Hugging Face's logo

Augusto777
/
Swin-DA2-final-AMD-Wet

TensorBoard
Safetensors
swinv2
Generated from Trainer
Eval Results
Model card Files Metrics Community
Augusto777 commited on
Commit
54c3d94
·
verified ·
1 Parent(s): b759391

End of training

Browse files
Files changed (11) hide show
  1. README.md +153 -0
  2. all_results.json +12 -0
  3. config.json +60 -0
  4. eval_results.json +8 -0
  5. model.safetensors +3 -0
  6. preprocessor_config.json +22 -0
  7. runs/Feb04_08-26-01_DESKTOP-SKBE9FB/events.out.tfevents.1738679162.DESKTOP-SKBE9FB.1684.0 +3 -0
  8. runs/Feb04_08-26-01_DESKTOP-SKBE9FB/events.out.tfevents.1738680059.DESKTOP-SKBE9FB.1684.1 +3 -0
  9. train_results.json +7 -0
  10. trainer_state.json +1338 -0
  11. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,153 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: microsoft/swinv2-tiny-patch4-window8-256
4
+ tags:
5
+ - generated_from_trainer
6
+ datasets:
7
+ - imagefolder
8
+ metrics:
9
+ - accuracy
10
+ model-index:
11
+ - name: Swin-DA2-final-AMD-Wet
12
+ results:
13
+ - task:
14
+ name: Image Classification
15
+ type: image-classification
16
+ dataset:
17
+ name: imagefolder
18
+ type: imagefolder
19
+ config: default
20
+ split: validation
21
+ args: default
22
+ metrics:
23
+ - name: Accuracy
24
+ type: accuracy
25
+ value: 0.788235294117647
26
+ ---
27
+
28
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
29
+ should probably proofread and complete it, then remove this comment. -->
30
+
31
+ # Swin-DA2-final-AMD-Wet
32
+
33
+ This model is a fine-tuned version of [microsoft/swinv2-tiny-patch4-window8-256](https://huggingface.co/microsoft/swinv2-tiny-patch4-window8-256) on the imagefolder dataset.
34
+ It achieves the following results on the evaluation set:
35
+ - Loss: 0.6258
36
+ - Accuracy: 0.7882
37
+
38
+ ## Model description
39
+
40
+ More information needed
41
+
42
+ ## Intended uses & limitations
43
+
44
+ More information needed
45
+
46
+ ## Training and evaluation data
47
+
48
+ More information needed
49
+
50
+ ## Training procedure
51
+
52
+ ### Training hyperparameters
53
+
54
+ The following hyperparameters were used during training:
55
+ - learning_rate: 5e-05
56
+ - train_batch_size: 16
57
+ - eval_batch_size: 16
58
+ - seed: 42
59
+ - gradient_accumulation_steps: 4
60
+ - total_train_batch_size: 64
61
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
62
+ - lr_scheduler_type: linear
63
+ - lr_scheduler_warmup_ratio: 0.1
64
+ - num_epochs: 80
65
+
66
+ ### Training results
67
+
68
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy |
69
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|
70
+ | 1.6093 | 0.95 | 13 | 1.6082 | 0.2118 |
71
+ | 1.6076 | 1.96 | 27 | 1.5981 | 0.1765 |
72
+ | 1.5847 | 2.98 | 41 | 1.5665 | 0.3059 |
73
+ | 1.5563 | 4.0 | 55 | 1.5250 | 0.2588 |
74
+ | 1.5127 | 4.95 | 68 | 1.3810 | 0.4353 |
75
+ | 1.2694 | 5.96 | 82 | 1.1126 | 0.5647 |
76
+ | 1.207 | 6.98 | 96 | 0.8636 | 0.7647 |
77
+ | 1.0502 | 8.0 | 110 | 0.9073 | 0.6471 |
78
+ | 0.9587 | 8.95 | 123 | 0.8657 | 0.6471 |
79
+ | 0.9694 | 9.96 | 137 | 0.7357 | 0.7412 |
80
+ | 0.8198 | 10.98 | 151 | 0.6258 | 0.7882 |
81
+ | 0.7846 | 12.0 | 165 | 0.6916 | 0.7294 |
82
+ | 0.7633 | 12.95 | 178 | 0.6798 | 0.6588 |
83
+ | 0.6097 | 13.96 | 192 | 0.6048 | 0.7294 |
84
+ | 0.5937 | 14.98 | 206 | 0.6778 | 0.7529 |
85
+ | 0.5789 | 16.0 | 220 | 0.8130 | 0.6941 |
86
+ | 0.5041 | 16.95 | 233 | 0.7081 | 0.7176 |
87
+ | 0.5102 | 17.96 | 247 | 0.7835 | 0.7176 |
88
+ | 0.4183 | 18.98 | 261 | 0.7358 | 0.7176 |
89
+ | 0.4632 | 20.0 | 275 | 0.7227 | 0.7176 |
90
+ | 0.3722 | 20.95 | 288 | 0.8009 | 0.7059 |
91
+ | 0.3962 | 21.96 | 302 | 0.7978 | 0.7529 |
92
+ | 0.3813 | 22.98 | 316 | 0.8392 | 0.7529 |
93
