step": 1960
1626
+ },
1627
+ {
1628
+ "epoch": 10.0,
1629
+ "learning_rate": 6.666666666666667e-05,
1630
+ "loss": 0.0148,
1631
+ "step": 1970
1632
+ },
1633
+ {
1634
+ "epoch": 10.05,
1635
+ "learning_rate": 6.598984771573604e-05,
1636
+ "loss": 0.0075,
1637
+ "step": 1980
1638
+ },
1639
+ {
1640
+ "epoch": 10.1,
1641
+ "learning_rate": 6.531302876480541e-05,
1642
+ "loss": 0.0232,
1643
+ "step": 1990
1644
+ },
1645
+ {
1646
+ "epoch": 10.15,
1647
+ "learning_rate": 6.463620981387479e-05,
1648
+ "loss": 0.0108,
1649
+ "step": 2000
1650
+ },
1651
+ {
1652
+ "epoch": 10.15,
1653
+ "eval_accuracy": 0.5043988269794721,
1654
+ "eval_loss": 2.407888889312744,
1655
+ "eval_runtime": 30.7257,
1656
+ "eval_samples_per_second": 55.491,
1657
+ "eval_steps_per_second": 6.965,
1658
+ "step": 2000
1659
+ },
1660
+ {
1661
+ "epoch": 10.2,
1662
+ "learning_rate": 6.395939086294417e-05,
1663
+ "loss": 0.0182,
1664
+ "step": 2010
1665
+ },
1666
+ {
1667
+ "epoch": 10.25,
1668
+ "learning_rate": 6.328257191201354e-05,
1669
+ "loss": 0.0219,
1670
+ "step": 2020
1671
+ },
1672
+ {
1673
+ "epoch": 10.3,
1674
+ "learning_rate": 6.26057529610829e-05,
1675
+ "loss": 0.0277,
1676
+ "step": 2030
1677
+ },
1678
+ {
1679
+ "epoch": 10.36,
1680
+ "learning_rate": 6.192893401015228e-05,
1681
+ "loss": 0.009,
1682
+ "step": 2040
1683
+ },
1684
+ {
1685
+ "epoch": 10.36,
1686
+ "eval_accuracy": 0.5067448680351906,
1687
+ "eval_loss": 2.3866255283355713,
1688
+ "eval_runtime": 30.7008,
1689
+ "eval_samples_per_second": 55.536,
1690
+ "eval_steps_per_second": 6.971,
1691
+ "step": 2040
1692
+ },
1693
+ {
1694
+ "epoch": 10.41,
1695
+ "learning_rate": 6.125211505922166e-05,
1696
+ "loss": 0.0212,
1697
+ "step": 2050
1698
+ },
1699
+ {
1700
+ "epoch": 10.46,
1701
+ "learning_rate": 6.057529610829103e-05,
1702
+ "loss": 0.0065,
1703
+ "step": 2060
1704
+ },
1705
+ {
1706
+ "epoch": 10.51,
1707
+ "learning_rate": 5.989847715736041e-05,
1708
+ "loss": 0.0306,
1709
+ "step": 2070
1710
+ },
1711
+ {
1712
+ "epoch": 10.56,
1713
+ "learning_rate": 5.9221658206429784e-05,
1714
+ "loss": 0.0112,
1715
+ "step": 2080
1716
+ },
1717
+ {
1718
+ "epoch": 10.56,
1719
+ "eval_accuracy": 0.5085043988269795,
1720
+ "eval_loss": 2.3716347217559814,
1721
+ "eval_runtime": 30.6367,
1722
+ "eval_samples_per_second": 55.652,
1723
+ "eval_steps_per_second": 6.985,
1724
+ "step": 2080
1725
+ },
1726
+ {
1727
+ "epoch": 10.61,
1728
+ "learning_rate": 5.854483925549916e-05,
1729
+ "loss": 0.0325,
1730
+ "step": 2090
1731
+ },
1732
+ {
1733
+ "epoch": 10.66,
1734
+ "learning_rate": 5.786802030456853e-05,
1735
+ "loss": 0.0335,
1736
+ "step": 2100
1737
+ },
1738
+ {
1739
+ "epoch": 10.71,
1740
+ "learning_rate": 5.71912013536379e-05,
1741
+ "loss": 0.0209,
1742
+ "step": 2110
1743
+ },
1744
+ {
1745
+ "epoch": 10.76,
1746
+ "learning_rate": 5.651438240270728e-05,
1747
+ "loss": 0.0066,
1748
+ "step": 2120
1749
+ },
1750
+ {
1751
+ "epoch": 10.76,
1752
+ "eval_accuracy": 0.509090909090909,
1753
+ "eval_loss": 2.3734991550445557,
1754
+ "eval_runtime": 30.6695,
1755
+ "eval_samples_per_second": 55.593,
1756
+ "eval_steps_per_second": 6.978,
1757
+ "step": 2120
1758
+ },
1759
+ {
1760
+ "epoch": 10.81,
1761
+ "learning_rate": 5.583756345177665e-05,
1762
+ "loss": 0.0146,
1763
+ "step": 2130
1764
+ },
1765
+ {
1766
+ "epoch": 10.86,
1767
+ "learning_rate": 5.516074450084603e-05,
1768
+ "loss": 0.0061,
1769
+ "step": 2140
1770
+ },
1771
+ {
1772
+ "epoch": 10.91,
1773
+ "learning_rate": 5.44839255499154e-05,
1774
+ "loss": 0.0163,
1775
+ "step": 2150
1776
+ },
1777
+ {
1778
+ "epoch": 10.96,
1779
+ "learning_rate": 5.380710659898477e-05,
1780
+ "loss": 0.0164,
1781
+ "step": 2160
1782
+ },
1783
+ {
1784
+ "epoch": 10.96,
1785
+ "eval_accuracy": 0.5055718475073314,
1786
+ "eval_loss": 2.399704694747925,
1787
+ "eval_runtime": 30.8256,
1788
+ "eval_samples_per_second": 55.311,
1789
+ "eval_steps_per_second": 6.942,
1790
+ "step": 2160
1791
+ },
1792
+ {
1793
+ "epoch": 11.02,
1794
+ "learning_rate": 5.313028764805415e-05,
1795
+ "loss": 0.017,
1796
+ "step": 2170
1797
+ },
1798
+ {
1799
+ "epoch": 11.07,
1800
+ "learning_rate": 5.245346869712352e-05,
1801
+ "loss": 0.0056,
1802
+ "step": 2180
1803
+ },
1804
+ {
1805
+ "epoch": 11.12,
1806
+ "learning_rate": 5.17766497461929e-05,
1807
+ "loss": 0.0191,
1808
+ "step": 2190
1809
+ },
1810
+ {
1811
+ "epoch": 11.17,
1812
+ "learning_rate": 5.109983079526227e-05,
1813
+ "loss": 0.0126,
1814
+ "step": 2200
1815
+ },
1816
+ {
1817
+ "epoch": 11.17,
1818
+ "eval_accuracy": 0.5055718475073314,
1819
+ "eval_loss": 2.4104385375976562,
1820
+ "eval_runtime": 30.7309,
1821
+ "eval_samples_per_second": 55.482,
1822
+ "eval_steps_per_second": 6.964,
1823
+ "step": 2200
1824
+ },
1825
+ {
1826
+ "epoch": 11.22,
1827
+ "learning_rate": 5.042301184433165e-05,
1828
+ "loss": 0.0125,
1829
+ "step": 2210
1830
+ },
1831
+ {
1832
+ "epoch": 11.27,
1833
+ "learning_rate": 4.9746192893401014e-05,
1834
+ "loss": 0.011,
1835
+ "step": 2220
1836
+ },
1837
+ {
1838
+ "epoch": 11.32,
1839
+ "learning_rate": 4.906937394247039e-05,
1840
+ "loss": 0.0305,
1841
+ "step": 2230
1842
+ },
1843
+ {
1844
+ "epoch": 11.37,
1845
+ "learning_rate": 4.8392554991539766e-05,
1846
+ "loss": 0.0174,
1847
+ "step": 2240
1848
+ },
1849
+ {
1850
+ "epoch": 11.37,
1851
+ "eval_accuracy": 0.5026392961876833,
1852
+ "eval_loss": 2.443206548690796,
1853
+ "eval_runtime": 30.8601,
1854
+ "eval_samples_per_second": 55.249,
1855
+ "eval_steps_per_second": 6.935,
1856
+ "step": 2240
1857
+ },
1858
+ {
1859
+ "epoch": 11.42,
1860
+ "learning_rate": 4.771573604060914e-05,
1861
+ "loss": 0.0125,
1862
+ "step": 2250
1863
+ },
1864
+ {
1865
+ "epoch": 11.47,
1866
+ "learning_rate": 4.7038917089678517e-05,
1867
+ "loss": 0.0127,
1868
+ "step": 2260
1869
+ },
1870
+ {
1871
+ "epoch": 11.52,
1872
+ "learning_rate": 4.636209813874789e-05,
1873
+ "loss": 0.0253,
1874
+ "step": 2270
1875
+ },
1876
+ {
1877
+ "epoch": 11.57,
1878
+ "learning_rate": 4.568527918781726e-05,
1879
+ "loss": 0.0241,
1880
+ "step": 2280
1881
+ },
1882
+ {
1883
+ "epoch": 11.57,
1884
+ "eval_accuracy": 0.5055718475073314,
1885
+ "eval_loss": 2.4564218521118164,
1886
+ "eval_runtime": 30.6781,
1887
+ "eval_samples_per_second": 55.577,
1888
+ "eval_steps_per_second": 6.976,
1889
+ "step": 2280
1890
+ },
1891
+ {
1892
+ "epoch": 11.62,
1893
+ "learning_rate": 4.500846023688663e-05,
1894
+ "loss": 0.0084,
1895
+ "step": 2290
1896
+ },
1897
+ {
1898
+ "epoch": 11.68,
1899
+ "learning_rate": 4.433164128595601e-05,
1900
+ "loss": 0.0054,
1901
+ "step": 2300
1902
+ },
1903
+ {
1904
+ "epoch": 11.73,
1905
+ "learning_rate": 4.365482233502538e-05,
1906
+ "loss": 0.0055,
1907
+ "step": 2310
1908
+ },
1909
+ {
1910
+ "epoch": 11.78,
1911
+ "learning_rate": 4.2978003384094756e-05,
1912
+ "loss": 0.0182,
1913
+ "step": 2320
1914
+ },
1915
+ {
1916
+ "epoch": 11.78,
1917
+ "eval_accuracy": 0.5038123167155425,
1918
+ "eval_loss": 2.481368064880371,
1919
+ "eval_runtime": 30.9495,
1920
+ "eval_samples_per_second": 55.09,
1921
+ "eval_steps_per_second": 6.914,
1922
+ "step": 2320
1923
+ },
1924
+ {
1925
+ "epoch": 11.83,
1926
+ "learning_rate": 4.230118443316413e-05,
1927
+ "loss": 0.0205,
1928
+ "step": 2330
1929
+ },
1930
+ {
1931
+ "epoch": 11.88,
1932
+ "learning_rate": 4.162436548223351e-05,
1933
+ "loss": 0.0053,
1934
+ "step": 2340
1935
+ },
1936
+ {
1937
+ "epoch": 11.93,
1938
+ "learning_rate": 4.094754653130288e-05,
1939
+ "loss": 0.0153,
1940
+ "step": 2350
1941
+ },
1942
+ {
1943
+ "epoch": 11.98,
1944
+ "learning_rate": 4.027072758037225e-05,
1945
+ "loss": 0.0196,
1946
+ "step": 2360
1947
+ },
1948
+ {
1949
+ "epoch": 11.98,
1950
+ "eval_accuracy": 0.5032258064516129,
1951
+ "eval_loss": 2.4746322631835938,
1952
+ "eval_runtime": 30.3112,
1953
+ "eval_samples_per_second": 56.25,
1954
+ "eval_steps_per_second": 7.06,
1955
+ "step": 2360
1956
+ },
1957
+ {
1958
+ "epoch": 12.03,
1959
+ "learning_rate": 3.959390862944163e-05,
1960
+ "loss": 0.0114,
1961
+ "step": 2370
1962
+ },
1963
+ {
1964
+ "epoch": 12.08,
1965
+ "learning_rate": 3.8917089678510996e-05,
1966
+ "loss": 0.0144,
1967
+ "step": 2380
1968
+ },
1969
+ {
1970
+ "epoch": 12.13,
1971
+ "learning_rate": 3.8240270727580375e-05,
1972
+ "loss": 0.0193,
1973
+ "step": 2390
1974
+ },
1975
+ {
1976
+ "epoch": 12.18,
1977
+ "learning_rate": 3.756345177664975e-05,
1978
+ "loss": 0.0104,
1979
+ "step": 2400
1980
+ },
1981
+ {
1982
+ "epoch": 12.18,
1983
+ "eval_accuracy": 0.5079178885630499,
1984
+ "eval_loss": 2.4812278747558594,
1985
+ "eval_runtime": 30.7018,
1986
+ "eval_samples_per_second": 55.534,
1987
+ "eval_steps_per_second": 6.97,
1988
+ "step": 2400
1989
+ },
1990
+ {
1991
+ "epoch": 12.23,
1992
+ "learning_rate": 3.688663282571912e-05,
1993
+ "loss": 0.015,
1994
+ "step": 2410
1995
+ },
1996
+ {
1997
+ "epoch": 12.28,
1998
+ "learning_rate": 3.62098138747885e-05,
1999
+ "loss": 0.0115,
2000
+ "step": 2420
2001
+ },
2002
+ {
2003
+ "epoch": 12.34,
2004
+ "learning_rate": 3.553299492385787e-05,
2005
+ "loss": 0.0152,
2006
+ "step": 2430
2007
+ },
2008
+ {
2009
+ "epoch": 12.39,
2010
+ "learning_rate": 3.485617597292725e-05,
2011
+ "loss": 0.0254,
2012
+ "step": 2440
2013
+ },
2014
+ {
2015
+ "epoch": 12.39,
2016
+ "eval_accuracy": 0.506158357771261,
2017
+ "eval_loss": 2.465996265411377,
2018
+ "eval_runtime": 30.4976,
2019
+ "eval_samples_per_second": 55.906,
2020
+ "eval_steps_per_second": 7.017,
2021
+ "step": 2440
2022
+ },
2023
+ {
2024
+ "epoch": 12.44,
2025
+ "learning_rate": 3.4179357021996615e-05,
2026
+ "loss": 0.005,
2027
+ "step": 2450
2028
+ },
2029
+ {
2030
+ "epoch": 12.49,
2031
+ "learning_rate": 3.3502538071065994e-05,
2032
+ "loss": 0.0046,
2033
+ "step": 2460
2034
+ },
2035
+ {
2036
+ "epoch": 12.54,
2037
+ "learning_rate": 3.2825719120135366e-05,
2038
+ "loss": 0.0048,
2039
+ "step": 2470
2040
+ },
2041
+ {
2042
+ "epoch": 12.59,
2043
+ "learning_rate": 3.214890016920474e-05,
2044
+ "loss": 0.0057,
2045
+ "step": 2480
2046
+ },
2047
+ {
2048
+ "epoch": 12.59,
2049
+ "eval_accuracy": 0.5096774193548387,
2050
+ "eval_loss": 2.4570250511169434,
2051
+ "eval_runtime": 30.647,
2052
+ "eval_samples_per_second": 55.634,
2053
+ "eval_steps_per_second": 6.983,
2054
+ "step": 2480
2055
+ },
2056
+ {
2057
+ "epoch": 12.64,
2058
+ "learning_rate": 3.147208121827411e-05,
2059
+ "loss": 0.0159,
2060
+ "step": 2490
2061
+ },
2062
+ {
2063
+ "epoch": 12.69,
2064
+ "learning_rate": 3.079526226734349e-05,
2065
+ "loss": 0.0156,
2066
+ "step": 2500
2067
+ },
2068
+ {
2069
+ "epoch": 12.74,
2070
+ "learning_rate": 3.0118443316412858e-05,
2071
+ "loss": 0.0089,
2072
+ "step": 2510
2073
+ },
2074
+ {
2075
+ "epoch": 12.79,
2076
+ "learning_rate": 2.9441624365482233e-05,
2077
+ "loss": 0.0161,
2078
+ "step": 2520
2079
+ },
2080
+ {
2081
+ "epoch": 12.79,
2082
+ "eval_accuracy": 0.509090909090909,
2083
+ "eval_loss": 2.452195644378662,
2084
+ "eval_runtime": 30.5139,
2085
+ "eval_samples_per_second": 55.876,
2086
+ "eval_steps_per_second": 7.013,
2087
+ "step": 2520
2088
+ },
2089
+ {
2090
+ "epoch": 12.84,
2091
+ "learning_rate": 2.876480541455161e-05,
2092
+ "loss": 0.0046,
2093
+ "step": 2530
2094
+ },
2095
+ {
2096
+ "epoch": 12.89,
2097
+ "learning_rate": 2.808798646362098e-05,
2098
+ "loss": 0.0181,
2099
+ "step": 2540
2100
+ },
2101
+ {
2102
+ "epoch": 12.94,
2103
+ "learning_rate": 2.7411167512690357e-05,
2104
+ "loss": 0.0131,
2105
+ "step": 2550
2106
+ },
2107
+ {
2108
+ "epoch": 12.99,
2109
+ "learning_rate": 2.6734348561759732e-05,
2110
+ "loss": 0.0237,
2111
+ "step": 2560
2112
+ },
2113
+ {
2114
+ "epoch": 12.99,
2115
+ "eval_accuracy": 0.506158357771261,
2116
+ "eval_loss": 2.461273670196533,
2117
+ "eval_runtime": 30.9259,
2118
+ "eval_samples_per_second": 55.132,
2119
+ "eval_steps_per_second": 6.92,
2120
+ "step": 2560
2121
+ },
2122
+ {
2123
+ "epoch": 13.05,
2124
+ "learning_rate": 2.6057529610829108e-05,
2125
+ "loss": 0.0098,
2126
+ "step": 2570
2127
+ },
2128
+ {
2129
+ "epoch": 13.1,
2130
+ "learning_rate": 2.5380710659898476e-05,
2131
+ "loss": 0.0135,
2132
+ "step": 2580
2133
+ },
2134
+ {
2135
+ "epoch": 13.15,
2136
+ "learning_rate": 2.4703891708967852e-05,
2137
+ "loss": 0.009,
2138
+ "step": 2590
2139
+ },
2140
+ {
2141
+ "epoch": 13.2,
2142
+ "learning_rate": 2.4027072758037224e-05,
2143
+ "loss": 0.0048,
2144
+ "step": 2600
2145
+ },
2146
+ {
2147
+ "epoch": 13.2,
2148
+ "eval_accuracy": 0.5055718475073314,
2149
+ "eval_loss": 2.4772698879241943,
2150
+ "eval_runtime": 30.4386,
2151
+ "eval_samples_per_second": 56.014,
2152
+ "eval_steps_per_second": 7.031,
2153
+ "step": 2600
2154
+ },
2155
+ {
2156
+ "epoch": 13.25,
2157
+ "learning_rate": 2.33502538071066e-05,
2158
+ "loss": 0.0138,
2159
+ "step": 2610
2160
+ },
2161
+ {
2162
+ "epoch": 13.3,
2163
+ "learning_rate": 2.2673434856175975e-05,
2164
+ "loss": 0.009,
2165
+ "step": 2620
2166
+ },
2167
+ {
2168
+ "epoch": 13.35,
2169
+ "learning_rate": 2.199661590524535e-05,
2170
+ "loss": 0.0044,
2171
+ "step": 2630
2172
+ },
2173
+ {
2174
+ "epoch": 13.4,
2175
+ "learning_rate": 2.1319796954314723e-05,
2176
+ "loss": 0.0067,
2177
+ "step": 2640
2178
+ },
2179
+ {
2180
+ "epoch": 13.4,
2181
+ "eval_accuracy": 0.5079178885630499,
2182
+ "eval_loss": 2.474796772003174,
2183
+ "eval_runtime": 30.9994,
2184
+ "eval_samples_per_second": 55.001,
2185
+ "eval_steps_per_second": 6.903,
2186
+ "step": 2640
2187
+ },
2188
+ {
2189
+ "epoch": 13.45,
2190
+ "learning_rate": 2.0642978003384095e-05,
2191
+ "loss": 0.0048,
2192
+ "step": 2650
2193
+ },
2194
+ {
2195
+ "epoch": 13.5,
2196
+ "learning_rate": 1.996615905245347e-05,
2197
+ "loss": 0.0201,
2198
+ "step": 2660
2199
+ },
2200
+ {
2201
+ "epoch": 13.55,
2202
+ "learning_rate": 1.9289340101522843e-05,
2203
+ "loss": 0.0144,
2204
+ "step": 2670
2205
+ },
2206
+ {
2207
+ "epoch": 13.6,
2208
+ "learning_rate": 1.8612521150592218e-05,
2209
+ "loss": 0.0105,
2210
+ "step": 2680
2211
+ },
2212
+ {
2213
+ "epoch": 13.6,
2214
+ "eval_accuracy": 0.5085043988269795,
2215
+ "eval_loss": 2.4759089946746826,
2216
+ "eval_runtime": 30.6746,
2217
+ "eval_samples_per_second": 55.584,
2218
+ "eval_steps_per_second": 6.976,
2219
+ "step": 2680
2220
+ },
2221
+ {
2222
+ "epoch": 13.65,
2223
+ "learning_rate": 1.793570219966159e-05,
2224
+ "loss": 0.0174,
2225
+ "step": 2690
2226
+ },
2227
+ {
2228
+ "epoch": 13.71,
2229
+ "learning_rate": 1.7258883248730966e-05,
2230
+ "loss": 0.0045,
2231
+ "step": 2700
2232
+ },
2233
+ {
2234
+ "epoch": 13.76,
2235
+ "learning_rate": 1.658206429780034e-05,
2236
+ "loss": 0.0044,
2237
+ "step": 2710
2238
+ },
2239
+ {
2240
+ "epoch": 13.81,
2241
+ "learning_rate": 1.5905245346869714e-05,
2242
+ "loss": 0.0209,
2243
+ "step": 2720
2244
+ },
2245
+ {
2246
+ "epoch": 13.81,
2247
+ "eval_accuracy": 0.506158357771261,
2248
+ "eval_loss": 2.4831149578094482,
2249
+ "eval_runtime": 30.8346,
2250
+ "eval_samples_per_second": 55.295,
2251
+ "eval_steps_per_second": 6.94,
2252
+ "step": 2720
2253
+ },
2254
+ {
2255
+ "epoch": 13.86,
2256
+ "learning_rate": 1.5228426395939088e-05,
2257
+ "loss": 0.0122,
2258
+ "step": 2730
2259
+ },
2260
+ {
2261
+ "epoch": 13.91,
2262
+ "learning_rate": 1.4551607445008461e-05,
2263
+ "loss": 0.0149,
2264
+ "step": 2740
2265
+ },
2266
+ {
2267
+ "epoch": 13.96,
2268
+ "learning_rate": 1.3874788494077834e-05,
2269
+ "loss": 0.0225,
2270
+ "step": 2750
2271
+ },
2272
+ {
2273
+ "epoch": 14.01,
2274
+ "learning_rate": 1.3197969543147209e-05,
2275
+ "loss": 0.0112,
2276
+ "step": 2760
2277
+ },
2278
+ {
2279
+ "epoch": 14.01,
2280
+ "eval_accuracy": 0.5085043988269795,
2281
+ "eval_loss": 2.482898473739624,
2282
+ "eval_runtime": 30.9465,
2283
+ "eval_samples_per_second": 55.095,
2284
+ "eval_steps_per_second": 6.915,
2285
+ "step": 2760
2286
+ },
2287
+ {
2288
+ "epoch": 14.06,
2289
+ "learning_rate": 1.2521150592216583e-05,
2290
+ "loss": 0.0084,
2291
+ "step": 2770
2292
+ },
2293
+ {
2294
+ "epoch": 14.11,
2295
+ "learning_rate": 1.1844331641285957e-05,
2296
+ "loss": 0.0127,
2297
+ "step": 2780
2298
+ },
2299
+ {
2300
+ "epoch": 14.16,
2301
+ "learning_rate": 1.116751269035533e-05,
2302
+ "loss": 0.0097,
2303
+ "step": 2790
2304
+ },
2305
+ {
2306
+ "epoch": 14.21,
2307
+ "learning_rate": 1.0490693739424704e-05,
2308
+ "loss": 0.0084,
2309
+ "step": 2800
2310
+ },
2311
+ {
2312
+ "epoch": 14.21,
2313
+ "eval_accuracy": 0.509090909090909,
2314
+ "eval_loss": 2.4852123260498047,
2315
+ "eval_runtime": 30.864,
2316
+ "eval_samples_per_second": 55.242,
2317
+ "eval_steps_per_second": 6.934,
2318
+ "step": 2800
2319
+ },
2320
+ {
2321
+ "epoch": 14.26,
2322
+ "learning_rate": 9.813874788494078e-06,
2323
+ "loss": 0.0078,
2324
+ "step": 2810
2325
+ },
2326
+ {
2327
+ "epoch": 14.31,
2328
+ "learning_rate": 9.137055837563452e-06,
2329
+ "loss": 0.0173,
2330
+ "step": 2820
2331
+ },
2332
+ {
2333
+ "epoch": 14.37,
2334
+ "learning_rate": 8.460236886632826e-06,
2335
+ "loss": 0.0145,
2336
+ "step": 2830
2337
+ },
2338
+ {
2339
+ "epoch": 14.42,
2340
+ "learning_rate": 7.7834179357022e-06,
2341
+ "loss": 0.0203,
2342
+ "step": 2840
2343
+ },
2344
+ {
2345
+ "epoch": 14.42,
2346
+ "eval_accuracy": 0.5096774193548387,
2347
+ "eval_loss": 2.4804649353027344,
2348
+ "eval_runtime": 30.8958,
2349
+ "eval_samples_per_second": 55.185,
2350
+ "eval_steps_per_second": 6.927,
2351
+ "step": 2840
2352
+ },
2353
+ {
2354
+ "epoch": 14.47,
2355
+ "learning_rate": 7.106598984771575e-06,
2356
+ "loss": 0.0143,
2357
+ "step": 2850
2358
+ },
2359
+ {
2360
+ "epoch": 14.52,
2361
+ "learning_rate": 6.429780033840948e-06,
2362
+ "loss": 0.0094,
2363
+ "step": 2860
2364
+ },
2365
+ {
2366
+ "epoch": 14.57,
2367
+ "learning_rate": 5.7529610829103214e-06,
2368
+ "loss": 0.0105,
2369
+ "step": 2870
2370
+ },
2371
+ {
2372
+ "epoch": 14.62,
2373
+ "learning_rate": 5.076142131979695e-06,
2374
+ "loss": 0.0086,
2375
+ "step": 2880
2376
+ },
2377
+ {
2378
+ "epoch": 14.62,
2379
+ "eval_accuracy": 0.5085043988269795,
2380
+ "eval_loss": 2.480623960494995,
2381
+ "eval_runtime": 30.8962,
2382
+ "eval_samples_per_second": 55.185,
2383
+ "eval_steps_per_second": 6.926,
2384
+ "step": 2880
2385
+ },
2386
+ {
2387
+ "epoch": 14.67,
2388
+ "learning_rate": 4.39932318104907e-06,
2389
+ "loss": 0.01,
2390
+ "step": 2890
2391
+ },
2392
+ {
2393
+ "epoch": 14.72,
2394
+ "learning_rate": 3.7225042301184434e-06,
2395
+ "loss": 0.0048,
2396
+ "step": 2900
2397
+ },
2398
+ {
2399
+ "epoch": 14.77,
2400
+ "learning_rate": 3.0456852791878177e-06,
2401
+ "loss": 0.0087,
2402
+ "step": 2910
2403
+ },
2404
+ {
2405
+ "epoch": 14.82,
2406
+ "learning_rate": 2.3688663282571915e-06,
2407
+ "loss": 0.0103,
2408
+ "step": 2920
2409
+ },
2410
+ {
2411
+ "epoch": 14.82,
2412
+ "eval_accuracy": 0.5085043988269795,
2413
+ "eval_loss": 2.4825499057769775,
2414
+ "eval_runtime": 30.9894,
2415
+ "eval_samples_per_second": 55.019,
2416
+ "eval_steps_per_second": 6.906,
2417
+ "step": 2920
2418
+ }
2419
+ ],
2420
+ "max_steps": 2955,
2421
+ "num_train_epochs": 15,
2422
+ "total_flos": 3.6066913919101256e+18,
2423
+ "trial_name": null,
2424
+ "trial_params": null
2425
+ }
checkpoint-2920/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:81bc596a5fb737e64cac2afcc4b39d05df7c2c9ae8aa4e8c80cb21301b1d396e
3
+ size 3963
checkpoint-640/config.json ADDED
@@ -0,0 +1,122 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/vit-base-patch16-224-in21k",
3
+ "architectures": [
4
+ "ViTForImageClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "encoder_stride": 16,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.0,
10
+ "hidden_size": 768,
11
+ "id2label": {
12
+ "0": "Abstract art",
13
+ "1": "African Art",
14
+ "10": "Constructivism art",
15
+ "11": "Constructivsm art",
16
+ "12": "Contemporary art",
17
+ "13": "Cubism art",
18
+ "14": "Ethiopian Art",
19
+ "15": "Expressionism paintings",
20
+ "16": "Fauvism paintings",
21
+ "17": "Gothic art",
22
+ "18": "Graffiti",
23
+ "19": "Impressionism and Post-Impressionism",
24
+ "2": "Amazonian Art",
25
+ "20": "Indochristian art",
26
+ "21": "Islamic Art",
27
+ "22": "Japanese Edo Period",
28
+ "23": "Japanese Heian Period",
29
+ "24": "Japanese momoyama",
30
+ "25": "Kamakura or Muromachi Art",
31
+ "26": "Korean Goryeo Dynasty Art",
32
+ "27": "Late Baroque (Rococo) art",
33
+ "28": "Latin American Modernism Art",
34
+ "29": "Latin American muralism",
35
+ "3": "Ancient Egyptian and Nubian Art",
36
+ "30": "Mannerism paintings",
37
+ "31": "Medieval art",
38
+ "32": "Minimalism paintings",
39
+ "33": "Modernism artwork",
40
+ "34": "Naive art",
41
+ "35": "Neo-expressionism art",
42
+ "36": "Neoclassical art",
43
+ "37": "Ottoman Art",
44
+ "38": "Polynesian Art",
45
+ "39": "Pop art",
46
+ "4": "Ancient and Traditional Chinese Art",
47
+ "40": "Realism paintings",
48
+ "41": "Renaissance paintings",
49
+ "42": "Romanticism paintings",
50
+ "43": "Surrealist paintings",
51
+ "44": "cubism",
52
+ "45": "expressionism",
53
+ "46": "romanticism",
54
+ "5": "Anime art",
55
+ "6": "Art Nouveau paintings",
56
+ "7": "Art deco paintings",
57
+ "8": "Baroque paintings",
58
+ "9": "Cave Art"
59
+ },
60
+ "image_size": 224,
61
+ "initializer_range": 0.02,
62
+ "intermediate_size": 3072,
63
+ "label2id": {
64
+ "Abstract art": "0",
65
+ "African Art": "1",
66
+ "Amazonian Art": "2",
67
+ "Ancient Egyptian and Nubian Art": "3",
68
+ "Ancient and Traditional Chinese Art": "4",
69
+ "Anime art": "5",
70
+ "Art Nouveau paintings": "6",
71
+ "Art deco paintings": "7",
72
+ "Baroque paintings": "8",
73
+ "Cave Art": "9",
74
+ "Constructivism art": "10",
75
+ "Constructivsm art": "11",
76
+ "Contemporary art": "12",
77
+ "Cubism art": "13",
78
+ "Ethiopian Art": "14",
79
+ "Expressionism paintings": "15",
80
+ "Fauvism paintings": "16",
81
+ "Gothic art": "17",
82
+ "Graffiti": "18",
83
+ "Impressionism and Post-Impressionism": "19",
84
+ "Indochristian art": "20",
85
+ "Islamic Art": "21",
86
+ "Japanese Edo Period": "22",
87
+ "Japanese Heian Period": "23",
88
+ "Japanese momoyama": "24",
89
+ "Kamakura or Muromachi Art": "25",
90
+ "Korean Goryeo Dynasty Art": "26",
91
+ "Late Baroque (Rococo) art": "27",
92
+ "Latin American Modernism Art": "28",
93
+ "Latin American muralism": "29",
94
+ "Mannerism paintings": "30",
95
+ "Medieval art": "31",
96
+ "Minimalism paintings": "32",
97
+ "Modernism artwork": "33",
98
+ "Naive art": "34",
99
+ "Neo-expressionism art": "35",
100
+ "Neoclassical art": "36",
101
+ "Ottoman Art": "37",
102
+ "Polynesian Art": "38",
103
+ "Pop art": "39",
104
+ "Realism paintings": "40",
105
+ "Renaissance paintings": "41",
106
+ "Romanticism paintings": "42",
107
+ "Surrealist paintings": "43",
108
+ "cubism": "44",
109
+ "expressionism": "45",
110
+ "romanticism": "46"
111
+ },
112
+ "layer_norm_eps": 1e-12,
113
+ "model_type": "vit",
114
+ "num_attention_heads": 12,
115
+ "num_channels": 3,
116
+ "num_hidden_layers": 12,
117
+ "patch_size": 16,
118
+ "problem_type": "single_label_classification",
119
+ "qkv_bias": true,
120
+ "torch_dtype": "float32",
121
+ "transformers_version": "4.31.0"
122
+ }
checkpoint-640/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b6cbc60d8597d8c8804f1875e4a460ec2f26a8e14f9e4b11928688a61c3b9d7
3
+ size 686795653
checkpoint-640/preprocessor_config.json ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "do_rescale": true,
4
+ "do_resize": true,
5
+ "image_mean": [
6
+ 0.5,
7
+ 0.5,
8
+ 0.5
9
+ ],
10
+ "image_processor_type": "ViTFeatureExtractor",
11
+ "image_std": [
12
+ 0.5,
13
+ 0.5,
14
+ 0.5
15
+ ],
16
+ "resample": 2,
17
+ "rescale_factor": 0.00392156862745098,
18
+ "size": {
19
+ "height": 224,
20
+ "width": 224
21
+ }
22
+ }
checkpoint-640/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a175054a69a0254c1fd94eabe1859be451fc0af4baeadc7f31803dc43992b88
3
+ size 343407149
checkpoint-640/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:370c3a07f37a8aae6ea141b54ca992b21699546baf7407eb587b6056f787333b
3
+ size 14575
checkpoint-640/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f99394df6cb7f1dde7cfef172df7859b229a2dd29be57071b7802e5f11def01
3
+ size 627
checkpoint-640/trainer_state.json ADDED
@@ -0,0 +1,544 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 1.8420542478561401,
3
+ "best_model_checkpoint": "./vit-base-beans/checkpoint-640",
4
+ "epoch": 3.248730964467005,
5
+ "global_step": 640,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.05,
12
+ "learning_rate": 0.00019932318104906937,
13
+ "loss": 3.8192,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.1,
18
+ "learning_rate": 0.00019864636209813877,
19
+ "loss": 3.7847,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.15,