+ | 0.3254 | 24.0 | 330 | 0.9322 | 0.6941 |
94
+ | 0.3575 | 24.95 | 343 | 0.8846 | 0.7294 |
95
+ | 0.3335 | 25.96 | 357 | 0.9651 | 0.7294 |
96
+ | 0.3055 | 26.98 | 371 | 0.8926 | 0.7412 |
97
+ | 0.2969 | 28.0 | 385 | 1.0529 | 0.6824 |
98
+ | 0.3243 | 28.95 | 398 | 1.0659 | 0.7294 |
99
+ | 0.3012 | 29.96 | 412 | 0.9155 | 0.7882 |
100
+ | 0.2582 | 30.98 | 426 | 1.0367 | 0.7176 |
101
+ | 0.2881 | 32.0 | 440 | 0.9626 | 0.7176 |
102
+ | 0.2538 | 32.95 | 453 | 0.9636 | 0.7294 |
103
+ | 0.2368 | 33.96 | 467 | 1.0624 | 0.7059 |
104
+ | 0.299 | 34.98 | 481 | 0.9683 | 0.7412 |
105
+ | 0.2553 | 36.0 | 495 | 0.9977 | 0.7176 |
106
+ | 0.2325 | 36.95 | 508 | 0.9951 | 0.7176 |
107
+ | 0.2363 | 37.96 | 522 | 1.1467 | 0.7176 |
108
+ | 0.2424 | 38.98 | 536 | 1.1336 | 0.6941 |
109
+ | 0.2096 | 40.0 | 550 | 1.0932 | 0.7294 |
110
+ | 0.1978 | 40.95 | 563 | 1.1394 | 0.6941 |
111
+ | 0.1746 | 41.96 | 577 | 1.0977 | 0.7529 |
112
+ | 0.2248 | 42.98 | 591 | 1.1301 | 0.7529 |
113
+ | 0.2484 | 44.0 | 605 | 1.2937 | 0.6941 |
114
+ | 0.2012 | 44.95 | 618 | 1.2553 | 0.7176 |
115
+ | 0.2347 | 45.96 | 632 | 1.1070 | 0.7412 |
116
+ | 0.2061 | 46.98 | 646 | 1.2093 | 0.7059 |
117
+ | 0.2057 | 48.0 | 660 | 1.1720 | 0.7294 |
118
+ | 0.1808 | 48.95 | 673 | 1.3337 | 0.6941 |
119
+ | 0.1872 | 49.96 | 687 | 1.3495 | 0.6824 |
120
+ | 0.1584 | 50.98 | 701 | 1.3836 | 0.6941 |
121
+ | 0.1237 | 52.0 | 715 | 1.3276 | 0.7176 |
122
+ | 0.2186 | 52.95 | 728 | 1.3548 | 0.6824 |
123
+ | 0.2053 | 53.96 | 742 | 1.3200 | 0.6941 |
124
+ | 0.1618 | 54.98 | 756 | 1.2399 | 0.7059 |
125
+ | 0.1568 | 56.0 | 770 | 1.3510 | 0.7059 |
126
+ | 0.175 | 56.95 | 783 | 1.3852 | 0.7059 |
127
+ | 0.15 | 57.96 | 797 | 1.3693 | 0.7176 |
128
+ | 0.2088 | 58.98 | 811 | 1.4166 | 0.6824 |
129
+ | 0.1635 | 60.0 | 825 | 1.2657 | 0.7059 |
130
+ | 0.1681 | 60.95 | 838 | 1.3034 | 0.7176 |
131
+ | 0.166 | 61.96 | 852 | 1.2800 | 0.7294 |
132
+ | 0.1585 | 62.98 | 866 | 1.3041 | 0.7294 |
133
+ | 0.152 | 64.0 | 880 | 1.4154 | 0.6824 |
134
+ | 0.155 | 64.95 | 893 | 1.3989 | 0.6824 |
135
+ | 0.1547 | 65.96 | 907 | 1.3432 | 0.7176 |
136
+ | 0.1113 | 66.98 | 921 | 1.3281 | 0.7294 |
137
+ | 0.159 | 68.0 | 935 | 1.3717 | 0.7294 |
138
+ | 0.1149 | 68.95 | 948 | 1.3776 | 0.7294 |
139
+ | 0.1478 | 69.96 | 962 | 1.3459 | 0.7294 |
140
+ | 0.1675 | 70.98 | 976 | 1.3941 | 0.7294 |
141
+ | 0.1476 | 72.0 | 990 | 1.4110 | 0.7059 |
142
+ | 0.1515 | 72.95 | 1003 | 1.3704 | 0.7059 |
143
+ | 0.1635 | 73.96 | 1017 | 1.3647 | 0.7176 |
144
+ | 0.1445 | 74.98 | 1031 | 1.3673 | 0.7176 |
145
+ | 0.1398 | 75.64 | 1040 | 1.3689 | 0.7176 |
146
+
147
+
148
+ ### Framework versions
149
+
150
+ - Transformers 4.36.2
151
+ - Pytorch 2.1.2+cu118
152
+ - Datasets 2.16.1
153
+ - Tokenizers 0.15.0
all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 75.64,
3
+ "eval_accuracy": 0.788235294117647,
4
+ "eval_loss": 0.62584388256073,
5
+ "eval_runtime": 0.4111,
6
+ "eval_samples_per_second": 206.767,
7
+ "eval_steps_per_second": 14.595,
8
+ "train_loss": 0.41205361909591237,
9
+ "train_runtime": 895.6069,
10
+ "train_samples_per_second": 78.338,
11
+ "train_steps_per_second": 1.161
12
+ }
config.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/swinv2-tiny-patch4-window8-256",
3
+ "architectures": [
4
+ "Swinv2ForImageClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "depths": [
8
+ 2,
9
+ 2,
10
+ 6,
11
+ 2
12
+ ],
13
+ "drop_path_rate": 0.1,
14
+ "embed_dim": 96,
15
+ "encoder_stride": 32,
16
+ "hidden_act": "gelu",
17
+ "hidden_dropout_prob": 0.0,
18
+ "hidden_size": 768,
19
+ "id2label": {
20
+ "0": "avanzada",
21
+ "1": "avanzada humeda",
22
+ "2": "leve",
23
+ "3": "moderada",
24
+ "4": "no dmae"
25
+ },
26
+ "image_size": 256,
27
+ "initializer_range": 0.02,
28
+ "label2id": {