24
+ "learning_rate": 0.00019796954314720813,
25
+ "loss": 3.6849,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.2,
30
+ "learning_rate": 0.00019729272419627753,
31
+ "loss": 3.6122,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.2,
36
+ "eval_accuracy": 0.10967741935483871,
37
+ "eval_loss": 3.561206340789795,
38
+ "eval_runtime": 31.2002,
39
+ "eval_samples_per_second": 54.647,
40
+ "eval_steps_per_second": 6.859,
41
+ "step": 40
42
+ },
43
+ {
44
+ "epoch": 0.25,
45
+ "learning_rate": 0.00019661590524534686,
46
+ "loss": 3.5748,
47
+ "step": 50
48
+ },
49
+ {
50
+ "epoch": 0.3,
51
+ "learning_rate": 0.00019593908629441626,
52
+ "loss": 3.491,
53
+ "step": 60
54
+ },
55
+ {
56
+ "epoch": 0.36,
57
+ "learning_rate": 0.00019526226734348562,
58
+ "loss": 3.3556,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.41,
63
+ "learning_rate": 0.000194585448392555,
64
+ "loss": 3.4008,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.41,
69
+ "eval_accuracy": 0.2475073313782991,
70
+ "eval_loss": 3.2414700984954834,
71
+ "eval_runtime": 30.544,
72
+ "eval_samples_per_second": 55.821,
73
+ "eval_steps_per_second": 7.006,
74
+ "step": 80
75
+ },
76
+ {
77
+ "epoch": 0.46,
78
+ "learning_rate": 0.00019390862944162438,
79
+ "loss": 3.2255,
80
+ "step": 90
81
+ },
82
+ {
83
+ "epoch": 0.51,
84
+ "learning_rate": 0.00019323181049069374,
85
+ "loss": 3.1171,
86
+ "step": 100
87
+ },
88
+ {
89
+ "epoch": 0.56,
90
+ "learning_rate": 0.00019255499153976314,
91
+ "loss": 3.1394,
92
+ "step": 110
93
+ },
94
+ {
95
+ "epoch": 0.61,
96
+ "learning_rate": 0.0001918781725888325,
97
+ "loss": 3.0872,
98
+ "step": 120
99
+ },
100
+ {
101
+ "epoch": 0.61,
102
+ "eval_accuracy": 0.2809384164222874,
103
+ "eval_loss": 2.9989874362945557,
104
+ "eval_runtime": 30.8814,
105
+ "eval_samples_per_second": 55.211,
106
+ "eval_steps_per_second": 6.93,
107
+ "step": 120
108
+ },
109
+ {
110
+ "epoch": 0.66,
111
+ "learning_rate": 0.00019120135363790187,
112
+ "loss": 2.9773,
113
+ "step": 130
114
+ },
115
+ {
116
+ "epoch": 0.71,
117
+ "learning_rate": 0.00019052453468697123,
118
+ "loss": 2.7993,
119
+ "step": 140
120
+ },
121
+ {
122
+ "epoch": 0.76,
123
+ "learning_rate": 0.0001898477157360406,
124
+ "loss": 2.8672,
125
+ "step": 150
126
+ },
127
+ {
128
+ "epoch": 0.81,
129
+ "learning_rate": 0.00018917089678511,
130
+ "loss": 2.8733,
131
+ "step": 160
132
+ },
133
+ {
134
+ "epoch": 0.81,
135
+ "eval_accuracy": 0.34252199413489737,
136
+ "eval_loss": 2.7833735942840576,
137
+ "eval_runtime": 30.3566,
138
+ "eval_samples_per_second": 56.166,
139
+ "eval_steps_per_second": 7.05,
140
+ "step": 160
141
+ },
142
+ {
143
+ "epoch": 0.86,
144
+ "learning_rate": 0.00018849407783417936,
145
+ "loss": 2.7912,
146
+ "step": 170
147
+ },
148
+ {
149
+ "epoch": 0.91,
150
+ "learning_rate": 0.00018781725888324875,
151
+ "loss": 2.7577,
152
+ "step": 180
153
+ },
154
+ {
155
+ "epoch": 0.96,
156
+ "learning_rate": 0.00018714043993231811,
157
+ "loss": 2.7865,
158
+ "step": 190
159
+ },
160
+ {
161
+ "epoch": 1.02,
162
+ "learning_rate": 0.00018646362098138748,
163
+ "loss": 2.591,
164
+ "step": 200
165
+ },
166
+ {
167
+ "epoch": 1.02,
168
+ "eval_accuracy": 0.37360703812316715,
169
+ "eval_loss": 2.6250743865966797,
170
+ "eval_runtime": 30.4336,
171
+ "eval_samples_per_second": 56.024,
172
+ "eval_steps_per_second": 7.032,
173
+ "step": 200
174
+ },
175
+ {
176
+ "epoch": 1.07,
177
+ "learning_rate": 0.00018578680203045687,
178
+ "loss": 2.3103,
179
+ "step": 210
180
+ },
181
+ {
182
+ "epoch": 1.12,
183
+ "learning_rate": 0.00018510998307952624,
184
+ "loss": 2.2942,
185
+ "step": 220
186
+ },
187
+ {
188
+ "epoch": 1.17,
189
+ "learning_rate": 0.0001844331641285956,
190
+ "loss": 2.3883,
191
+ "step": 230
192
+ },
193
+ {
194
+ "epoch": 1.22,
195
+ "learning_rate": 0.00018375634517766497,
196
+ "loss": 2.1643,
197
+ "step": 240
198
+ },
199
+ {
200
+ "epoch": 1.22,
201
+ "eval_accuracy": 0.40527859237536656,
202
+ "eval_loss": 2.423466920852661,
203
+ "eval_runtime": 30.3039,
204
+ "eval_samples_per_second": 56.263,
205
+ "eval_steps_per_second": 7.062,
206
+ "step": 240
207
+ },
208
+ {
209
+ "epoch": 1.27,
210
+ "learning_rate": 0.00018307952622673436,
211
+ "loss": 2.2077,
212
+ "step": 250
213
+ },
214
+ {
215
+ "epoch": 1.32,
216
+ "learning_rate": 0.00018240270727580373,
217
+ "loss": 2.1227,
218
+ "step": 260
219
+ },
220
+ {
221
+ "epoch": 1.37,
222
+ "learning_rate": 0.0001817258883248731,
223
+ "loss": 2.0657,
224
+ "step": 270
225
+ },
226
+ {
227
+ "epoch": 1.42,
228
+ "learning_rate": 0.00018104906937394248,
229
+ "loss": 2.0669,
230
+ "step": 280
231
+ },
232
+ {
233
+ "epoch": 1.42,
234
+ "eval_accuracy": 0.3882697947214076,
235
+ "eval_loss": 2.3226494789123535,
236
+ "eval_runtime": 30.2544,
237
+ "eval_samples_per_second": 56.355,
238
+ "eval_steps_per_second": 7.073,
239
+ "step": 280
240
+ },
241
+ {
242
+ "epoch": 1.47,
243
+ "learning_rate": 0.00018037225042301185,
244
+ "loss": 2.1189,
245
+ "step": 290
246
+ },
247
+ {
248
+ "epoch": 1.52,
249
+ "learning_rate": 0.00017969543147208124,
250
+ "loss": 2.1857,
251
+ "step": 300
252
+ },
253
+ {
254
+ "epoch": 1.57,
255
+ "learning_rate": 0.0001790186125211506,
256
+ "loss": 2.272,
257
+ "step": 310
258
+ },
259
+ {
260
+ "epoch": 1.62,
261
+ "learning_rate": 0.00017834179357021997,
262
+ "loss": 1.9717,
263
+ "step": 320
264
+ },
265
+ {
266
+ "epoch": 1.62,
267
+ "eval_accuracy": 0.40821114369501466,
268
+ "eval_loss": 2.2509706020355225,
269
+ "eval_runtime": 30.2595,
270
+ "eval_samples_per_second": 56.346,
271
+ "eval_steps_per_second": 7.072,
272
+ "step": 320
273
+ },
274
+ {
275
+ "epoch": 1.68,
276
+ "learning_rate": 0.00017766497461928934,
277
+ "loss": 2.0974,
278
+ "step": 330
279
+ },
280
+ {
281
+ "epoch": 1.73,
282
+ "learning_rate": 0.0001769881556683587,
283
+ "loss": 2.1316,
284
+ "step": 340
285
+ },
286
+ {
287
+ "epoch": 1.78,
288
+ "learning_rate": 0.0001763113367174281,
289
+ "loss": 2.0845,
290
+ "step": 350
291
+ },
292
+ {
293
+ "epoch": 1.83,
294
+ "learning_rate": 0.00017563451776649746,
295
+ "loss": 2.1714,
296
+ "step": 360
297
+ },
298
+ {
299
+ "epoch": 1.83,
300
+ "eval_accuracy": 0.4351906158357771,
301
+ "eval_loss": 2.203674077987671,
302
+ "eval_runtime": 30.4099,
303
+ "eval_samples_per_second": 56.067,
304
+ "eval_steps_per_second": 7.037,
305
+ "step": 360
306
+ },
307
+ {
308
+ "epoch": 1.88,
309
+ "learning_rate": 0.00017495769881556685,
310
+ "loss": 2.0881,
311
+ "step": 370
312
+ },
313
+ {
314
+ "epoch": 1.93,
315
+ "learning_rate": 0.00017428087986463622,
316
+ "loss": 1.9428,
317
+ "step": 380
318
+ },
319
+ {
320
+ "epoch": 1.98,
321
+ "learning_rate": 0.0001736040609137056,
322
+ "loss": 1.8203,
323
+ "step": 390
324
+ },
325
+ {
326
+ "epoch": 2.03,
327
+ "learning_rate": 0.00017292724196277498,
328
+ "loss": 1.6183,
329
+ "step": 400
330
+ },
331
+ {
332
+ "epoch": 2.03,
333
+ "eval_accuracy": 0.4533724340175953,
334
+ "eval_loss": 2.0731186866760254,
335
+ "eval_runtime": 30.3675,
336
+ "eval_samples_per_second": 56.146,
337
+ "eval_steps_per_second": 7.047,
338
+ "step": 400
339
+ },
340
+ {
341
+ "epoch": 2.08,
342
+ "learning_rate": 0.00017225042301184434,
343
+ "loss": 1.3395,
344
+ "step": 410
345
+ },
346
+ {
347
+ "epoch": 2.13,
348
+ "learning_rate": 0.0001715736040609137,
349
+ "loss": 1.4052,
350
+ "step": 420
351
+ },
352
+ {
353
+ "epoch": 2.18,
354
+ "learning_rate": 0.00017089678510998307,
355
+ "loss": 1.3804,
356
+ "step": 430
357
+ },
358
+ {
359
+ "epoch": 2.23,
360
+ "learning_rate": 0.00017021996615905247,
361
+ "loss": 1.2771,
362
+ "step": 440
363
+ },
364
+ {
365
+ "epoch": 2.23,
366
+ "eval_accuracy": 0.42052785923753666,
367
+ "eval_loss": 2.0928714275360107,
368
+ "eval_runtime": 30.2484,
369
+ "eval_samples_per_second": 56.367,
370
+ "eval_steps_per_second": 7.075,
371
+ "step": 440
372
+ },
373
+ {
374
+ "epoch": 2.28,
375
+ "learning_rate": 0.00016954314720812183,
376
+ "loss": 1.431,
377
+ "step": 450
378
+ },
379
+ {
380
+ "epoch": 2.34,
381
+ "learning_rate": 0.00016886632825719122,
382
+ "loss": 1.353,
383
+ "step": 460
384
+ },
385
+ {
386
+ "epoch": 2.39,
387
+ "learning_rate": 0.0001681895093062606,
388
+ "loss": 1.307,
389
+ "step": 470
390
+ },
391
+ {
392
+ "epoch": 2.44,
393
+ "learning_rate": 0.00016751269035532995,
394
+ "loss": 1.2116,
395
+ "step": 480
396
+ },
397
+ {
398
+ "epoch": 2.44,
399
+ "eval_accuracy": 0.4504398826979472,
400
+ "eval_loss": 2.0171947479248047,
401
+ "eval_runtime": 30.2989,
402
+ "eval_samples_per_second": 56.273,
403
+ "eval_steps_per_second": 7.063,
404
+ "step": 480
405
+ },
406
+ {
407
+ "epoch": 2.49,
408
+ "learning_rate": 0.00016683587140439935,
409
+ "loss": 1.2762,
410
+ "step": 490
411
+ },
412
+ {
413
+ "epoch": 2.54,
414
+ "learning_rate": 0.0001661590524534687,
415
+ "loss": 1.2253,
416
+ "step": 500
417
+ },
418
+ {
419
+ "epoch": 2.59,
420
+ "learning_rate": 0.00016548223350253808,
421
+ "loss": 1.238,
422
+ "step": 510
423
+ },
424
+ {
425
+ "epoch": 2.64,
426
+ "learning_rate": 0.00016480541455160744,
427
+ "loss": 1.158,
428
+ "step": 520
429
+ },
430
+ {
431
+ "epoch": 2.64,
432
+ "eval_accuracy": 0.45689149560117304,
433
+ "eval_loss": 1.9670987129211426,
434
+ "eval_runtime": 30.2767,
435
+ "eval_samples_per_second": 56.314,
436
+ "eval_steps_per_second": 7.068,
437
+ "step": 520
438
+ },
439
+ {
440
+ "epoch": 2.69,
441
+ "learning_rate": 0.00016412859560067684,
442
+ "loss": 1.1488,
443
+ "step": 530
444
+ },
445
+ {
446
+ "epoch": 2.74,
447
+ "learning_rate": 0.0001634517766497462,
448
+ "loss": 1.294,
449
+ "step": 540
450
+ },
451
+ {
452
+ "epoch": 2.79,
453
+ "learning_rate": 0.00016277495769881557,
454
+ "loss": 1.0431,
455
+ "step": 550
456
+ },
457
+ {
458
+ "epoch": 2.84,
459
+ "learning_rate": 0.00016209813874788496,
460
+ "loss": 1.1864,
461
+ "step": 560
462
+ },
463
+ {
464
+ "epoch": 2.84,
465
+ "eval_accuracy": 0.4563049853372434,
466
+ "eval_loss": 1.9570399522781372,
467
+ "eval_runtime": 30.2455,
468
+ "eval_samples_per_second": 56.372,
469
+ "eval_steps_per_second": 7.075,
470
+ "step": 560
471
+ },
472
+ {
473
+ "epoch": 2.89,
474
+ "learning_rate": 0.00016142131979695432,
475
+ "loss": 1.1271,
476
+ "step": 570
477
+ },
478
+ {
479
+ "epoch": 2.94,
480
+ "learning_rate": 0.00016074450084602372,
481
+ "loss": 1.1006,
482
+ "step": 580
483
+ },
484
+ {
485
+ "epoch": 2.99,
486
+ "learning_rate": 0.00016006768189509308,
487
+ "loss": 1.1867,
488
+ "step": 590
489
+ },
490
+ {
491
+ "epoch": 3.05,
492
+ "learning_rate": 0.00015939086294416242,
493
+ "loss": 0.7778,
494
+ "step": 600
495
+ },
496
+ {
497
+ "epoch": 3.05,
498
+ "eval_accuracy": 0.49560117302052786,
499
+ "eval_loss": 1.8706681728363037,
500
+ "eval_runtime": 30.3606,
501
+ "eval_samples_per_second": 56.158,
502
+ "eval_steps_per_second": 7.049,
503
+ "step": 600
504
+ },
505
+ {
506
+ "epoch": 3.1,
507
+ "learning_rate": 0.0001587140439932318,
508
+ "loss": 0.6935,
509
+ "step": 610
510
+ },
511
+ {
512
+ "epoch": 3.15,
513
+ "learning_rate": 0.00015803722504230118,
514
+ "loss": 0.5692,
515
+ "step": 620
516
+ },
517
+ {
518
+ "epoch": 3.2,
519
+ "learning_rate": 0.00015736040609137057,
520
+ "loss": 0.5985,
521
+ "step": 630
522
+ },
523
+ {
524
+ "epoch": 3.25,
525
+ "learning_rate": 0.00015668358714043994,
526
+ "loss": 0.6021,
527
+ "step": 640
528
+ },
529
+ {
530
+ "epoch": 3.25,
531
+ "eval_accuracy": 0.4950146627565982,
532
+ "eval_loss": 1.8420542478561401,
533
+ "eval_runtime": 30.3002,
534
+ "eval_samples_per_second": 56.27,
535
+ "eval_steps_per_second": 7.063,
536
+ "step": 640
537
+ }
538
+ ],
539
+ "max_steps": 2955,
540
+ "num_train_epochs": 15,
541
+ "total_flos": 7.9058203969348e+17,
542
+ "trial_name": null,
543
+ "trial_params": null
544
+ }
checkpoint-640/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:81bc596a5fb737e64cac2afcc4b39d05df7c2c9ae8aa4e8c80cb21301b1d396e
3
+ size 3963
config.json ADDED
@@ -0,0 +1,122 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/vit-base-patch16-224-in21k",
3
+ "architectures": [
4
+ "ViTForImageClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "encoder_stride": 16,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.0,
10
+ "hidden_size": 768,
11
+ "id2label": {
12
+ "0": "Abstract art",
13
+ "1": "African Art",
14
+ "10": "Constructivism art",
15
+ "11": "Constructivsm art",
16
+ "12": "Contemporary art",
17
+ "13": "Cubism art",
18
+ "14": "Ethiopian Art",
19
+ "15": "Expressionism paintings",
20
+ "16": "Fauvism paintings",
21
+ "17": "Gothic art",
22
+ "18": "Graffiti",
23
+ "19": "Impressionism and Post-Impressionism",
24
+ "2": "Amazonian Art",
25
+ "20": "Indochristian art",
26
+ "21": "Islamic Art",
27
+ "22": "Japanese Edo Period",
28
+ "23": "Japanese Heian Period",
29
+ "24": "Japanese momoyama",
30
+ "25": "Kamakura or Muromachi Art",
31
+ "26": "Korean Goryeo Dynasty Art",
32
+ "27": "Late Baroque (Rococo) art",
33
+ "28": "Latin American Modernism Art",
34
+ "29": "Latin American muralism",
35
+ "3": "Ancient Egyptian and Nubian Art",
36
+ "30": "Mannerism paintings",
37
+ "31": "Medieval art",
38
+ "32": "Minimalism paintings",
39
+ "33": "Modernism artwork",
40
+ "34": "Naive art",
41
+ "35": "Neo-expressionism art",
42
+ "36": "Neoclassical art",
43
+ "37": "Ottoman Art",
44
+ "38": "Polynesian Art",
45
+ "39": "Pop art",
46
+ "4": "Ancient and Traditional Chinese Art",
47
+ "40": "Realism paintings",
48
+ "41": "Renaissance paintings",
49
+ "42": "Romanticism paintings",
50
+ "43": "Surrealist paintings",
51
+ "44": "cubism",
52
+ "45": "expressionism",
53
+ "46": "romanticism",
54
+ "5": "Anime art",
55
+ "6": "Art Nouveau paintings",
56
+ "7": "Art deco paintings",
57
+ "8": "Baroque paintings",
58
+ "9": "Cave Art"
59
+ },
60
+ "image_size": 224,
61
+ "initializer_range": 0.02,
62
+ "intermediate_size": 3072,
63
+ "label2id": {
64
+ "Abstract art": "0",
65
+ "African Art": "1",
66
+ "Amazonian Art": "2",
67
+ "Ancient Egyptian and Nubian Art": "3",
68
+ "Ancient and Traditional Chinese Art": "4",
69
+ "Anime art": "5",
70
+ "Art Nouveau paintings": "6",
71
+ "Art deco paintings": "7",
72
+ "Baroque paintings": "8",
73
+ "Cave Art": "9",
74
+ "Constructivism art": "10",
75
+ "Constructivsm art": "11",
76
+ "Contemporary art": "12",
77
+ "Cubism art": "13",
78
+ "Ethiopian Art": "14",
79
+ "Expressionism paintings": "15",
80
+ "Fauvism paintings": "16",
81
+ "Gothic art": "17",
82
+ "Graffiti": "18",
83
+ "Impressionism and Post-Impressionism": "19",
84
+ "Indochristian art": "20",
85
+ "Islamic Art": "21",
86
+ "Japanese Edo Period": "22",
87
+ "Japanese Heian Period": "23",
88
+ "Japanese momoyama": "24",
89
+ "Kamakura or Muromachi Art": "25",
90
+ "Korean Goryeo Dynasty Art": "26",
91
+ "Late Baroque (Rococo) art": "27",
92
+ "Latin American Modernism Art": "28",
93
+ "Latin American muralism": "29",
94
+ "Mannerism paintings": "30",
95
+ "Medieval art": "31",
96
+ "Minimalism paintings": "32",
97
+ "Modernism artwork": "33",
98
+ "Naive art": "34",
99
+ "Neo-expressionism art": "35",
100
+ "Neoclassical art": "36",
101
+ "Ottoman Art": "37",
102
+ "Polynesian Art": "38",
103
+ "Pop art": "39",
104
+ "Realism paintings": "40",
105
+ "Renaissance paintings": "41",
106
+ "Romanticism paintings": "42",
107
+ "Surrealist paintings": "43",
108
+ "cubism": "44",
109
+ "expressionism": "45",
110
+ "romanticism": "46"
111
+ },
112
+ "layer_norm_eps": 1e-12,
113
+ "model_type": "vit",
114
+ "num_attention_heads": 12,
115
+ "num_channels": 3,
116
+ "num_hidden_layers": 12,
117
+ "patch_size": 16,
118
+ "problem_type": "single_label_classification",
119
+ "qkv_bias": true,
120
+ "torch_dtype": "float32",
121
+ "transformers_version": "4.31.0"
122
+ }
preprocessor_config.json ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "do_rescale": true,
4
+ "do_resize": true,
5
+ "image_mean": [
6
+ 0.5,
7
+ 0.5,
8
+ 0.5
9
+ ],
10
+ "image_processor_type": "ViTFeatureExtractor",
11
+ "image_std": [
12
+ 0.5,
13
+ 0.5,
14
+ 0.5
15
+ ],
16
+ "resample": 2,
17
+ "rescale_factor": 0.00392156862745098,
18
+ "size": {
19
+ "height": 224,
20
+ "width": 224
21
+ }
22
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a175054a69a0254c1fd94eabe1859be451fc0af4baeadc7f31803dc43992b88
3
+ size 343407149
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 15.0,
3
+ "total_flos": 3.64901908299393e+18,
4
+ "train_loss": 0.5233747142333972,
5
+ "train_runtime": 4896.0515,
6
+ "train_samples_per_second": 9.614,
7
+ "train_steps_per_second": 0.604
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,2452 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 1.8420542478561401,
3
+ "best_model_checkpoint": "./vit-base-beans/checkpoint-640",
4
+ "epoch": 15.0,
5
+ "global_step": 2955,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.05,
12
+ "learning_rate": 0.00019932318104906937,
13
+ "loss": 3.8192,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.1,
18
+ "learning_rate": 0.00019864636209813877,
19
+ "loss": 3.7847,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.15,
24
+ "learning_rate": 0.00019796954314720813,
25
+ "loss": 3.6849,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.2,
30
+ "learning_rate": 0.00019729272419627753,
31
+ "loss": 3.6122,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.2,
36
+ "eval_accuracy": 0.10967741935483871,
37
+ "eval_loss": 3.561206340789795,
38
+ "eval_runtime": 31.2002,
39
+ "eval_samples_per_second": 54.647,
40
+ "eval_steps_per_second": 6.859,
41
+ "step": 40
42
+ },
43
+ {
44
+ "epoch": 0.25,
45
+ "learning_rate": 0.00019661590524534686,
46
+ "loss": 3.5748,
47
+ "step": 50
48
+ },
49
+ {
50
+ "epoch": 0.3,
51
+ "learning_rate": 0.00019593908629441626,
52
+ "loss": 3.491,
53
+ "step": 60
54
+ },
55
+ {
56
+ "epoch": 0.36,
57
+ "learning_rate": 0.00019526226734348562,
58
+ "loss": 3.3556,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.41,
63
+ "learning_rate": 0.000194585448392555,
64
+ "loss": 3.4008,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.41,
69
+ "eval_accuracy": 0.2475073313782991,
70
+ "eval_loss": 3.2414700984954834,
71
+ "eval_runtime": 30.544,
72
+ "eval_samples_per_second": 55.821,
73
+ "eval_steps_per_second": 7.006,
74
+ "step": 80
75
+ },
76
+ {
77
+ "epoch": 0.46,
78
+ "learning_rate": 0.00019390862944162438,
79
+ "loss": 3.2255,
80
+ "step": 90
81
+ },
82
+ {
83
+ "epoch": 0.51,
84
+ "learning_rate": 0.00019323181049069374,
85
+ "loss": 3.1171,
86
+ "step": 100
87
+ },
88
+ {
89
+ "epoch": 0.56,
90
+ "learning_rate": 0.00019255499153976314,
91
+ "loss": 3.1394,
92
+ "step": 110
93
+ },
94
+ {
95
+ "epoch": 0.61,
96
+ "learning_rate": 0.0001918781725888325,
97
+ "loss": 3.0872,
98
+ "step": 120
99
+ },
100
+ {
101
+ "epoch": 0.61,
102
+ "eval_accuracy": 0.2809384164222874,
103
+ "eval_loss": 2.9989874362945557,
104
+ "eval_runtime": 30.8814,
105
+ "eval_samples_per_second": 55.211,
106
+ "eval_steps_per_second": 6.93,
107
+ "step": 120
108
+ },
109
+ {
110
+ "epoch": 0.66,
111
+ "learning_rate": 0.00019120135363790187,
112
+ "loss": 2.9773,
113
+ "step": 130
114
+ },
115
+ {
116
+ "epoch": 0.71,
117
+ "learning_rate": 0.00019052453468697123,
118
+ "loss": 2.7993,
119
+ "step": 140
120
+ },
121
+ {
122
+ "epoch": 0.76,
123
+ "learning_rate": 0.0001898477157360406,
124
+ "loss": 2.8672,
125
+ "step": 150
126
+ },
127
+ {
128
+ "epoch": 0.81,
129
+ "learning_rate": 0.00018917089678511,
130
+ "loss": 2.8733,
131
+ "step": 160
132
+ },
133
+ {
134
+ "epoch": 0.81,
135
+ "eval_accuracy": 0.34252199413489737,
136
+ "eval_loss": 2.7833735942840576,
137
+ "eval_runtime": 30.3566,
138
+ "eval_samples_per_second": 56.166,
139
+ "eval_steps_per_second": 7.05,
140
+ "step": 160
141
+ },
142
+ {
143
+ "epoch": 0.86,
144
+ "learning_rate": 0.00018849407783417936,
145
+ "loss": 2.7912,
146
+ "step": 170
147
+ },
148
+ {
149
+ "epoch": 0.91,
150
+ "learning_rate": 0.00018781725888324875,
151
+ "loss": 2.7577,
152
+ "step": 180
153
+ },
154
+ {
155
+ "epoch": 0.96,
156
+ "learning_rate": 0.00018714043993231811,
157
+ "loss": 2.7865,
158
+ "step": 190
159
+ },
160
+ {
161
+ "epoch": 1.02,
162
+ "learning_rate": 0.00018646362098138748,
163
+ "loss": 2.591,
164
+ "step": 200
165
+ },
166
+ {
167
+ "epoch": 1.02,
168
+ "eval_accuracy": 0.37360703812316715,
169
+ "eval_loss": 2.6250743865966797,
170
+ "eval_runtime": 30.4336,
171
+ "eval_samples_per_second": 56.024,
172
+ "eval_steps_per_second": 7.032,
173
+ "step": 200
174
+ },
175
+ {
176
+ "epoch": 1.07,
177
+ "learning_rate": 0.00018578680203045687,
178
+ "loss": 2.3103,
179
+ "step": 210
180
+ },
181
+ {
182
+ "epoch": 1.12,
183
+ "learning_rate": 0.00018510998307952624,
184
+ "loss": 2.2942,
185
+ "step": 220
186
+ },
187
+ {
188
+ "epoch": 1.17,
189
+ "learning_rate": 0.0001844331641285956,
190
+ "loss": 2.3883,
191
+ "step": 230
192
+ },
193
+ {
194
+ "epoch": 1.22,
195
+ "learning_rate": 0.00018375634517766497,
196
+ "loss": 2.1643,
197
+ "step": 240
198
+ },
199
+ {
200
+ "epoch": 1.22,
201
+ "eval_accuracy": 0.40527859237536656,
202
+ "eval_loss": 2.423466920852661,
203
+ "eval_runtime": 30.3039,
204
+ "eval_samples_per_second": 56.263,
205
+ "eval_steps_per_second": 7.062,
206
+ "step": 240
207
+ },
208
+ {
209
+ "epoch": 1.27,
210
+ "learning_rate": 0.00018307952622673436,
211
+ "loss": 2.2077,
212
+ "step": 250
213
+ },
214
+ {
215
+ "epoch": 1.32,
216
+ "learning_rate": 0.00018240270727580373,
217
+ "loss": 2.1227,
218
+ "step": 260
219
+ },
220
+ {
221
+ "epoch": 1.37,
222
+ "learning_rate": 0.0001817258883248731,
223
+ "loss": 2.0657,
224
+ "step": 270
225
+ },
226
+ {
227
+ "epoch": 1.42,
228
+ "learning_rate": 0.00018104906937394248,
229
+ "loss": 2.0669,
230
+ "step": 280
231
+ },
232
+ {
233
+ "epoch": 1.42,
234
+ "eval_accuracy": 0.3882697947214076,
235
+ "eval_loss": 2.3226494789123535,
236
+ "eval_runtime": 30.2544,
237
+ "eval_samples_per_second": 56.355,
238
+ "eval_steps_per_second": 7.073,
239
+ "step": 280
240
+ },
241
+ {
242
+ "epoch": 1.47,
243
+ "learning_rate": 0.00018037225042301185,
244
+ "loss": 2.1189,
245
+ "step": 290
246
+ },
247
+ {
248
+ "epoch": 1.52,
249
+ "learning_rate": 0.00017969543147208124,
250
+ "loss": 2.1857,
251
+ "step": 300
252
+ },
253
+ {
254
+ "epoch": 1.57,
255
+ "learning_rate": 0.0001790186125211506,
256
+ "loss": 2.272,
257
+ "step": 310
258
+ },
259
+ {
260
+ "epoch": 1.62,
261
+ "learning_rate": 0.00017834179357021997,
262
+ "loss": 1.9717,
263
+ "step": 320
264
+ },
265
+ {
266
+ "epoch": 1.62,
267
+ "eval_accuracy": 0.40821114369501466,
268
+ "eval_loss": 2.2509706020355225,
269
+ "eval_runtime": 30.2595,
270
+ "eval_samples_per_second": 56.346,
271
+ "eval_steps_per_second": 7.072,
272
+ "step": 320
273
+ },
274
+ {
275
+ "epoch": 1.68,
276
+ "learning_rate": 0.00017766497461928934,
277
+ "loss": 2.0974,
278
+ "step": 330
279
+ },
280
+ {
281
+ "epoch": 1.73,
282
+ "learning_rate": 0.0001769881556683587,
283
+ "loss": 2.1316,
284
+ "step": 340
285
+ },
286
+ {
287
+ "epoch": 1.78,
288
+ "learning_rate": 0.0001763113367174281,
289
+ "loss": 2.0845,
290
+ "step": 350
291
+ },
292
+ {
293
+ "epoch": 1.83,
294
+ "learning_rate": 0.00017563451776649746,
295
+ "loss": 2.1714,
296
+ "step": 360
297
+ },
298
+ {
299
+ "epoch": 1.83,
300
+ "eval_accuracy": 0.4351906158357771,
301
+ "eval_loss": 2.203674077987671,
302
+ "eval_runtime": 30.4099,
303
+ "eval_samples_per_second": 56.067,
304
+ "eval_steps_per_second": 7.037,
305
+ "step": 360
306
+ },
307
+ {
308
+ "epoch": 1.88,
309
+ "learning_rate": 0.00017495769881556685,
310
+ "loss": 2.0881,
311
+ "step": 370
312
+ },
313
+ {
314
+ "epoch": 1.93,
315
+ "learning_rate": 0.00017428087986463622,
316
+ "loss": 1.9428,
317
+ "step": 380
318
+ },
319
+ {
320
+ "epoch": 1.98,
321
+ "learning_rate": 0.0001736040609137056,
322
+ "loss": 1.8203,
323
+ "step": 390
324
+ },
325
+ {
326
+ "epoch": 2.03,
327
+ "learning_rate": 0.00017292724196277498,
328
+ "loss": 1.6183,
329
+ "step": 400
330
+ },
331
+ {
332
+ "epoch": 2.03,
333
+ "eval_accuracy": 0.4533724340175953,
334
+ "eval_loss": 2.0731186866760254,
335
+ "eval_runtime": 30.3675,
336
+ "eval_samples_per_second": 56.146,
337
+ "eval_steps_per_second": 7.047,
338
+ "step": 400
339
+ },
340
+ {
341
+ "epoch": 2.08,
342
+ "learning_rate": 0.00017225042301184434,
343
+ "loss": 1.3395,
344
+ "step": 410
345
+ },
346
+ {
347
+ "epoch": 2.13,
348
+ "learning_rate": 0.0001715736040609137,
349
+ "loss": 1.4052,
350
+ "step": 420
351
+ },
352
+ {
353
+ "epoch": 2.18,
354
+ "learning_rate": 0.00017089678510998307,
355
+ "loss": 1.3804,
356
+ "step": 430
357
+ },
358
+ {
359
+ "epoch": 2.23,
360