29
+ "avanzada": 0,
30
+ "avanzada humeda": 1,
31
+ "leve": 2,
32
+ "moderada": 3,
33
+ "no dmae": 4
34
+ },
35
+ "layer_norm_eps": 1e-05,
36
+ "mlp_ratio": 4.0,
37
+ "model_type": "swinv2",
38
+ "num_channels": 3,
39
+ "num_heads": [
40
+ 3,
41
+ 6,
42
+ 12,
43
+ 24
44
+ ],
45
+ "num_layers": 4,
46
+ "patch_size": 4,
47
+ "path_norm": true,
48
+ "pretrained_window_sizes": [
49
+ 0,
50
+ 0,
51
+ 0,
52
+ 0
53
+ ],
54
+ "problem_type": "single_label_classification",
55
+ "qkv_bias": true,
56
+ "torch_dtype": "float32",
57
+ "transformers_version": "4.36.2",
58
+ "use_absolute_embeddings": false,
59
+ "window_size": 8
60
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 75.64,
3
+ "eval_accuracy": 0.788235294117647,
4
+ "eval_loss": 0.62584388256073,
5
+ "eval_runtime": 0.4111,
6
+ "eval_samples_per_second": 206.767,
7
+ "eval_steps_per_second": 14.595
8
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:80a3fa42743f4d37da768220e98dbfb5bfea11894170f72e61ca1190019d35c1
3
+ size 110359372
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
+ 0.485,
7
+ 0.456,
8
+ 0.406
9
+ ],
10
+ "image_processor_type": "ViTImageProcessor",
11
+ "image_std": [
12
+ 0.229,
13
+ 0.224,
14
+ 0.225
15
+ ],
16
+ "resample": 3,
17
+ "rescale_factor": 0.00392156862745098,
18
+ "size": {
19
+ "height": 256,
20
+ "width": 256
21
+ }
22
+ }
runs/Feb04_08-26-01_DESKTOP-SKBE9FB/events.out.tfevents.1738679162.DESKTOP-SKBE9FB.1684.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e24ab7fea7132744eb4e1ae2e777591c7a906dbc70f6371f002b7e47ff9be3d0
3
+ size 45974
runs/Feb04_08-26-01_DESKTOP-SKBE9FB/events.out.tfevents.1738680059.DESKTOP-SKBE9FB.1684.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f46936e26eaf6930e9b3c7c5282051a90bc7c726dfc90131b3853f542e9e9c1c
3
+ size 411
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 75.64,
3
+ "train_loss": 0.41205361909591237,
4
+ "train_runtime": 895.6069,
5
+ "train_samples_per_second": 78.338,
6
+ "train_steps_per_second": 1.161
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,1338 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.788235294117647,
3
+ "best_model_checkpoint": "/Modelos/Swin-DA2-final-AMD-Wet\\checkpoint-151",
4
+ "epoch": 75.63636363636364,
5
+ "eval_steps": 500,
6
+ "global_step": 1040,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.73,
13
+ "learning_rate": 4.807692307692308e-06,
14
+ "loss": 1.6093,
15
+ "step": 10
16
+ },
17
+ {
18
+ "epoch": 0.95,
19
+ "eval_accuracy": 0.21176470588235294,
20
+ "eval_loss": 1.6081717014312744,
21
+ "eval_runtime": 0.8052,
22
+ "eval_samples_per_second": 105.566,
23
+ "eval_steps_per_second": 7.452,
24
+ "step": 13
25
+ },
26
+ {
27
+ "epoch": 1.45,
28
+ "learning_rate": 9.615384615384616e-06,
29
+ "loss": 1.6076,
30
+ "step": 20
31
+ },
32
+ {
33
+ "epoch": 1.96,
34
+ "eval_accuracy": 0.17647058823529413,
35
+ "eval_loss": 1.598115086555481,
36
+ "eval_runtime": 0.4141,
37
+ "eval_samples_per_second": 205.268,
38
+ "eval_steps_per_second": 14.489,
39
+ "step": 27
40
+ },
41
+ {
42
+ "epoch": 2.18,
43
+ "learning_rate": 1.4423076923076923e-05,
44
+ "loss": 1.602,
45
+ "step": 30
46
+ },
47
+ {
48
+ "epoch": 2.91,
49
+ "learning_rate": 1.923076923076923e-05,
50
+ "loss": 1.5847,
51
+ "step": 40
52
+ },
53
+ {
54
+ "epoch": 2.98,
55
+ "eval_accuracy": 0.3058823529411765,
56
+ "eval_loss": 1.566546082496643,
57
+ "eval_runtime": 0.4121,
58
+ "eval_samples_per_second": 206.264,
59
+ "eval_steps_per_second": 14.56,
60
+ "step": 41
61
+ },
62
+ {
63
+ "epoch": 3.64,
64
+ "learning_rate": 2.4038461538461542e-05,
65
+ "loss": 1.5563,
66
+ "step": 50
67
+ },
68
+ {
69
+ "epoch": 4.0,
70
+ "eval_accuracy": 0.25882352941176473,
71
+ "eval_loss": 1.525040626525879,
72