+ "learning_rate": 0.00017021996615905247,
361
+ "loss": 1.2771,
362
+ "step": 440
363
+ },
364
+ {
365
+ "epoch": 2.23,
366
+ "eval_accuracy": 0.42052785923753666,
367
+ "eval_loss": 2.0928714275360107,
368
+ "eval_runtime": 30.2484,
369
+ "eval_samples_per_second": 56.367,
370
+ "eval_steps_per_second": 7.075,
371
+ "step": 440
372
+ },
373
+ {
374
+ "epoch": 2.28,
375
+ "learning_rate": 0.00016954314720812183,
376
+ "loss": 1.431,
377
+ "step": 450
378
+ },
379
+ {
380
+ "epoch": 2.34,
381
+ "learning_rate": 0.00016886632825719122,
382
+ "loss": 1.353,
383
+ "step": 460
384
+ },
385
+ {
386
+ "epoch": 2.39,
387
+ "learning_rate": 0.0001681895093062606,
388
+ "loss": 1.307,
389
+ "step": 470
390
+ },
391
+ {
392
+ "epoch": 2.44,
393
+ "learning_rate": 0.00016751269035532995,
394
+ "loss": 1.2116,
395
+ "step": 480
396
+ },
397
+ {
398
+ "epoch": 2.44,
399
+ "eval_accuracy": 0.4504398826979472,
400
+ "eval_loss": 2.0171947479248047,
401
+ "eval_runtime": 30.2989,
402
+ "eval_samples_per_second": 56.273,
403
+ "eval_steps_per_second": 7.063,
404
+ "step": 480
405
+ },
406
+ {
407
+ "epoch": 2.49,
408
+ "learning_rate": 0.00016683587140439935,
409
+ "loss": 1.2762,
410
+ "step": 490
411
+ },
412
+ {
413
+ "epoch": 2.54,
414
+ "learning_rate": 0.0001661590524534687,
415
+ "loss": 1.2253,
416
+ "step": 500
417
+ },
418
+ {
419
+ "epoch": 2.59,
420
+ "learning_rate": 0.00016548223350253808,
421
+ "loss": 1.238,
422
+ "step": 510
423
+ },
424
+ {
425
+ "epoch": 2.64,
426
+ "learning_rate": 0.00016480541455160744,
427
+ "loss": 1.158,
428
+ "step": 520
429
+ },
430
+ {
431
+ "epoch": 2.64,
432
+ "eval_accuracy": 0.45689149560117304,
433
+ "eval_loss": 1.9670987129211426,
434
+ "eval_runtime": 30.2767,
435
+ "eval_samples_per_second": 56.314,
436
+ "eval_steps_per_second": 7.068,
437
+ "step": 520
438
+ },
439
+ {
440
+ "epoch": 2.69,
441
+ "learning_rate": 0.00016412859560067684,
442
+ "loss": 1.1488,
443
+ "step": 530
444
+ },
445
+ {
446
+ "epoch": 2.74,
447
+ "learning_rate": 0.0001634517766497462,
448
+ "loss": 1.294,
449
+ "step": 540
450
+ },
451
+ {
452
+ "epoch": 2.79,
453
+ "learning_rate": 0.00016277495769881557,
454
+ "loss": 1.0431,
455
+ "step": 550
456
+ },
457
+ {
458
+ "epoch": 2.84,
459
+ "learning_rate": 0.00016209813874788496,
460
+ "loss": 1.1864,
461
+ "step": 560
462
+ },
463
+ {
464
+ "epoch": 2.84,
465
+ "eval_accuracy": 0.4563049853372434,
466
+ "eval_loss": 1.9570399522781372,
467
+ "eval_runtime": 30.2455,
468
+ "eval_samples_per_second": 56.372,
469
+ "eval_steps_per_second": 7.075,
470
+ "step": 560
471
+ },
472
+ {
473
+ "epoch": 2.89,
474
+ "learning_rate": 0.00016142131979695432,
475
+ "loss": 1.1271,
476
+ "step": 570
477
+ },
478
+ {
479
+ "epoch": 2.94,
480
+ "learning_rate": 0.00016074450084602372,
481
+ "loss": 1.1006,
482
+ "step": 580
483
+ },
484
+ {
485
+ "epoch": 2.99,
486
+ "learning_rate": 0.00016006768189509308,
487
+ "loss": 1.1867,
488
+ "step": 590
489
+ },
490
+ {
491
+ "epoch": 3.05,
492
+ "learning_rate": 0.00015939086294416242,
493
+ "loss": 0.7778,
494
+ "step": 600
495
+ },
496
+ {
497
+ "epoch": 3.05,
498
+ "eval_accuracy": 0.49560117302052786,
499
+ "eval_loss": 1.8706681728363037,
500
+ "eval_runtime": 30.3606,
501
+ "eval_samples_per_second": 56.158,
502
+ "eval_steps_per_second": 7.049,
503
+ "step": 600
504
+ },
505
+ {
506
+ "epoch": 3.1,
507
+ "learning_rate": 0.0001587140439932318,
508
+ "loss": 0.6935,
509
+ "step": 610
510
+ },
511
+ {
512
+ "epoch": 3.15,
513
+ "learning_rate": 0.00015803722504230118,
514
+ "loss": 0.5692,
515
+ "step": 620
516
+ },
517
+ {
518
+ "epoch": 3.2,
519
+ "learning_rate": 0.00015736040609137057,
520
+ "loss": 0.5985,
521
+ "step": 630
522
+ },
523
+ {
524
+ "epoch": 3.25,
525
+ "learning_rate": 0.00015668358714043994,
526
+ "loss": 0.6021,
527
+ "step": 640
528
+ },
529
+ {
530
+ "epoch": 3.25,
531
+ "eval_accuracy": 0.4950146627565982,
532
+ "eval_loss": 1.8420542478561401,
533
+ "eval_runtime": 30.3002,
534
+ "eval_samples_per_second": 56.27,
535
+ "eval_steps_per_second": 7.063,
536
+ "step": 640
537
+ },
538
+ {
539
+ "epoch": 3.3,
540
+ "learning_rate": 0.00015600676818950933,
541
+ "loss": 0.5766,
542
+ "step": 650
543
+ },
544
+ {
545
+ "epoch": 3.35,
546
+ "learning_rate": 0.0001553299492385787,
547
+ "loss": 0.7232,
548
+ "step": 660
549
+ },
550
+ {
551
+ "epoch": 3.4,
552
+ "learning_rate": 0.00015465313028764806,
553
+ "loss": 0.5083,
554
+ "step": 670
555
+ },
556
+ {
557
+ "epoch": 3.45,
558
+ "learning_rate": 0.00015397631133671742,
559
+ "loss": 0.5362,
560
+ "step": 680
561
+ },
562
+ {
563
+ "epoch": 3.45,
564
+ "eval_accuracy": 0.4756598240469208,
565
+ "eval_loss": 1.9472501277923584,
566
+ "eval_runtime": 30.3609,
567
+ "eval_samples_per_second": 56.158,
568
+ "eval_steps_per_second": 7.049,
569
+ "step": 680
570
+ },
571
+ {
572
+ "epoch": 3.5,
573
+ "learning_rate": 0.0001532994923857868,
574
+ "loss": 0.6784,
575
+ "step": 690
576
+ },
577
+ {
578
+ "epoch": 3.55,
579
+ "learning_rate": 0.00015262267343485618,
580
+ "loss": 0.5623,
581
+ "step": 700
582
+ },
583
+ {
584
+ "epoch": 3.6,
585
+ "learning_rate": 0.00015194585448392555,
586
+ "loss": 0.5619,
587
+ "step": 710
588
+ },
589
+ {
590
+ "epoch": 3.65,
591
+ "learning_rate": 0.00015126903553299494,
592
+ "loss": 0.5469,
593
+ "step": 720
594
+ },
595
+ {
596
+ "epoch": 3.65,
597
+ "eval_accuracy": 0.4621700879765396,
598
+ "eval_loss": 1.9724935293197632,
599
+ "eval_runtime": 30.833,
600
+ "eval_samples_per_second": 55.298,
601
+ "eval_steps_per_second": 6.941,
602
+ "step": 720
603
+ },
604
+ {
605
+ "epoch": 3.71,
606
+ "learning_rate": 0.0001505922165820643,
607
+ "loss": 0.6058,
608
+ "step": 730
609
+ },
610
+ {
611
+ "epoch": 3.76,
612
+ "learning_rate": 0.00014991539763113367,
613
+ "loss": 0.5332,
614
+ "step": 740
615
+ },
616
+ {
617
+ "epoch": 3.81,
618
+ "learning_rate": 0.00014923857868020306,
619
+ "loss": 0.5891,
620
+ "step": 750
621
+ },
622
+ {
623
+ "epoch": 3.86,
624
+ "learning_rate": 0.00014856175972927243,
625
+ "loss": 0.5486,
626
+ "step": 760
627
+ },
628
+ {
629
+ "epoch": 3.86,
630
+ "eval_accuracy": 0.4744868035190616,
631
+ "eval_loss": 1.9740947484970093,
632
+ "eval_runtime": 30.1738,
633
+ "eval_samples_per_second": 56.506,
634
+ "eval_steps_per_second": 7.092,
635
+ "step": 760
636
+ },
637
+ {
638
+ "epoch": 3.91,
639
+ "learning_rate": 0.0001478849407783418,
640
+ "loss": 0.624,
641
+ "step": 770
642
+ },
643
+ {
644
+ "epoch": 3.96,
645
+ "learning_rate": 0.00014720812182741116,
646
+ "loss": 0.5719,
647
+ "step": 780
648
+ },
649
+ {
650
+ "epoch": 4.01,
651
+ "learning_rate": 0.00014653130287648055,
652
+ "loss": 0.6461,
653
+ "step": 790
654
+ },
655
+ {
656
+ "epoch": 4.06,
657
+ "learning_rate": 0.00014585448392554992,
658
+ "loss": 0.339,
659
+ "step": 800
660
+ },
661
+ {
662
+ "epoch": 4.06,
663
+ "eval_accuracy": 0.47038123167155427,
664
+ "eval_loss": 1.9945420026779175,
665
+ "eval_runtime": 30.3298,
666
+ "eval_samples_per_second": 56.215,
667
+ "eval_steps_per_second": 7.056,
668
+ "step": 800
669
+ },
670
+ {
671
+ "epoch": 4.11,
672
+ "learning_rate": 0.00014517766497461928,
673
+ "loss": 0.2895,
674
+ "step": 810
675
+ },
676
+ {
677
+ "epoch": 4.16,
678
+ "learning_rate": 0.00014450084602368868,
679
+ "loss": 0.3324,
680
+ "step": 820
681
+ },
682
+ {
683
+ "epoch": 4.21,
684
+ "learning_rate": 0.00014382402707275804,
685
+ "loss": 0.3296,
686
+ "step": 830
687
+ },
688
+ {
689
+ "epoch": 4.26,
690
+ "learning_rate": 0.00014314720812182743,
691
+ "loss": 0.2798,
692
+ "step": 840
693
+ },
694
+ {
695
+ "epoch": 4.26,
696
+ "eval_accuracy": 0.46686217008797654,
697
+ "eval_loss": 2.03950572013855,
698
+ "eval_runtime": 30.2799,
699
+ "eval_samples_per_second": 56.308,
700
+ "eval_steps_per_second": 7.067,
701
+ "step": 840
702
+ },
703
+ {
704
+ "epoch": 4.31,
705
+ "learning_rate": 0.0001424703891708968,
706
+ "loss": 0.254,
707
+ "step": 850
708
+ },
709
+ {
710
+ "epoch": 4.37,
711
+ "learning_rate": 0.00014179357021996616,
712
+ "loss": 0.2533,
713
+ "step": 860
714
+ },
715
+ {
716
+ "epoch": 4.42,
717
+ "learning_rate": 0.00014111675126903553,
718
+ "loss": 0.1897,
719
+ "step": 870
720
+ },
721
+ {
722
+ "epoch": 4.47,
723
+ "learning_rate": 0.0001404399323181049,
724
+ "loss": 0.2813,
725
+ "step": 880
726
+ },
727
+ {
728
+ "epoch": 4.47,
729
+ "eval_accuracy": 0.47390029325513194,
730
+ "eval_loss": 2.0258262157440186,
731
+ "eval_runtime": 30.5513,
732
+ "eval_samples_per_second": 55.808,
733
+ "eval_steps_per_second": 7.005,
734
+ "step": 880
735
+ },
736
+ {
737
+ "epoch": 4.52,
738
+ "learning_rate": 0.0001397631133671743,
739
+ "loss": 0.3286,
740
+ "step": 890
741
+ },
742
+ {
743
+ "epoch": 4.57,
744
+ "learning_rate": 0.00013908629441624365,
745
+ "loss": 0.2575,
746
+ "step": 900
747
+ },
748
+ {
749
+ "epoch": 4.62,
750
+ "learning_rate": 0.00013840947546531305,
751
+ "loss": 0.201,
752
+ "step": 910
753
+ },
754
+ {
755
+ "epoch": 4.67,
756
+ "learning_rate": 0.0001377326565143824,
757
+ "loss": 0.209,
758
+ "step": 920
759
+ },
760
+ {
761
+ "epoch": 4.67,
762
+ "eval_accuracy": 0.47624633431085045,
763
+ "eval_loss": 2.0693771839141846,
764
+ "eval_runtime": 30.3252,
765
+ "eval_samples_per_second": 56.224,
766
+ "eval_steps_per_second": 7.057,
767
+ "step": 920
768
+ },
769
+ {
770
+ "epoch": 4.72,
771
+ "learning_rate": 0.00013705583756345178,
772
+ "loss": 0.2873,
773
+ "step": 930
774
+ },
775
+ {
776
+ "epoch": 4.77,
777
+ "learning_rate": 0.00013637901861252117,
778
+ "loss": 0.33,
779
+ "step": 940
780
+ },
781
+ {
782
+ "epoch": 4.82,
783
+ "learning_rate": 0.00013570219966159053,
784
+ "loss": 0.2578,
785
+ "step": 950
786
+ },
787
+ {
788
+ "epoch": 4.87,
789
+ "learning_rate": 0.0001350253807106599,
790
+ "loss": 0.3733,
791
+ "step": 960
792
+ },
793
+ {
794
+ "epoch": 4.87,
795
+ "eval_accuracy": 0.4797653958944281,
796
+ "eval_loss": 2.0537850856781006,
797
+ "eval_runtime": 30.0679,
798
+ "eval_samples_per_second": 56.705,
799
+ "eval_steps_per_second": 7.117,
800
+ "step": 960
801
+ },
802
+ {
803
+ "epoch": 4.92,
804
+ "learning_rate": 0.00013434856175972926,
805
+ "loss": 0.2922,
806
+ "step": 970
807
+ },
808
+ {
809
+ "epoch": 4.97,
810
+ "learning_rate": 0.00013367174280879866,
811
+ "loss": 0.2236,
812
+ "step": 980
813
+ },
814
+ {
815
+ "epoch": 5.03,
816
+ "learning_rate": 0.00013299492385786802,
817
+ "loss": 0.2532,
818
+ "step": 990
819
+ },
820
+ {
821
+ "epoch": 5.08,
822
+ "learning_rate": 0.0001323181049069374,
823
+ "loss": 0.1729,
824
+ "step": 1000
825
+ },
826
+ {
827
+ "epoch": 5.08,
828
+ "eval_accuracy": 0.4563049853372434,
829
+ "eval_loss": 2.179201126098633,
830
+ "eval_runtime": 30.3122,
831
+ "eval_samples_per_second": 56.248,
832
+ "eval_steps_per_second": 7.06,
833
+ "step": 1000
834
+ },
835
+ {
836
+ "epoch": 5.13,
837
+ "learning_rate": 0.00013164128595600678,
838
+ "loss": 0.1751,
839
+ "step": 1010
840
+ },
841
+ {
842
+ "epoch": 5.18,
843
+ "learning_rate": 0.00013096446700507615,
844
+ "loss": 0.1609,
845
+ "step": 1020
846
+ },
847
+ {
848
+ "epoch": 5.23,
849
+ "learning_rate": 0.00013028764805414554,
850
+ "loss": 0.1254,
851
+ "step": 1030
852
+ },
853
+ {
854
+ "epoch": 5.28,
855
+ "learning_rate": 0.0001296108291032149,
856
+ "loss": 0.0891,
857
+ "step": 1040
858
+ },
859
+ {
860
+ "epoch": 5.28,
861
+ "eval_accuracy": 0.4715542521994135,
862
+ "eval_loss": 2.140515089035034,
863
+ "eval_runtime": 30.4734,
864
+ "eval_samples_per_second": 55.95,
865
+ "eval_steps_per_second": 7.023,
866
+ "step": 1040
867
+ },
868
+ {
869
+ "epoch": 5.33,
870
+ "learning_rate": 0.00012893401015228427,
871
+ "loss": 0.0719,
872
+ "step": 1050
873
+ },
874
+ {
875
+ "epoch": 5.38,
876
+ "learning_rate": 0.00012825719120135363,
877
+ "loss": 0.1418,
878
+ "step": 1060
879
+ },
880
+ {
881
+ "epoch": 5.43,
882
+ "learning_rate": 0.000127580372250423,
883
+ "loss": 0.1633,
884
+ "step": 1070
885
+ },
886
+ {
887
+ "epoch": 5.48,
888
+ "learning_rate": 0.0001269035532994924,
889
+ "loss": 0.114,
890
+ "step": 1080
891
+ },
892
+ {
893
+ "epoch": 5.48,
894
+ "eval_accuracy": 0.4774193548387097,
895
+ "eval_loss": 2.172985315322876,
896
+ "eval_runtime": 30.271,
897
+ "eval_samples_per_second": 56.325,
898
+ "eval_steps_per_second": 7.069,
899
+ "step": 1080
900
+ },
901
+ {
902
+ "epoch": 5.53,
903
+ "learning_rate": 0.00012622673434856176,
904
+ "loss": 0.1615,
905
+ "step": 1090
906
+ },
907
+ {
908
+ "epoch": 5.58,
909
+ "learning_rate": 0.00012554991539763115,
910
+ "loss": 0.1589,
911
+ "step": 1100
912
+ },
913
+ {
914
+ "epoch": 5.63,
915
+ "learning_rate": 0.00012487309644670052,
916
+ "loss": 0.1249,
917
+ "step": 1110
918
+ },
919
+ {
920
+ "epoch": 5.69,
921
+ "learning_rate": 0.0001241962774957699,
922
+ "loss": 0.207,
923
+ "step": 1120
924
+ },
925
+ {
926
+ "epoch": 5.69,
927
+ "eval_accuracy": 0.4475073313782991,
928
+ "eval_loss": 2.32855486869812,
929
+ "eval_runtime": 30.3218,
930
+ "eval_samples_per_second": 56.23,
931
+ "eval_steps_per_second": 7.058,
932
+ "step": 1120
933
+ },
934
+ {
935
+ "epoch": 5.74,
936
+ "learning_rate": 0.00012351945854483927,
937
+ "loss": 0.1255,
938
+ "step": 1130
939
+ },
940
+ {
941
+ "epoch": 5.79,
942
+ "learning_rate": 0.00012284263959390864,
943
+ "loss": 0.0847,
944
+ "step": 1140
945
+ },
946
+ {
947
+ "epoch": 5.84,
948
+ "learning_rate": 0.000122165820642978,
949
+ "loss": 0.2418,
950
+ "step": 1150
951
+ },
952
+ {
953
+ "epoch": 5.89,
954
+ "learning_rate": 0.00012148900169204738,
955
+ "loss": 0.0984,
956
+ "step": 1160
957
+ },
958
+ {
959
+ "epoch": 5.89,
960
+ "eval_accuracy": 0.4826979472140763,
961
+ "eval_loss": 2.280658006668091,
962
+ "eval_runtime": 30.3832,
963
+ "eval_samples_per_second": 56.116,
964
+ "eval_steps_per_second": 7.043,
965
+ "step": 1160
966
+ },
967
+ {
968
+ "epoch": 5.94,
969
+ "learning_rate": 0.00012081218274111676,
970
+ "loss": 0.2079,
971
+ "step": 1170
972
+ },
973
+ {
974
+ "epoch": 5.99,
975
+ "learning_rate": 0.00012013536379018613,
976
+ "loss": 0.1239,
977
+ "step": 1180
978
+ },
979
+ {
980
+ "epoch": 6.04,
981
+ "learning_rate": 0.00011945854483925552,
982
+ "loss": 0.1165,
983
+ "step": 1190
984
+ },
985
+ {
986
+ "epoch": 6.09,
987
+ "learning_rate": 0.00011878172588832489,
988
+ "loss": 0.0544,
989
+ "step": 1200
990
+ },
991
+ {
992
+ "epoch": 6.09,
993
+ "eval_accuracy": 0.44868035190615835,
994
+ "eval_loss": 2.3984222412109375,
995
+ "eval_runtime": 30.0364,
996
+ "eval_samples_per_second": 56.764,
997
+ "eval_steps_per_second": 7.125,
998
+ "step": 1200
999
+ },
1000
+ {
1001
+ "epoch": 6.14,
1002
+ "learning_rate": 0.00011810490693739425,
1003
+ "loss": 0.1029,
1004
+ "step": 1210
1005
+ },
1006
+ {
1007
+ "epoch": 6.19,
1008
+ "learning_rate": 0.00011742808798646363,
1009
+ "loss": 0.1151,
1010
+ "step": 1220
1011
+ },
1012
+ {
1013
+ "epoch": 6.24,
1014
+ "learning_rate": 0.000116751269035533,
1015
+ "loss": 0.0611,
1016
+ "step": 1230
1017
+ },
1018
+ {
1019
+ "epoch": 6.29,
1020
+ "learning_rate": 0.00011607445008460239,
1021
+ "loss": 0.1109,
1022
+ "step": 1240
1023
+ },
1024
+ {
1025
+ "epoch": 6.29,
1026
+ "eval_accuracy": 0.4780058651026393,
1027
+ "eval_loss": 2.314086437225342,
1028
+ "eval_runtime": 30.0544,
1029
+ "eval_samples_per_second": 56.73,
1030
+ "eval_steps_per_second": 7.12,
1031
+ "step": 1240
1032
+ },
1033
+ {
1034
+ "epoch": 6.35,
1035
+ "learning_rate": 0.00011539763113367175,
1036
+ "loss": 0.064,
1037
+ "step": 1250
1038
+ },
1039
+ {
1040
+ "epoch": 6.4,
1041
+ "learning_rate": 0.00011472081218274113,
1042
+ "loss": 0.0555,
1043
+ "step": 1260
1044
+ },
1045
+ {
1046
+ "epoch": 6.45,
1047
+ "learning_rate": 0.0001140439932318105,
1048
+ "loss": 0.0774,
1049
+ "step": 1270
1050
+ },
1051
+ {
1052
+ "epoch": 6.5,
1053
+ "learning_rate": 0.00011336717428087986,
1054
+ "loss": 0.121,
1055
+ "step": 1280
1056
+ },
1057
+ {
1058
+ "epoch": 6.5,
1059
+ "eval_accuracy": 0.49032258064516127,
1060
+ "eval_loss": 2.225583791732788,
1061
+ "eval_runtime": 30.2835,
1062
+ "eval_samples_per_second": 56.301,
1063
+ "eval_steps_per_second": 7.067,
1064
+ "step": 1280
1065
+ },
1066
+ {
1067
+ "epoch": 6.55,
1068
+ "learning_rate": 0.00011269035532994925,
1069
+ "loss": 0.0801,
1070
+ "step": 1290
1071
+ },
1072
+ {
1073
+ "epoch": 6.6,
1074
+ "learning_rate": 0.0001120135363790186,
1075
+ "loss": 0.045,
1076
+ "step": 1300
1077
+ },
1078
+ {
1079
+ "epoch": 6.65,
1080
+ "learning_rate": 0.000111336717428088,
1081
+ "loss": 0.0725,
1082
+ "step": 1310
1083
+ },
1084
+ {
1085
+ "epoch": 6.7,
1086
+ "learning_rate": 0.00011065989847715736,
1087
+ "loss": 0.0387,
1088
+ "step": 1320
1089
+ },
1090
+ {
1091
+ "epoch": 6.7,
1092
+ "eval_accuracy": 0.5032258064516129,
1093
+ "eval_loss": 2.191972494125366,
1094
+ "eval_runtime": 30.6293,
1095
+ "eval_samples_per_second": 55.666,
1096
+ "eval_steps_per_second": 6.987,
1097
+ "step": 1320
1098
+ },
1099
+ {
1100
+ "epoch": 6.75,
1101
+ "learning_rate": 0.00010998307952622673,
1102
+ "loss": 0.0376,
1103
+ "step": 1330
1104
+ },
1105
+ {
1106
+ "epoch": 6.8,
1107
+ "learning_rate": 0.00010930626057529611,
1108
+ "loss": 0.082,
1109
+ "step": 1340
1110
+ },
1111
+ {
1112
+ "epoch": 6.85,
1113
+ "learning_rate": 0.00010862944162436547,
1114
+ "loss": 0.026,
1115
+ "step": 1350
1116
+ },
1117
+ {
1118
+ "epoch": 6.9,
1119
+ "learning_rate": 0.00010795262267343487,
1120
+ "loss": 0.0469,
1121
+ "step": 1360
1122
+ },
1123
+ {
1124
+ "epoch": 6.9,
1125
+ "eval_accuracy": 0.481524926686217,
1126
+ "eval_loss": 2.3258745670318604,
1127
+ "eval_runtime": 30.3501,
1128
+ "eval_samples_per_second": 56.178,
1129
+ "eval_steps_per_second": 7.051,
1130
+ "step": 1360
1131
+ },
1132
+ {
1133
+ "epoch": 6.95,
1134
+ "learning_rate": 0.00010727580372250423,
1135
+ "loss": 0.0221,
1136
+ "step": 1370
1137
+ },
1138
+ {
1139
+ "epoch": 7.01,
1140
+ "learning_rate": 0.00010659898477157362,
1141
+ "loss": 0.0234,
1142
+ "step": 1380
1143
+ },
1144
+ {
1145
+ "epoch": 7.06,
1146
+ "learning_rate": 0.00010592216582064298,
1147
+ "loss": 0.0536,
1148
+ "step": 1390
1149
+ },
1150
+ {
1151
+ "epoch": 7.11,
1152
+ "learning_rate": 0.00010524534686971234,
1153
+ "loss": 0.0269,
1154
+ "step": 1400
1155
+ },
1156
+ {
1157
+ "epoch": 7.11,
1158
+ "eval_accuracy": 0.49325513196480936,
1159
+ "eval_loss": 2.3082282543182373,
1160
+ "eval_runtime": 30.3956,
1161
+ "eval_samples_per_second": 56.094,
1162
+ "eval_steps_per_second": 7.04,
1163
+ "step": 1400
1164
+ },
1165
+ {
1166
+ "epoch": 7.16,
1167
+ "learning_rate": 0.00010456852791878173,
1168
+ "loss": 0.0626,
1169
+ "step": 1410
1170
+ },
1171
+ {
1172
+ "epoch": 7.21,
1173
+ "learning_rate": 0.0001038917089678511,
1174
+ "loss": 0.0338,
1175
+ "step": 1420
1176
+ },
1177
+ {
1178
+ "epoch": 7.26,
1179
+ "learning_rate": 0.00010321489001692048,
1180
+ "loss": 0.0534,
1181
+ "step": 1430
1182
+ },
1183
+ {
1184
+ "epoch": 7.31,
1185
+ "learning_rate": 0.00010253807106598984,
1186
+ "loss": 0.0482,
1187
+ "step": 1440
1188
+ },
1189
+ {
1190
+ "epoch": 7.31,
1191
+ "eval_accuracy": 0.47624633431085045,
1192
+ "eval_loss": 2.3973476886749268,
1193
+ "eval_runtime": 30.8314,
1194
+ "eval_samples_per_second": 55.301,
1195
+ "eval_steps_per_second": 6.941,
1196
+ "step": 1440
1197
+ },
1198
+ {
1199
+ "epoch": 7.36,
1200
+ "learning_rate": 0.00010186125211505924,
1201
+ "loss": 0.043,
1202
+ "step": 1450
1203
+ },
1204
+ {
1205
+ "epoch": 7.41,
1206
+ "learning_rate": 0.0001011844331641286,
1207
+ "loss": 0.0217,
1208
+ "step": 1460
1209
+ },
1210
+ {
1211
+ "epoch": 7.46,
1212
+ "learning_rate": 0.00010050761421319797,
1213
+ "loss": 0.0253,
1214
+ "step": 1470
1215
+ },
1216
+ {
1217
+ "epoch": 7.51,
1218
+ "learning_rate": 9.983079526226735e-05,
1219
+ "loss": 0.052,
1220
+ "step": 1480
1221
+ },
1222
+ {
1223
+ "epoch": 7.51,
1224
+ "eval_accuracy": 0.4750733137829912,
1225
+ "eval_loss": 2.3328230381011963,
1226
+ "eval_runtime": 30.3845,
1227
+ "eval_samples_per_second": 56.114,
1228
+ "eval_steps_per_second": 7.043,
1229
+ "step": 1480
1230
+ },
1231
+ {
1232
+ "epoch": 7.56,
1233
+ "learning_rate": 9.915397631133673e-05,
1234
+ "loss": 0.0445,
1235
+ "step": 1490
1236
+ },
1237
+ {
1238
+ "epoch": 7.61,
1239
+ "learning_rate": 9.847715736040609e-05,
1240
+ "loss": 0.0927,
1241
+ "step": 1500
1242
+ },
1243
+ {
1244
+ "epoch": 7.66,
1245
+ "learning_rate": 9.780033840947547e-05,
1246
+ "loss": 0.0418,
1247
+ "step": 1510
1248
+ },
1249
+ {
1250
+ "epoch": 7.72,
1251
+ "learning_rate": 9.712351945854485e-05,
1252
+ "loss": 0.0342,
1253
+ "step": 1520
1254
+ },
1255
+ {
1256
+ "epoch": 7.72,
1257
+ "eval_accuracy": 0.5020527859237537,
1258
+ "eval_loss": 2.301551580429077,
1259
+ "eval_runtime": 31.0377,
1260
+ "eval_samples_per_second": 54.933,
1261
+ "eval_steps_per_second": 6.895,
1262
+ "step": 1520
1263
+ },
1264
+ {
1265
+ "epoch": 7.77,
1266
+ "learning_rate": 9.644670050761421e-05,
1267
+ "loss": 0.0925,
1268
+ "step": 1530
1269
+ },
1270
+ {
1271
+ "epoch": 7.82,
1272
+ "learning_rate": 9.576988155668359e-05,
1273
+ "loss": 0.0187,
1274
+ "step": 1540
1275
+ },
1276
+ {
1277
+ "epoch": 7.87,
1278
+ "learning_rate": 9.509306260575297e-05,
1279
+ "loss": 0.0816,
1280
+ "step": 1550
1281
+ },
1282
+ {
1283
+ "epoch": 7.92,
1284
+ "learning_rate": 9.441624365482235e-05,
1285
+ "loss": 0.0374,
1286
+ "step": 1560
1287
+ },
1288
+ {
1289
+ "epoch": 7.92,
1290
+ "eval_accuracy": 0.49736070381231673,
1291
+ "eval_loss": 2.250404119491577,
1292
+ "eval_runtime": 31.1263,
1293
+ "eval_samples_per_second": 54.777,
1294
+ "eval_steps_per_second": 6.875,
1295
+ "step": 1560
1296
+ },
1297
+ {
1298
+ "epoch": 7.97,
1299
+ "learning_rate": 9.373942470389172e-05,
1300
+ "loss": 0.0521,
1301
+ "step": 1570
1302
+ },
1303
+ {
1304
+ "epoch": 8.02,
1305
+ "learning_rate": 9.306260575296108e-05,
1306
+ "loss": 0.0443,
1307
+ "step": 1580
1308
+ },
1309
+ {
1310
+ "epoch": 8.07,
1311
+ "learning_rate": 9.238578680203046e-05,
1312
+ "loss": 0.019,
1313
+ "step": 1590
1314
+ },
1315
+ {
1316
+ "epoch": 8.12,
1317
+ "learning_rate": 9.170896785109984e-05,
1318
+ "loss": 0.0169,
1319
+ "step": 1600
1320
+ },
1321
+ {
1322
+ "epoch": 8.12,
1323
+ "eval_accuracy": 0.48328445747800586,
1324
+ "eval_loss": 2.3862178325653076,
1325
+ "eval_runtime": 30.7127,
1326
+ "eval_samples_per_second": 55.514,
1327
+ "eval_steps_per_second": 6.968,
1328
+ "step": 1600
1329
+ },
1330
+ {
1331
+ "epoch": 8.17,
1332
+ "learning_rate": 9.103214890016922e-05,
1333
+ "loss": 0.0239,
1334
+ "step": 1610
1335
+ },
1336
+ {
1337
+ "epoch": 8.22,
1338
+ "learning_rate": 9.035532994923858e-05,
1339
+ "loss": 0.0145,
1340
+ "step": 1620
1341
+ },
1342
+ {
1343
+ "epoch": 8.27,
1344
+ "learning_rate": 8.967851099830795e-05,
1345
+ "loss": 0.0398,
1346
+ "step": 1630
1347
+ },
1348
+ {
1349
+ "epoch": 8.32,
1350
+ "learning_rate": 8.900169204737733e-05,
1351
+ "loss": 0.084,
1352
+ "step": 1640
1353
+ },
1354
+ {
1355
+ "epoch": 8.32,
1356
+ "eval_accuracy": 0.49442815249266864,
1357
+ "eval_loss": 2.4026331901550293,
1358
+ "eval_runtime": 30.8185,
1359
+ "eval_samples_per_second": 55.324,
1360
+ "eval_steps_per_second": 6.944,
1361
+ "step": 1640
1362
+ },
1363
+ {
1364
+ "epoch": 8.38,
1365
+ "learning_rate": 8.83248730964467e-05,
1366
+ "loss": 0.0215,
1367
+ "step": 1650
1368
+ },
1369
+ {
1370
+ "epoch": 8.43,
1371
+ "learning_rate": 8.764805414551607e-05,
1372
+ "loss": 0.0115,
1373
+ "step": 1660
1374
+ },
1375
+ {
1376
+ "epoch": 8.48,
1377
+ "learning_rate": 8.697123519458545e-05,
1378
+ "loss": 0.0331,
1379
+ "step": 1670
1380
+ },
1381
+ {
1382
+ "epoch": 8.53,
1383
+ "learning_rate": 8.629441624365483e-05,
1384
+ "loss": 0.0115,
1385
+ "step": 1680
1386
+ },
1387
+ {
1388
+ "epoch": 8.53,
1389
+ "eval_accuracy": 0.49149560117302055,
1390
+ "eval_loss": 2.382765293121338,
1391
+ "eval_runtime": 30.6052,
1392
+ "eval_samples_per_second": 55.71,
1393
+ "eval_steps_per_second": 6.992,
1394
+ "step": 1680
1395
+ },
1396
+ {
1397
+ "epoch": 8.58,
1398
+ "learning_rate": 8.561759729272421e-05,
1399
+ "loss": 0.0304,
1400
+ "step": 1690
1401
+ },
1402
+ {
1403
+ "epoch": 8.63,
1404
+ "learning_rate": 8.494077834179357e-05,
1405
+ "loss": 0.0108,
1406
+ "step": 1700
1407
+ },
1408
+ {
1409
+ "epoch": 8.68,
1410
+ "learning_rate": 8.426395939086294e-05,
1411
+ "loss": 0.0252,
1412
+ "step": 1710
1413
+ },
1414
+ {
1415
+ "epoch": 8.73,
1416
+ "learning_rate": 8.358714043993232e-05,
1417
+ "loss": 0.0285,
1418
+ "step": 1720
1419
+ },
1420
+ {
1421
+ "epoch": 8.73,
1422
+ "eval_accuracy": 0.49560117302052786,
1423
+ "eval_loss": 2.352389097213745,
1424
+ "eval_runtime": 30.3007,
1425