+ "eval_runtime": 0.4171,
73
+ "eval_samples_per_second": 203.791,
74
+ "eval_steps_per_second": 14.385,
75
+ "step": 55
76
+ },
77
+ {
78
+ "epoch": 4.36,
79
+ "learning_rate": 2.8846153846153845e-05,
80
+ "loss": 1.5127,
81
+ "step": 60
82
+ },
83
+ {
84
+ "epoch": 4.95,
85
+ "eval_accuracy": 0.43529411764705883,
86
+ "eval_loss": 1.3809895515441895,
87
+ "eval_runtime": 0.4181,
88
+ "eval_samples_per_second": 203.303,
89
+ "eval_steps_per_second": 14.351,
90
+ "step": 68
91
+ },
92
+ {
93
+ "epoch": 5.09,
94
+ "learning_rate": 3.365384615384616e-05,
95
+ "loss": 1.4057,
96
+ "step": 70
97
+ },
98
+ {
99
+ "epoch": 5.82,
100
+ "learning_rate": 3.846153846153846e-05,
101
+ "loss": 1.2694,
102
+ "step": 80
103
+ },
104
+ {
105
+ "epoch": 5.96,
106
+ "eval_accuracy": 0.5647058823529412,
107
+ "eval_loss": 1.1125926971435547,
108
+ "eval_runtime": 0.4161,
109
+ "eval_samples_per_second": 204.281,
110
+ "eval_steps_per_second": 14.42,
111
+ "step": 82
112
+ },
113
+ {
114
+ "epoch": 6.55,
115
+ "learning_rate": 4.326923076923077e-05,
116
+ "loss": 1.207,
117
+ "step": 90
118
+ },
119
+ {
120
+ "epoch": 6.98,
121
+ "eval_accuracy": 0.7647058823529411,
122
+ "eval_loss": 0.8636198043823242,
123
+ "eval_runtime": 0.4131,
124
+ "eval_samples_per_second": 205.765,
125
+ "eval_steps_per_second": 14.525,
126
+ "step": 96
127
+ },
128
+ {
129
+ "epoch": 7.27,
130
+ "learning_rate": 4.8076923076923084e-05,
131
+ "loss": 1.0562,
132
+ "step": 100
133
+ },
134
+ {
135
+ "epoch": 8.0,
136
+ "learning_rate": 4.9679487179487185e-05,
137
+ "loss": 1.0502,
138
+ "step": 110
139
+ },
140
+ {
141
+ "epoch": 8.0,
142
+ "eval_accuracy": 0.6470588235294118,
143
+ "eval_loss": 0.9073464870452881,
144
+ "eval_runtime": 0.4161,
145
+ "eval_samples_per_second": 204.281,
146
+ "eval_steps_per_second": 14.42,
147
+ "step": 110
148
+ },
149
+ {
150
+ "epoch": 8.73,
151
+ "learning_rate": 4.9145299145299147e-05,
152
+ "loss": 0.9587,
153
+ "step": 120
154
+ },
155
+ {
156
+ "epoch": 8.95,
157
+ "eval_accuracy": 0.6470588235294118,
158
+ "eval_loss": 0.8657189011573792,
159
+ "eval_runtime": 0.4311,
160
+ "eval_samples_per_second": 197.171,
161
+ "eval_steps_per_second": 13.918,
162
+ "step": 123
163
+ },
164
+ {
165
+ "epoch": 9.45,
166
+ "learning_rate": 4.8611111111111115e-05,
167
+ "loss": 0.9694,
168
+ "step": 130
169
+ },
170
+ {
171
+ "epoch": 9.96,
172
+ "eval_accuracy": 0.7411764705882353,
173
+ "eval_loss": 0.7357165217399597,
174
+ "eval_runtime": 0.4171,
175
+ "eval_samples_per_second": 203.791,
176
+ "eval_steps_per_second": 14.385,
177
+ "step": 137
178
+ },
179
+ {
180
+ "epoch": 10.18,
181
+ "learning_rate": 4.8076923076923084e-05,
182
+ "loss": 0.8916,
183
+ "step": 140
184
+ },
185
+ {
186
+ "epoch": 10.91,
187
+ "learning_rate": 4.7542735042735045e-05,
188
+ "loss": 0.8198,
189
+ "step": 150
190
+ },
191
+ {
192
+ "epoch": 10.98,
193
+ "eval_accuracy": 0.788235294117647,
194
+ "eval_loss": 0.62584388256073,
195
+ "eval_runtime": 0.4151,
196
+ "eval_samples_per_second": 204.773,
197
+ "eval_steps_per_second": 14.455,
198
+ "step": 151
199
+ },
200
+ {
201
+ "epoch": 11.64,
202
+ "learning_rate": 4.700854700854701e-05,
203
+ "loss": 0.7846,
204
+ "step": 160
205
+ },
206
+ {
207
+ "epoch": 12.0,
208
+ "eval_accuracy": 0.7294117647058823,
209
+ "eval_loss": 0.6916422247886658,
210
+ "eval_runtime": 0.4591,
211
+ "eval_samples_per_second": 185.143,
212
+ "eval_steps_per_second": 13.069,
213
+ "step": 165
214
+ },
215
+ {
216
+ "epoch": 12.36,
217
+ "learning_rate": 4.6474358974358976e-05,
218
+ "loss": 0.7633,
219
+ "step": 170
220
+ },
221
+ {
222
+ "epoch": 12.95,
223
+ "eval_accuracy": 0.6588235294117647,
224
+ "eval_loss": 0.6798376441001892,
225
+ "eval_runtime": 0.4411,
226
+ "eval_samples_per_second": 192.7,
227
+ "eval_steps_per_second": 13.602,
228
+ "step": 178
229
+ },
230
+ {
231
+ "epoch": 13.09,
232
+ "learning_rate": 4.594017094017094e-05,
233
+ "loss": 0.6697,
234
+ "step": 180
235
+ },
236
+ {
237
+ "epoch": 13.82,
238
+ "learning_rate": 4.5405982905982906e-05,
239
+ "loss": 0.6097,
240
+ "step": 190
241
+ },
242
+ {
243
+ "epoch": 13.96,
244
+ "eval_accuracy": 0.7294117647058823,
245
+ "eval_loss": 0.6048377752304077,
246
+ "eval_runtime": 0.4151,
247
+ "eval_samples_per_second": 204.773,
248
+ "eval_steps_per_second": 14.455,
249
+ "step": 192
250
+ },
251
+ {
252
+ "epoch": 14.55,
253
+ "learning_rate": 4.4871794871794874e-05,
254
+ "loss": 0.5937,
255
+ "step": 200
256
+ },
257
+ {
258
+ "epoch": 14.98,
259
+ "eval_accuracy": 0.7529411764705882,
260
+ "eval_loss": 0.6778049468994141,
261
+ "eval_runtime": 0.4131,
262
+ "eval_samples_per_second": 205.765,
263
+ "eval_steps_per_second": 14.525,
264
+ "step": 206
265
+ },
266
+ {
267
+ "epoch": 15.27,
268
+ "learning_rate": 4.4337606837606836e-05,
269
+ "loss": 0.5214,
270
+ "step": 210
271
+ },
272
+ {
273
+ "epoch": 16.0,
274
+ "learning_rate": 4.3803418803418805e-05,
275
+ "loss": 0.5789,
276
+ "step": 220
277
+ },
278
+ {
279
+ "epoch": 16.0,
280
+ "eval_accuracy": 0.6941176470588235,
281
+ "eval_loss": 0.8130398988723755,
282
+ "eval_runtime": 0.4201,
283
+ "eval_samples_per_second": 202.336,
284
+ "eval_steps_per_second": 14.283,
285
+ "step": 220
286
+ },
287
+ {
288
+ "epoch": 16.73,
289
+ "learning_rate": 4.326923076923077e-05,
290
+ "loss": 0.5041,
291
+ "step": 230
292
+ },
293
+ {
294
+ "epoch": 16.95,
295
+ "eval_accuracy": 0.7176470588235294,
296
+ "eval_loss": 0.7080594897270203,
297
+ "eval_runtime": 0.4181,
298
+ "eval_samples_per_second": 203.303,
299
+ "eval_steps_per_second": 14.351,
300
+ "step": 233
301
+ },
302
+ {
303
+ "epoch": 17.45,
304
+ "learning_rate": 4.2735042735042735e-05,
305
+ "loss": 0.5102,
306
+ "step": 240
307
+ },
308
+ {
309
+ "epoch": 17.96,
310
+ "eval_accuracy": 0.7176470588235294,
311
+ "eval_loss": 0.7834978103637695,
312
+ "eval_runtime": 0.4211,
313
+ "eval_samples_per_second": 201.855,
314
+ "eval_steps_per_second": 14.249,
315
+ "step": 247
316
+ },
317
+ {
318
+ "epoch": 18.18,
319
+ "learning_rate": 4.2200854700854704e-05,
320
+ "loss": 0.4691,
321
+ "step": 250
322
+ },
323
+ {
324
+ "epoch": 18.91,
325
+ "learning_rate": 4.166666666666667e-05,
326
+ "loss": 0.4183,
327
+ "step": 260
328
+ },
329
+ {
330
+ "epoch": 18.98,
331
+ "eval_accuracy": 0.7176470588235294,
332
+ "eval_loss": 0.7358449697494507,
333
+ "eval_runtime": 0.4141,
334
+ "eval_samples_per_second": 205.268,
335
+ "eval_steps_per_second": 14.489,
336
+ "step": 261
337
+ },
338
+ {
339
+ "epoch": 19.64,
340
+ "learning_rate": 4.1132478632478634e-05,
341
+ "loss": 0.4632,
342
+ "step": 270
343
+ },
344
+ {
345
+ "epoch": 20.0,
346
+ "eval_accuracy": 0.7176470588235294,
347
+ "eval_loss": 0.7226832509040833,
348
+ "eval_runtime": 0.4531,
349
+ "eval_samples_per_second": 187.596,
350
+ "eval_steps_per_second": 13.242,
351
+ "step": 275
352
+ },
353
+ {
354
+ "epoch": 20.36,
355
+ "learning_rate": 4.05982905982906e-05,
356
+ "loss": 0.3722,
357
+ "step": 280
358
+ },
359
+ {
360
+ "epoch": 20.95,
361
+ "eval_accuracy": 0.7058823529411765,
362
+ "eval_loss": 0.8008852005004883,
363
+ "eval_runtime": 0.4181,
364
+ "eval_samples_per_second": 203.303,
365
+ "eval_steps_per_second": 14.351,
366
+ "step": 288
367
+ },
368
+ {
369
+ "epoch": 21.09,
370
+ "learning_rate": 4.006410256410257e-05,
371
+ "loss": 0.4272,
372
+ "step": 290
373
+ },
374
+ {
375
+ "epoch": 21.82,
376
+ "learning_rate": 3.952991452991453e-05,
377
+ "loss": 0.3962,
378
+ "step": 300
379
+ },
380
+ {
381