+ "eval_samples_per_second": 56.269,
1426
+ "eval_steps_per_second": 7.063,
1427
+ "step": 1720
1428
+ },
1429
+ {
1430
+ "epoch": 8.78,
1431
+ "learning_rate": 8.29103214890017e-05,
1432
+ "loss": 0.0605,
1433
+ "step": 1730
1434
+ },
1435
+ {
1436
+ "epoch": 8.83,
1437
+ "learning_rate": 8.223350253807108e-05,
1438
+ "loss": 0.0314,
1439
+ "step": 1740
1440
+ },
1441
+ {
1442
+ "epoch": 8.88,
1443
+ "learning_rate": 8.155668358714044e-05,
1444
+ "loss": 0.0135,
1445
+ "step": 1750
1446
+ },
1447
+ {
1448
+ "epoch": 8.93,
1449
+ "learning_rate": 8.087986463620981e-05,
1450
+ "loss": 0.1062,
1451
+ "step": 1760
1452
+ },
1453
+ {
1454
+ "epoch": 8.93,
1455
+ "eval_accuracy": 0.49149560117302055,
1456
+ "eval_loss": 2.412898063659668,
1457
+ "eval_runtime": 30.498,
1458
+ "eval_samples_per_second": 55.905,
1459
+ "eval_steps_per_second": 7.017,
1460
+ "step": 1760
1461
+ },
1462
+ {
1463
+ "epoch": 8.98,
1464
+ "learning_rate": 8.020304568527919e-05,
1465
+ "loss": 0.025,
1466
+ "step": 1770
1467
+ },
1468
+ {
1469
+ "epoch": 9.04,
1470
+ "learning_rate": 7.952622673434857e-05,
1471
+ "loss": 0.0383,
1472
+ "step": 1780
1473
+ },
1474
+ {
1475
+ "epoch": 9.09,
1476
+ "learning_rate": 7.884940778341794e-05,
1477
+ "loss": 0.0116,
1478
+ "step": 1790
1479
+ },
1480
+ {
1481
+ "epoch": 9.14,
1482
+ "learning_rate": 7.817258883248731e-05,
1483
+ "loss": 0.0113,
1484
+ "step": 1800
1485
+ },
1486
+ {
1487
+ "epoch": 9.14,
1488
+ "eval_accuracy": 0.49208211143695013,
1489
+ "eval_loss": 2.3613696098327637,
1490
+ "eval_runtime": 30.6596,
1491
+ "eval_samples_per_second": 55.611,
1492
+ "eval_steps_per_second": 6.98,
1493
+ "step": 1800
1494
+ },
1495
+ {
1496
+ "epoch": 9.19,
1497
+ "learning_rate": 7.749576988155669e-05,
1498
+ "loss": 0.0226,
1499
+ "step": 1810
1500
+ },
1501
+ {
1502
+ "epoch": 9.24,
1503
+ "learning_rate": 7.681895093062607e-05,
1504
+ "loss": 0.0203,
1505
+ "step": 1820
1506
+ },
1507
+ {
1508
+ "epoch": 9.29,
1509
+ "learning_rate": 7.614213197969543e-05,
1510
+ "loss": 0.0354,
1511
+ "step": 1830
1512
+ },
1513
+ {
1514
+ "epoch": 9.34,
1515
+ "learning_rate": 7.546531302876481e-05,
1516
+ "loss": 0.0101,
1517
+ "step": 1840
1518
+ },
1519
+ {
1520
+ "epoch": 9.34,
1521
+ "eval_accuracy": 0.5002932551319648,
1522
+ "eval_loss": 2.3706891536712646,
1523
+ "eval_runtime": 30.567,
1524
+ "eval_samples_per_second": 55.779,
1525
+ "eval_steps_per_second": 7.001,
1526
+ "step": 1840
1527
+ },
1528
+ {
1529
+ "epoch": 9.39,
1530
+ "learning_rate": 7.478849407783418e-05,
1531
+ "loss": 0.0181,
1532
+ "step": 1850
1533
+ },
1534
+ {
1535
+ "epoch": 9.44,
1536
+ "learning_rate": 7.411167512690356e-05,
1537
+ "loss": 0.0368,
1538
+ "step": 1860
1539
+ },
1540
+ {
1541
+ "epoch": 9.49,
1542
+ "learning_rate": 7.343485617597293e-05,
1543
+ "loss": 0.0082,
1544
+ "step": 1870
1545
+ },
1546
+ {
1547
+ "epoch": 9.54,
1548
+ "learning_rate": 7.275803722504231e-05,
1549
+ "loss": 0.028,
1550
+ "step": 1880
1551
+ },
1552
+ {
1553
+ "epoch": 9.54,
1554
+ "eval_accuracy": 0.4967741935483871,
1555
+ "eval_loss": 2.414238452911377,
1556
+ "eval_runtime": 30.6609,
1557
+ "eval_samples_per_second": 55.608,
1558
+ "eval_steps_per_second": 6.98,
1559
+ "step": 1880
1560
+ },
1561
+ {
1562
+ "epoch": 9.59,
1563
+ "learning_rate": 7.208121827411168e-05,
1564
+ "loss": 0.0404,
1565
+ "step": 1890
1566
+ },
1567
+ {
1568
+ "epoch": 9.64,
1569
+ "learning_rate": 7.140439932318104e-05,
1570
+ "loss": 0.0486,
1571
+ "step": 1900
1572
+ },
1573
+ {
1574
+ "epoch": 9.7,
1575
+ "learning_rate": 7.072758037225042e-05,
1576
+ "loss": 0.0197,
1577
+ "step": 1910
1578
+ },
1579
+ {
1580
+ "epoch": 9.75,
1581
+ "learning_rate": 7.00507614213198e-05,
1582
+ "loss": 0.0131,
1583
+ "step": 1920
1584
+ },
1585
+ {
1586
+ "epoch": 9.75,
1587
+ "eval_accuracy": 0.501466275659824,
1588
+ "eval_loss": 2.4358253479003906,
1589
+ "eval_runtime": 30.4732,
1590
+ "eval_samples_per_second": 55.951,
1591
+ "eval_steps_per_second": 7.023,
1592
+ "step": 1920
1593
+ },
1594
+ {
1595
+ "epoch": 9.8,
1596
+ "learning_rate": 6.937394247038918e-05,
1597
+ "loss": 0.0394,
1598
+ "step": 1930
1599
+ },
1600
+ {
1601
+ "epoch": 9.85,
1602
+ "learning_rate": 6.869712351945855e-05,
1603
+ "loss": 0.0402,
1604
+ "step": 1940
1605
+ },
1606
+ {
1607
+ "epoch": 9.9,
1608
+ "learning_rate": 6.802030456852793e-05,
1609
+ "loss": 0.0215,
1610
+ "step": 1950
1611
+ },
1612
+ {
1613
+ "epoch": 9.95,
1614
+ "learning_rate": 6.734348561759729e-05,
1615
+ "loss": 0.0115,
1616
+ "step": 1960
1617
+ },
1618
+ {
1619
+ "epoch": 9.95,
1620
+ "eval_accuracy": 0.4997067448680352,
1621
+ "eval_loss": 2.3999907970428467,
1622
+ "eval_runtime": 30.5014,
1623
+ "eval_samples_per_second": 55.899,
1624
+ "eval_steps_per_second": 7.016,
1625
+ "step": 1960
1626
+ },
1627
+ {
1628
+ "epoch": 10.0,
1629
+ "learning_rate": 6.666666666666667e-05,
1630
+ "loss": 0.0148,
1631
+ "step": 1970
1632
+ },
1633
+ {
1634
+ "epoch": 10.05,
1635
+ "learning_rate": 6.598984771573604e-05,
1636
+ "loss": 0.0075,
1637
+ "step": 1980
1638
+ },
1639
+ {
1640
+ "epoch": 10.1,
1641
+ "learning_rate": 6.531302876480541e-05,
1642
+ "loss": 0.0232,
1643
+ "step": 1990
1644
+ },
1645
+ {
1646
+ "epoch": 10.15,
1647
+ "learning_rate": 6.463620981387479e-05,
1648
+ "loss": 0.0108,
1649
+ "step": 2000
1650
+ },
1651
+ {
1652
+ "epoch": 10.15,
1653
+ "eval_accuracy": 0.5043988269794721,
1654
+ "eval_loss": 2.407888889312744,
1655
+ "eval_runtime": 30.7257,
1656
+ "eval_samples_per_second": 55.491,
1657
+ "eval_steps_per_second": 6.965,
1658
+ "step": 2000
1659
+ },
1660
+ {
1661
+ "epoch": 10.2,
1662
+ "learning_rate": 6.395939086294417e-05,
1663
+ "loss": 0.0182,
1664
+ "step": 2010
1665
+ },
1666
+ {
1667
+ "epoch": 10.25,
1668
+ "learning_rate": 6.328257191201354e-05,
1669
+ "loss": 0.0219,
1670
+ "step": 2020
1671
+ },
1672
+ {
1673
+ "epoch": 10.3,
1674
+ "learning_rate": 6.26057529610829e-05,
1675
+ "loss": 0.0277,
1676
+ "step": 2030
1677
+ },
1678
+ {
1679
+ "epoch": 10.36,
1680
+ "learning_rate": 6.192893401015228e-05,
1681
+ "loss": 0.009,
1682
+ "step": 2040
1683
+ },
1684
+ {
1685
+ "epoch": 10.36,
1686
+ "eval_accuracy": 0.5067448680351906,
1687
+ "eval_loss": 2.3866255283355713,
1688
+ "eval_runtime": 30.7008,
1689
+ "eval_samples_per_second": 55.536,
1690
+ "eval_steps_per_second": 6.971,
1691
+ "step": 2040
1692
+ },
1693
+ {
1694
+ "epoch": 10.41,
1695
+ "learning_rate": 6.125211505922166e-05,
1696
+ "loss": 0.0212,
1697
+ "step": 2050
1698
+ },
1699
+ {
1700
+ "epoch": 10.46,
1701
+ "learning_rate": 6.057529610829103e-05,
1702
+ "loss": 0.0065,
1703
+ "step": 2060
1704
+ },
1705
+ {
1706
+ "epoch": 10.51,
1707
+ "learning_rate": 5.989847715736041e-05,
1708
+ "loss": 0.0306,
1709
+ "step": 2070
1710
+ },
1711
+ {
1712
+ "epoch": 10.56,
1713
+ "learning_rate": 5.9221658206429784e-05,
1714
+ "loss": 0.0112,
1715
+ "step": 2080
1716
+ },
1717
+ {
1718
+ "epoch": 10.56,
1719
+ "eval_accuracy": 0.5085043988269795,
1720
+ "eval_loss": 2.3716347217559814,
1721
+ "eval_runtime": 30.6367,
1722
+ "eval_samples_per_second": 55.652,
1723
+ "eval_steps_per_second": 6.985,
1724
+ "step": 2080
1725
+ },
1726
+ {
1727
+ "epoch": 10.61,
1728
+ "learning_rate": 5.854483925549916e-05,
1729
+ "loss": 0.0325,
1730
+ "step": 2090
1731
+ },
1732
+ {
1733
+ "epoch": 10.66,
1734
+ "learning_rate": 5.786802030456853e-05,
1735
+ "loss": 0.0335,
1736
+ "step": 2100
1737
+ },
1738
+ {
1739
+ "epoch": 10.71,
1740
+ "learning_rate": 5.71912013536379e-05,
1741
+ "loss": 0.0209,
1742
+ "step": 2110
1743
+ },
1744
+ {
1745
+ "epoch": 10.76,
1746
+ "learning_rate": 5.651438240270728e-05,
1747
+ "loss": 0.0066,
1748
+ "step": 2120
1749
+ },
1750
+ {
1751
+ "epoch": 10.76,
1752
+ "eval_accuracy": 0.509090909090909,
1753
+ "eval_loss": 2.3734991550445557,
1754
+ "eval_runtime": 30.6695,
1755
+ "eval_samples_per_second": 55.593,
1756
+ "eval_steps_per_second": 6.978,
1757
+ "step": 2120
1758
+ },
1759
+ {
1760
+ "epoch": 10.81,
1761
+ "learning_rate": 5.583756345177665e-05,
1762
+ "loss": 0.0146,
1763
+ "step": 2130
1764
+ },
1765
+ {
1766
+ "epoch": 10.86,
1767
+ "learning_rate": 5.516074450084603e-05,
1768
+ "loss": 0.0061,
1769
+ "step": 2140
1770
+ },
1771
+ {
1772
+ "epoch": 10.91,
1773
+ "learning_rate": 5.44839255499154e-05,
1774
+ "loss": 0.0163,
1775
+ "step": 2150
1776
+ },
1777
+ {
1778
+ "epoch": 10.96,
1779
+ "learning_rate": 5.380710659898477e-05,
1780
+ "loss": 0.0164,
1781
+ "step": 2160
1782
+ },
1783
+ {
1784
+ "epoch": 10.96,
1785
+ "eval_accuracy": 0.5055718475073314,
1786
+ "eval_loss": 2.399704694747925,
1787
+ "eval_runtime": 30.8256,
1788
+ "eval_samples_per_second": 55.311,
1789
+ "eval_steps_per_second": 6.942,
1790
+ "step": 2160
1791
+ },
1792
+ {
1793
+ "epoch": 11.02,
1794
+ "learning_rate": 5.313028764805415e-05,
1795
+ "loss": 0.017,
1796
+ "step": 2170
1797
+ },
1798
+ {
1799
+ "epoch": 11.07,
1800
+ "learning_rate": 5.245346869712352e-05,
1801
+ "loss": 0.0056,
1802
+ "step": 2180
1803
+ },
1804
+ {
1805
+ "epoch": 11.12,
1806
+ "learning_rate": 5.17766497461929e-05,
1807
+ "loss": 0.0191,
1808
+ "step": 2190
1809
+ },
1810
+ {
1811
+ "epoch": 11.17,
1812
+ "learning_rate": 5.109983079526227e-05,
1813
+ "loss": 0.0126,
1814
+ "step": 2200
1815
+ },
1816
+ {
1817
+ "epoch": 11.17,
1818
+ "eval_accuracy": 0.5055718475073314,
1819
+ "eval_loss": 2.4104385375976562,
1820
+ "eval_runtime": 30.7309,
1821
+ "eval_samples_per_second": 55.482,
1822
+ "eval_steps_per_second": 6.964,
1823
+ "step": 2200
1824
+ },
1825
+ {
1826
+ "epoch": 11.22,
1827
+ "learning_rate": 5.042301184433165e-05,
1828
+ "loss": 0.0125,
1829
+ "step": 2210
1830
+ },
1831
+ {
1832
+ "epoch": 11.27,
1833
+ "learning_rate": 4.9746192893401014e-05,
1834
+ "loss": 0.011,
1835
+ "step": 2220
1836
+ },
1837
+ {
1838
+ "epoch": 11.32,
1839
+ "learning_rate": 4.906937394247039e-05,
1840
+ "loss": 0.0305,
1841
+ "step": 2230
1842
+ },
1843
+ {
1844
+ "epoch": 11.37,
1845
+ "learning_rate": 4.8392554991539766e-05,
1846
+ "loss": 0.0174,
1847
+ "step": 2240
1848
+ },
1849
+ {
1850
+ "epoch": 11.37,
1851
+ "eval_accuracy": 0.5026392961876833,
1852
+ "eval_loss": 2.443206548690796,
1853
+ "eval_runtime": 30.8601,
1854
+ "eval_samples_per_second": 55.249,
1855
+ "eval_steps_per_second": 6.935,
1856
+ "step": 2240
1857
+ },
1858
+ {
1859
+ "epoch": 11.42,
1860
+ "learning_rate": 4.771573604060914e-05,
1861
+ "loss": 0.0125,
1862
+ "step": 2250
1863
+ },
1864
+ {
1865
+ "epoch": 11.47,
1866
+ "learning_rate": 4.7038917089678517e-05,
1867
+ "loss": 0.0127,
1868
+ "step": 2260
1869
+ },
1870
+ {
1871
+ "epoch": 11.52,
1872
+ "learning_rate": 4.636209813874789e-05,
1873
+ "loss": 0.0253,
1874
+ "step": 2270
1875
+ },
1876
+ {
1877
+ "epoch": 11.57,
1878
+ "learning_rate": 4.568527918781726e-05,
1879
+ "loss": 0.0241,
1880
+ "step": 2280
1881
+ },
1882
+ {
1883
+ "epoch": 11.57,
1884
+ "eval_accuracy": 0.5055718475073314,
1885
+ "eval_loss": 2.4564218521118164,
1886
+ "eval_runtime": 30.6781,
1887
+ "eval_samples_per_second": 55.577,
1888
+ "eval_steps_per_second": 6.976,
1889
+ "step": 2280
1890
+ },
1891
+ {
1892
+ "epoch": 11.62,
1893
+ "learning_rate": 4.500846023688663e-05,
1894
+ "loss": 0.0084,
1895
+ "step": 2290
1896
+ },
1897
+ {
1898
+ "epoch": 11.68,
1899
+ "learning_rate": 4.433164128595601e-05,
1900
+ "loss": 0.0054,
1901
+ "step": 2300
1902
+ },
1903
+ {
1904
+ "epoch": 11.73,
1905
+ "learning_rate": 4.365482233502538e-05,
1906
+ "loss": 0.0055,
1907
+ "step": 2310
1908
+ },
1909
+ {
1910
+ "epoch": 11.78,
1911
+ "learning_rate": 4.2978003384094756e-05,
1912
+ "loss": 0.0182,
1913
+ "step": 2320
1914
+ },
1915
+ {
1916
+ "epoch": 11.78,
1917
+ "eval_accuracy": 0.5038123167155425,
1918
+ "eval_loss": 2.481368064880371,
1919
+ "eval_runtime": 30.9495,
1920
+ "eval_samples_per_second": 55.09,
1921
+ "eval_steps_per_second": 6.914,
1922
+ "step": 2320
1923
+ },
1924
+ {
1925
+ "epoch": 11.83,
1926
+ "learning_rate": 4.230118443316413e-05,
1927
+ "loss": 0.0205,
1928
+ "step": 2330
1929
+ },
1930
+ {
1931
+ "epoch": 11.88,
1932
+ "learning_rate": 4.162436548223351e-05,
1933
+ "loss": 0.0053,
1934
+ "step": 2340
1935
+ },
1936
+ {
1937
+ "epoch": 11.93,
1938
+ "learning_rate": 4.094754653130288e-05,
1939
+ "loss": 0.0153,
1940
+ "step": 2350
1941
+ },
1942
+ {
1943
+ "epoch": 11.98,
1944
+ "learning_rate": 4.027072758037225e-05,