+ "epoch": 21.96,
382
+ "eval_accuracy": 0.7529411764705882,
383
+ "eval_loss": 0.7977811098098755,
384
+ "eval_runtime": 0.4411,
385
+ "eval_samples_per_second": 192.7,
386
+ "eval_steps_per_second": 13.602,
387
+ "step": 302
388
+ },
389
+ {
390
+ "epoch": 22.55,
391
+ "learning_rate": 3.89957264957265e-05,
392
+ "loss": 0.3813,
393
+ "step": 310
394
+ },
395
+ {
396
+ "epoch": 22.98,
397
+ "eval_accuracy": 0.7529411764705882,
398
+ "eval_loss": 0.8392243981361389,
399
+ "eval_runtime": 0.4171,
400
+ "eval_samples_per_second": 203.791,
401
+ "eval_steps_per_second": 14.385,
402
+ "step": 316
403
+ },
404
+ {
405
+ "epoch": 23.27,
406
+ "learning_rate": 3.846153846153846e-05,
407
+ "loss": 0.3852,
408
+ "step": 320
409
+ },
410
+ {
411
+ "epoch": 24.0,
412
+ "learning_rate": 3.7927350427350425e-05,
413
+ "loss": 0.3254,
414
+ "step": 330
415
+ },
416
+ {
417
+ "epoch": 24.0,
418
+ "eval_accuracy": 0.6941176470588235,
419
+ "eval_loss": 0.9321720600128174,
420
+ "eval_runtime": 0.4281,
421
+ "eval_samples_per_second": 198.553,
422
+ "eval_steps_per_second": 14.016,
423
+ "step": 330
424
+ },
425
+ {
426
+ "epoch": 24.73,
427
+ "learning_rate": 3.739316239316239e-05,
428
+ "loss": 0.3575,
429
+ "step": 340
430
+ },
431
+ {
432
+ "epoch": 24.95,
433
+ "eval_accuracy": 0.7294117647058823,
434
+ "eval_loss": 0.8846125602722168,
435
+ "eval_runtime": 0.4341,
436
+ "eval_samples_per_second": 195.808,
437
+ "eval_steps_per_second": 13.822,
438
+ "step": 343
439
+ },
440
+ {
441
+ "epoch": 25.45,
442
+ "learning_rate": 3.685897435897436e-05,
443
+ "loss": 0.3335,
444
+ "step": 350
445
+ },
446
+ {
447
+ "epoch": 25.96,
448
+ "eval_accuracy": 0.7294117647058823,
449
+ "eval_loss": 0.9651345014572144,
450
+ "eval_runtime": 0.7482,
451
+ "eval_samples_per_second": 113.611,
452
+ "eval_steps_per_second": 8.02,
453
+ "step": 357
454
+ },
455
+ {
456
+ "epoch": 26.18,
457
+ "learning_rate": 3.6324786324786323e-05,
458
+ "loss": 0.3756,
459
+ "step": 360
460
+ },
461
+ {
462
+ "epoch": 26.91,
463
+ "learning_rate": 3.579059829059829e-05,
464
+ "loss": 0.3055,
465
+ "step": 370
466
+ },
467
+ {
468
+ "epoch": 26.98,
469
+ "eval_accuracy": 0.7411764705882353,
470
+ "eval_loss": 0.8926057815551758,
471
+ "eval_runtime": 0.4221,
472
+ "eval_samples_per_second": 201.376,
473
+ "eval_steps_per_second": 14.215,
474
+ "step": 371
475
+ },
476
+ {
477
+ "epoch": 27.64,
478
+ "learning_rate": 3.525641025641026e-05,
479
+ "loss": 0.2969,
480
+ "step": 380
481
+ },
482
+ {
483
+ "epoch": 28.0,
484
+ "eval_accuracy": 0.6823529411764706,
485
+ "eval_loss": 1.05293869972229,
486
+ "eval_runtime": 0.4271,
487
+ "eval_samples_per_second": 199.018,
488
+ "eval_steps_per_second": 14.048,
489
+ "step": 385
490
+ },
491
+ {
492
+ "epoch": 28.36,
493
+ "learning_rate": 3.472222222222222e-05,
494
+ "loss": 0.3243,
495
+ "step": 390
496
+ },
497
+ {
498
+ "epoch": 28.95,
499
+ "eval_accuracy": 0.7294117647058823,
500
+ "eval_loss": 1.065946340560913,
501
+ "eval_runtime": 0.4271,
502
+ "eval_samples_per_second": 199.018,
503
+ "eval_steps_per_second": 14.048,
504
+ "step": 398
505
+ },
506
+ {
507
+ "epoch": 29.09,
508
+ "learning_rate": 3.418803418803419e-05,
509
+ "loss": 0.3214,
510
+ "step": 400
511
+ },
512
+ {
513
+ "epoch": 29.82,
514
+ "learning_rate": 3.365384615384616e-05,
515
+ "loss": 0.3012,
516
+ "step": 410
517
+ },
518
+ {
519
+ "epoch": 29.96,
520
+ "eval_accuracy": 0.788235294117647,
521
+ "eval_loss": 0.9155008792877197,
522
+ "eval_runtime": 0.4251,
523
+ "eval_samples_per_second": 199.955,
524
+ "eval_steps_per_second": 14.114,
525
+ "step": 412
526
+ },
527
+ {
528
+ "epoch": 30.55,
529
+ "learning_rate": 3.311965811965812e-05,
530
+ "loss": 0.2582,
531
+ "step": 420
532
+ },
533
+ {
534
+ "epoch": 30.98,
535
+ "eval_accuracy": 0.7176470588235294,
536
+ "eval_loss": 1.0367075204849243,
537
+ "eval_runtime": 0.4241,
538
+ "eval_samples_per_second": 200.426,
539
+ "eval_steps_per_second": 14.148,
540
+ "step": 426
541
+ },
542
+ {
543
+ "epoch": 31.27,
544
+ "learning_rate": 3.258547008547009e-05,
545
+ "loss": 0.3075,
546
+ "step": 430
547
+ },
548
+ {
549
+ "epoch": 32.0,
550
+ "learning_rate": 3.205128205128206e-05,
551
+ "loss": 0.2881,
552
+ "step": 440
553
+ },
554
+ {
555
+ "epoch": 32.0,
556
+ "eval_accuracy": 0.7176470588235294,
557
+ "eval_loss": 0.9625994563102722,
558
+ "eval_runtime": 0.4311,
559
+ "eval_samples_per_second": 197.171,
560
+ "eval_steps_per_second": 13.918,
561
+ "step": 440
562
+ },
563
+ {
564
+ "epoch": 32.73,
565
+ "learning_rate": 3.151709401709402e-05,
566
+ "loss": 0.2538,
567
+ "step": 450
568
+ },
569
+ {
570
+ "epoch": 32.95,
571
+ "eval_accuracy": 0.7294117647058823,
572
+ "eval_loss": 0.9636082649230957,
573
+ "eval_runtime": 0.4271,
574
+ "eval_samples_per_second": 199.018,
575
+ "eval_steps_per_second": 14.048,
576
+ "step": 453
577
+ },
578
+ {
579
+ "epoch": 33.45,
580
+ "learning_rate": 3.098290598290599e-05,
581
+ "loss": 0.2368,
582
+ "step": 460
583
+ },
584
+ {
585
+ "epoch": 33.96,
586
+ "eval_accuracy": 0.7058823529411765,
587
+ "eval_loss": 1.0623918771743774,
588
+ "eval_runtime": 0.4231,
589
+ "eval_samples_per_second": 200.9,
590
+ "eval_steps_per_second": 14.181,
591
+ "step": 467
592
+ },
593
+ {
594
+ "epoch": 34.18,
595
+ "learning_rate": 3.0448717948717947e-05,
596
+ "loss": 0.2655,
597
+ "step": 470
598
+ },
599
+ {
600
+ "epoch": 34.91,
601
+ "learning_rate": 2.9914529914529915e-05,
602
+ "loss": 0.299,
603
+ "step": 480
604
+ },
605
+ {
606
+ "epoch": 34.98,
607
+ "eval_accuracy": 0.7411764705882353,
608
+ "eval_loss": 0.968308687210083,
609
+ "eval_runtime": 0.4381,
610
+ "eval_samples_per_second": 194.02,
611
+ "eval_steps_per_second": 13.696,
612
+ "step": 481
613
+ },
614
+ {
615
+ "epoch": 35.64,
616
+ "learning_rate": 2.9380341880341884e-05,
617
+ "loss": 0.2553,
618
+ "step": 490
619
+ },
620
+ {
621
+ "epoch": 36.0,
622
+ "eval_accuracy": 0.7176470588235294,
623
+ "eval_loss": 0.9977430701255798,
624
+ "eval_runtime": 0.4541,
625
+ "eval_samples_per_second": 187.182,
626
+ "eval_steps_per_second": 13.213,
627
+ "step": 495
628
+ },
629
+ {
630
+ "epoch": 36.36,
631
+ "learning_rate": 2.8846153846153845e-05,
632
+ "loss": 0.2325,
633
+ "step": 500
634
+ },
635
+ {
636
+ "epoch": 36.95,
637
+ "eval_accuracy": 0.7176470588235294,
638
+ "eval_loss": 0.9950642585754395,
639
+ "eval_runtime": 0.4741,
640
+ "eval_samples_per_second": 179.284,
641
+ "eval_steps_per_second": 12.655,
642
+ "step": 508
643
+ },
644
+ {
645
+ "epoch": 37.09,
646
+ "learning_rate": 2.8311965811965814e-05,
647
+ "loss": 0.1955,
648
+ "step": 510
649
+ },
650
+ {
651
+ "epoch": 37.82,
652
+ "learning_rate": 2.777777777777778e-05,
653
+ "loss": 0.2363,
654
+ "step": 520
655
+ },
656
+ {
657
+ "epoch": 37.96,
658
+ "eval_accuracy": 0.7176470588235294,
659
+ "eval_loss": 1.1467396020889282,
660
+ "eval_runtime": 1.0832,
661
+ "eval_samples_per_second": 78.468,
662
+ "eval_steps_per_second": 5.539,
663
+ "step": 522
664
+ },
665
+ {
666
+ "epoch": 38.55,
667
+ "learning_rate": 2.724358974358974e-05,
668
+ "loss": 0.2424,
669
+ "step": 530
670
+ },
671
+ {
672
+ "epoch": 38.98,
673
+ "eval_accuracy": 0.6941176470588235,
674
+ "eval_loss": 1.133606195449829,
675
+ "eval_runtime": 1.6874,
676
+ "eval_samples_per_second": 50.374,
677
+ "eval_steps_per_second": 3.556,
678
+ "step": 536
679
+ },
680
+ {
681
+ "epoch": 39.27,
682
+ "learning_rate": 2.670940170940171e-05,
683
+ "loss": 0.2323,
684
+ "step": 540