1945
+ "loss": 0.0196,
1946
+ "step": 2360
1947
+ },
1948
+ {
1949
+ "epoch": 11.98,
1950
+ "eval_accuracy": 0.5032258064516129,
1951
+ "eval_loss": 2.4746322631835938,
1952
+ "eval_runtime": 30.3112,
1953
+ "eval_samples_per_second": 56.25,
1954
+ "eval_steps_per_second": 7.06,
1955
+ "step": 2360
1956
+ },
1957
+ {
1958
+ "epoch": 12.03,
1959
+ "learning_rate": 3.959390862944163e-05,
1960
+ "loss": 0.0114,
1961
+ "step": 2370
1962
+ },
1963
+ {
1964
+ "epoch": 12.08,
1965
+ "learning_rate": 3.8917089678510996e-05,
1966
+ "loss": 0.0144,
1967
+ "step": 2380
1968
+ },
1969
+ {
1970
+ "epoch": 12.13,
1971
+ "learning_rate": 3.8240270727580375e-05,
1972
+ "loss": 0.0193,
1973
+ "step": 2390
1974
+ },
1975
+ {
1976
+ "epoch": 12.18,
1977
+ "learning_rate": 3.756345177664975e-05,
1978
+ "loss": 0.0104,
1979
+ "step": 2400
1980
+ },
1981
+ {
1982
+ "epoch": 12.18,
1983
+ "eval_accuracy": 0.5079178885630499,
1984
+ "eval_loss": 2.4812278747558594,
1985
+ "eval_runtime": 30.7018,
1986
+ "eval_samples_per_second": 55.534,
1987
+ "eval_steps_per_second": 6.97,
1988
+ "step": 2400
1989
+ },
1990
+ {
1991
+ "epoch": 12.23,
1992
+ "learning_rate": 3.688663282571912e-05,
1993
+ "loss": 0.015,
1994
+ "step": 2410
1995
+ },
1996
+ {
1997
+ "epoch": 12.28,
1998
+ "learning_rate": 3.62098138747885e-05,
1999
+ "loss": 0.0115,
2000
+ "step": 2420
2001
+ },
2002
+ {
2003
+ "epoch": 12.34,
2004
+ "learning_rate": 3.553299492385787e-05,
2005
+ "loss": 0.0152,
2006
+ "step": 2430
2007
+ },
2008
+ {
2009
+ "epoch": 12.39,
2010
+ "learning_rate": 3.485617597292725e-05,
2011
+ "loss": 0.0254,
2012
+ "step": 2440
2013
+ },
2014
+ {
2015
+ "epoch": 12.39,
2016
+ "eval_accuracy": 0.506158357771261,
2017
+ "eval_loss": 2.465996265411377,
2018
+ "eval_runtime": 30.4976,
2019
+ "eval_samples_per_second": 55.906,
2020
+ "eval_steps_per_second": 7.017,
2021
+ "step": 2440
2022
+ },
2023
+ {
2024
+ "epoch": 12.44,
2025
+ "learning_rate": 3.4179357021996615e-05,
2026
+ "loss": 0.005,
2027
+ "step": 2450
2028
+ },
2029
+ {
2030
+ "epoch": 12.49,
2031
+ "learning_rate": 3.3502538071065994e-05,
2032
+ "loss": 0.0046,
2033
+ "step": 2460
2034
+ },
2035
+ {
2036
+ "epoch": 12.54,
2037
+ "learning_rate": 3.2825719120135366e-05,
2038
+ "loss": 0.0048,
2039
+ "step": 2470
2040
+ },
2041
+ {
2042
+ "epoch": 12.59,
2043
+ "learning_rate": 3.214890016920474e-05,
2044
+ "loss": 0.0057,
2045
+ "step": 2480
2046
+ },
2047
+ {
2048
+ "epoch": 12.59,
2049
+ "eval_accuracy": 0.5096774193548387,
2050
+ "eval_loss": 2.4570250511169434,
2051
+ "eval_runtime": 30.647,
2052
+ "eval_samples_per_second": 55.634,
2053
+ "eval_steps_per_second": 6.983,
2054
+ "step": 2480
2055
+ },
2056
+ {
2057
+ "epoch": 12.64,
2058
+ "learning_rate": 3.147208121827411e-05,
2059
+ "loss": 0.0159,
2060
+ "step": 2490
2061
+ },
2062
+ {
2063
+ "epoch": 12.69,
2064
+ "learning_rate": 3.079526226734349e-05,
2065
+ "loss": 0.0156,
2066
+ "step": 2500
2067
+ },
2068
+ {
2069
+ "epoch": 12.74,
2070
+ "learning_rate": 3.0118443316412858e-05,
2071
+ "loss": 0.0089,
2072
+ "step": 2510
2073
+ },
2074
+ {
2075
+ "epoch": 12.79,
2076
+ "learning_rate": 2.9441624365482233e-05,
2077
+ "loss": 0.0161,
2078
+ "step": 2520
2079
+ },
2080
+ {
2081
+ "epoch": 12.79,
2082
+ "eval_accuracy": 0.509090909090909,
2083
+ "eval_loss": 2.452195644378662,
2084
+ "eval_runtime": 30.5139,
2085
+ "eval_samples_per_second": 55.876,
2086
+ "eval_steps_per_second": 7.013,
2087
+ "step": 2520
2088
+ },
2089
+ {
2090
+ "epoch": 12.84,
2091
+ "learning_rate": 2.876480541455161e-05,
2092
+ "loss": 0.0046,
2093
+ "step": 2530
2094
+ },
2095
+ {
2096
+ "epoch": 12.89,
2097
+ "learning_rate": 2.808798646362098e-05,
2098
+ "loss": 0.0181,
2099
+ "step": 2540
2100
+ },
2101
+ {
2102
+ "epoch": 12.94,
2103
+ "learning_rate": 2.7411167512690357e-05,
2104
+ "loss": 0.0131,
2105
+ "step": 2550
2106
+ },
2107
+ {
2108
+ "epoch": 12.99,
2109
+ "learning_rate": 2.6734348561759732e-05,
2110
+ "loss": 0.0237,
2111
+ "step": 2560
2112
+ },
2113
+ {
2114
+ "epoch": 12.99,
2115
+ "eval_accuracy": 0.506158357771261,
2116
+ "eval_loss": 2.461273670196533,
2117
+ "eval_runtime": 30.9259,
2118
+ "eval_samples_per_second": 55.132,
2119
+ "eval_steps_per_second": 6.92,
2120
+ "step": 2560
2121
+ },
2122
+ {
2123
+ "epoch": 13.05,
2124
+ "learning_rate": 2.6057529610829108e-05,
2125
+ "loss": 0.0098,
2126
+ "step": 2570
2127
+ },
2128
+ {
2129
+ "epoch": 13.1,
2130
+ "learning_rate": 2.5380710659898476e-05,
2131
+ "loss": 0.0135,
2132
+ "step": 2580
2133
+ },
2134
+ {
2135
+ "epoch": 13.15,
2136
+ "learning_rate": 2.4703891708967852e-05,
2137
+ "loss": 0.009,
2138
+ "step": 2590
2139
+ },
2140
+ {
2141
+ "epoch": 13.2,
2142
+ "learning_rate": 2.4027072758037224e-05,
2143
+ "loss": 0.0048,
2144
+ "step": 2600
2145
+ },
2146
+ {
2147
+ "epoch": 13.2,
2148
+ "eval_accuracy": 0.5055718475073314,
2149
+ "eval_loss": 2.4772698879241943,
2150
+ "eval_runtime": 30.4386,
2151
+ "eval_samples_per_second": 56.014,
2152
+ "eval_steps_per_second": 7.031,
2153
+ "step": 2600
2154
+ },
2155
+ {
2156
+ "epoch": 13.25,
2157
+ "learning_rate": 2.33502538071066e-05,
2158
+ "loss": 0.0138,
2159
+ "step": 2610
2160
+ },
2161
+ {
2162
+ "epoch": 13.3,
2163
+ "learning_rate": 2.2673434856175975e-05,
2164
+ "loss": 0.009,
2165
+ "step": 2620
2166
+ },
2167
+ {
2168
+ "epoch": 13.35,
2169
+ "learning_rate": 2.199661590524535e-05,
2170
+ "loss": 0.0044,
2171
+ "step": 2630
2172
+ },
2173
+ {
2174
+ "epoch": 13.4,
2175
+ "learning_rate": 2.1319796954314723e-05,
2176
+ "loss": 0.0067,
2177
+ "step": 2640
2178
+ },
2179
+ {
2180
+ "epoch": 13.4,
2181
+ "eval_accuracy": 0.5079178885630499,
2182
+ "eval_loss": 2.474796772003174,
2183
+ "eval_runtime": 30.9994,
2184
+ "eval_samples_per_second": 55.001,
2185
+ "eval_steps_per_second": 6.903,
2186
+ "step": 2640
2187
+ },
2188
+ {
2189
+ "epoch": 13.45,
2190
+ "learning_rate": 2.0642978003384095e-05,
2191
+ "loss": 0.0048,
2192
+ "step": 2650
2193
+ },
2194
+ {
2195
+ "epoch": 13.5,
2196
+ "learning_rate": 1.996615905245347e-05,
2197
+ "loss": 0.0201,
2198
+ "step": 2660
2199
+ },
2200
+ {
2201
+ "epoch": 13.55,
2202
+ "learning_rate": 1.9289340101522843e-05,
2203
+ "loss": 0.0144,
2204
+ "step": 2670
2205
+ },
2206
+ {
2207
+ "epoch": 13.6,
2208
+ "learning_rate": 1.8612521150592218e-05,
2209
+ "loss": 0.0105,
2210
+ "step": 2680
2211
+ },
2212
+ {
2213
+ "epoch": 13.6,
2214
+ "eval_accuracy": 0.5085043988269795,
2215
+ "eval_loss": 2.4759089946746826,
2216
+ "eval_runtime": 30.6746,
2217
+ "eval_samples_per_second": 55.584,
2218
+ "eval_steps_per_second": 6.976,
2219
+ "step": 2680
2220
+ },
2221
+ {
2222
+ "epoch": 13.65,
2223
+ "learning_rate": 1.793570219966159e-05,
2224
+ "loss": 0.0174,
2225
+ "step": 2690
2226
+ },
2227
+ {
2228
+ "epoch": 13.71,
2229
+ "learning_rate": 1.7258883248730966e-05,
2230
+ "loss": 0.0045,
2231
+ "step": 2700
2232
+ },
2233
+ {
2234
+ "epoch": 13.76,
2235
+ "learning_rate": 1.658206429780034e-05,
2236
+ "loss": 0.0044,
2237
+ "step": 2710
2238
+ },
2239
+ {
2240
+ "epoch": 13.81,
2241
+ "learning_rate": 1.5905245346869714e-05,
2242
+ "loss": 0.0209,
2243
+ "step": 2720
2244
+ },
2245
+ {
2246
+ "epoch": 13.81,
2247
+ "eval_accuracy": 0.506158357771261,
2248
+ "eval_loss": 2.4831149578094482,
2249
+ "eval_runtime": 30.8346,
2250
+ "eval_samples_per_second": 55.295,
2251
+ "eval_steps_per_second": 6.94,
2252
+ "step": 2720
2253
+ },
2254
+ {
2255
+ "epoch": 13.86,
2256
+ "learning_rate": 1.5228426395939088e-05,
2257
+ "loss": 0.0122,
2258
+ "step": 2730
2259
+ },
2260
+ {
2261
+ "epoch": 13.91,
2262
+ "learning_rate": 1.4551607445008461e-05,
2263
+ "loss": 0.0149,
2264
+ "step": 2740
2265
+ },
2266
+ {
2267
+ "epoch": 13.96,
2268
+ "learning_rate": 1.3874788494077834e-05,
2269
+ "loss": 0.0225,
2270
+ "step": 2750
2271
+ },
2272
+ {
2273
+ "epoch": 14.01,
2274
+ "learning_rate": 1.3197969543147209e-05,
2275
+ "loss": 0.0112,
2276
+ "step": 2760
2277
+ },
2278
+ {
2279
+ "epoch": 14.01,
2280
+ "eval_accuracy": 0.5085043988269795,
2281
+ "eval_loss": 2.482898473739624,
2282
+ "eval_runtime": 30.9465,
2283
+ "eval_samples_per_second": 55.095,
2284
+ "eval_steps_per_second": 6.915,
2285
+ "step": 2760
2286
+ },
2287
+ {
2288
+ "epoch": 14.06,
2289
+ "learning_rate": 1.2521150592216583e-05,
2290
+ "loss": 0.0084,
2291
+ "step": 2770
2292
+ },
2293
+ {
2294
+ "epoch": 14.11,
2295
+ "learning_rate": 1.1844331641285957e-05,
2296
+ "loss": 0.0127,
2297
+ "step": 2780
2298
+ },
2299
+ {
2300
+ "epoch": 14.16,
2301
+ "learning_rate": 1.116751269035533e-05,
2302
+ "loss": 0.0097,
2303
+ "step": 2790
2304
+ },
2305
+ {
2306
+ "epoch": 14.21,
2307
+ "learning_rate": 1.0490693739424704e-05,
2308
+ "loss": 0.0084,
2309
+ "step": 2800
2310
+ },
2311
+ {
2312
+ "epoch": 14.21,
2313
+ "eval_accuracy": 0.509090909090909,
2314
+ "eval_loss": 2.4852123260498047,
2315
+ "eval_runtime": 30.864,
2316
+ "eval_samples_per_second": 55.242,
2317
+ "eval_steps_per_second": 6.934,
2318
+ "step": 2800
2319
+ },
2320
+ {
2321
+ "epoch": 14.26,
2322
+ "learning_rate": 9.813874788494078e-06,
2323
+ "loss": 0.0078,
2324
+ "step": 2810
2325
+ },
2326
+ {
2327
+ "epoch": 14.31,
2328
+ "learning_rate": 9.137055837563452e-06,
2329
+ "loss": 0.0173,
2330
+ "step": 2820
2331
+ },
2332
+ {
2333
+ "epoch": 14.37,
2334
+ "learning_rate": 8.460236886632826e-06,
2335
+ "loss": 0.0145,
2336
+ "step": 2830
2337
+ },
2338
+ {
2339
+ "epoch": 14.42,
2340
+ "learning_rate": 7.7834179357022e-06,
2341
+ "loss": 0.0203,
2342
+ "step": 2840
2343
+ },
2344
+ {
2345
+ "epoch": 14.42,
2346
+ "eval_accuracy": 0.5096774193548387,
2347
+ "eval_loss": 2.4804649353027344,
2348
+ "eval_runtime": 30.8958,
2349
+ "eval_samples_per_second": 55.185,
2350
+ "eval_steps_per_second": 6.927,
2351
+ "step": 2840
2352
+ },
2353
+ {
2354
+ "epoch": 14.47,
2355
+ "learning_rate": 7.106598984771575e-06,
2356
+ "loss": 0.0143,
2357
+ "step": 2850
2358
+ },
2359
+ {
2360
+ "epoch": 14.52,
2361
+ "learning_rate": 6.429780033840948e-06,
2362
+ "loss": 0.0094,
2363
+ "step": 2860
2364
+ },
2365
+ {
2366
+ "epoch": 14.57,
2367
+ "learning_rate": 5.7529610829103214e-06,
2368
+ "loss": 0.0105,
2369
+ "step": 2870
2370
+ },
2371
+ {
2372
+ "epoch": 14.62,
2373
+ "learning_rate": 5.076142131979695e-06,
2374
+ "loss": 0.0086,
2375
+ "step": 2880
2376
+ },
2377
+ {
2378
+ "epoch": 14.62,
2379
+ "eval_accuracy": 0.5085043988269795,
2380
+ "eval_loss": 2.480623960494995,
2381
+ "eval_runtime": 30.8962,
2382
+ "eval_samples_per_second": 55.185,
2383
+ "eval_steps_per_second": 6.926,
2384
+ "step": 2880
2385
+ },
2386
+ {
2387
+ "epoch": 14.67,
2388
+ "learning_rate": 4.39932318104907e-06,
2389
+ "loss": 0.01,
2390
+ "step": 2890
2391
+ },
2392
+ {
2393
+ "epoch": 14.72,
2394
+ "learning_rate": 3.7225042301184434e-06,
2395
+ "loss": 0.0048,
2396
+ "step": 2900
2397
+ },
2398
+ {
2399
+ "epoch": 14.77,
2400
+ "learning_rate": 3.0456852791878177e-06,
2401
+ "loss": 0.0087,
2402
+ "step": 2910
2403
+ },
2404
+ {
2405
+ "epoch": 14.82,
2406
+ "learning_rate": 2.3688663282571915e-06,
2407
+ "loss": 0.0103,
2408
+ "step": 2920
2409
+ },
2410
+ {
2411
+ "epoch": 14.82,
2412
+ "eval_accuracy": 0.5085043988269795,
2413
+ "eval_loss": 2.4825499057769775,
2414
+ "eval_runtime": 30.9894,
2415
+ "eval_samples_per_second": 55.019,
2416
+ "eval_steps_per_second": 6.906,
2417
+ "step": 2920
2418
+ },
2419
+ {
2420
+ "epoch": 14.87,
2421
+ "learning_rate": 1.6920473773265652e-06,
2422
+ "loss": 0.0095,
2423
+ "step": 2930
2424
+ },
2425
+ {
2426
+ "epoch": 14.92,
2427
+ "learning_rate": 1.015228426395939e-06,
2428
+ "loss": 0.0046,
2429
+ "step": 2940
2430
+ },
2431
+ {
2432
+ "epoch": 14.97,
2433
+ "learning_rate": 3.38409475465313e-07,
2434
+ "loss": 0.0113,
2435
+ "step": 2950
2436
+ },
2437
+ {
2438
+ "epoch": 15.0,
2439
+ "step": 2955,
2440
+ "total_flos": 3.64901908299393e+18,
2441
+ "train_loss": 0.5233747142333972,
2442
+ "train_runtime": 4896.0515,
2443
+ "train_samples_per_second": 9.614,
2444
+ "train_steps_per_second": 0.604
2445
+ }
2446
+ ],
2447
+ "max_steps": 2955,
2448
+ "num_train_epochs": 15,
2449
+ "total_flos": 3.64901908299393e+18,
2450
+ "trial_name": null,
2451
+ "trial_params": null
2452
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:81bc596a5fb737e64cac2afcc4b39d05df7c2c9ae8aa4e8c80cb21301b1d396e
3
+ size 3963