685
+ },
686
+ {
687
+ "epoch": 40.0,
688
+ "learning_rate": 2.6175213675213678e-05,
689
+ "loss": 0.2096,
690
+ "step": 550
691
+ },
692
+ {
693
+ "epoch": 40.0,
694
+ "eval_accuracy": 0.7294117647058823,
695
+ "eval_loss": 1.0931720733642578,
696
+ "eval_runtime": 0.4311,
697
+ "eval_samples_per_second": 197.171,
698
+ "eval_steps_per_second": 13.918,
699
+ "step": 550
700
+ },
701
+ {
702
+ "epoch": 40.73,
703
+ "learning_rate": 2.564102564102564e-05,
704
+ "loss": 0.1978,
705
+ "step": 560
706
+ },
707
+ {
708
+ "epoch": 40.95,
709
+ "eval_accuracy": 0.6941176470588235,
710
+ "eval_loss": 1.139365792274475,
711
+ "eval_runtime": 0.4251,
712
+ "eval_samples_per_second": 199.955,
713
+ "eval_steps_per_second": 14.114,
714
+ "step": 563
715
+ },
716
+ {
717
+ "epoch": 41.45,
718
+ "learning_rate": 2.5106837606837608e-05,
719
+ "loss": 0.1746,
720
+ "step": 570
721
+ },
722
+ {
723
+ "epoch": 41.96,
724
+ "eval_accuracy": 0.7529411764705882,
725
+ "eval_loss": 1.0977457761764526,
726
+ "eval_runtime": 0.4151,
727
+ "eval_samples_per_second": 204.773,
728
+ "eval_steps_per_second": 14.455,
729
+ "step": 577
730
+ },
731
+ {
732
+ "epoch": 42.18,
733
+ "learning_rate": 2.4572649572649573e-05,
734
+ "loss": 0.1675,
735
+ "step": 580
736
+ },
737
+ {
738
+ "epoch": 42.91,
739
+ "learning_rate": 2.4038461538461542e-05,
740
+ "loss": 0.2248,
741
+ "step": 590
742
+ },
743
+ {
744
+ "epoch": 42.98,
745
+ "eval_accuracy": 0.7529411764705882,
746
+ "eval_loss": 1.1300934553146362,
747
+ "eval_runtime": 0.4191,
748
+ "eval_samples_per_second": 202.818,
749
+ "eval_steps_per_second": 14.317,
750
+ "step": 591
751
+ },
752
+ {
753
+ "epoch": 43.64,
754
+ "learning_rate": 2.3504273504273504e-05,
755
+ "loss": 0.2484,
756
+ "step": 600
757
+ },
758
+ {
759
+ "epoch": 44.0,
760
+ "eval_accuracy": 0.6941176470588235,
761
+ "eval_loss": 1.2936742305755615,
762
+ "eval_runtime": 0.4331,
763
+ "eval_samples_per_second": 196.261,
764
+ "eval_steps_per_second": 13.854,
765
+ "step": 605
766
+ },
767
+ {
768
+ "epoch": 44.36,
769
+ "learning_rate": 2.297008547008547e-05,
770
+ "loss": 0.2012,
771
+ "step": 610
772
+ },
773
+ {
774
+ "epoch": 44.95,
775
+ "eval_accuracy": 0.7176470588235294,
776
+ "eval_loss": 1.2552695274353027,
777
+ "eval_runtime": 0.4221,
778
+ "eval_samples_per_second": 201.376,
779
+ "eval_steps_per_second": 14.215,
780
+ "step": 618
781
+ },
782
+ {
783
+ "epoch": 45.09,
784
+ "learning_rate": 2.2435897435897437e-05,
785
+ "loss": 0.2492,
786
+ "step": 620
787
+ },
788
+ {
789
+ "epoch": 45.82,
790
+ "learning_rate": 2.1901709401709402e-05,
791
+ "loss": 0.2347,
792
+ "step": 630
793
+ },
794
+ {
795
+ "epoch": 45.96,
796
+ "eval_accuracy": 0.7411764705882353,
797
+ "eval_loss": 1.1069552898406982,
798
+ "eval_runtime": 0.4191,
799
+ "eval_samples_per_second": 202.818,
800
+ "eval_steps_per_second": 14.317,
801
+ "step": 632
802
+ },
803
+ {
804
+ "epoch": 46.55,
805
+ "learning_rate": 2.1367521367521368e-05,
806
+ "loss": 0.2061,
807
+ "step": 640
808
+ },
809
+ {
810
+ "epoch": 46.98,
811
+ "eval_accuracy": 0.7058823529411765,
812
+ "eval_loss": 1.2092617750167847,
813
+ "eval_runtime": 0.4241,
814
+ "eval_samples_per_second": 200.426,
815
+ "eval_steps_per_second": 14.148,
816
+ "step": 646
817
+ },
818
+ {
819
+ "epoch": 47.27,
820
+ "learning_rate": 2.0833333333333336e-05,
821
+ "loss": 0.1832,
822
+ "step": 650
823
+ },
824
+ {
825
+ "epoch": 48.0,
826
+ "learning_rate": 2.02991452991453e-05,
827
+ "loss": 0.2057,
828
+ "step": 660
829
+ },
830
+ {
831
+ "epoch": 48.0,
832
+ "eval_accuracy": 0.7294117647058823,
833
+ "eval_loss": 1.1720176935195923,
834
+ "eval_runtime": 0.4261,
835
+ "eval_samples_per_second": 199.486,
836
+ "eval_steps_per_second": 14.081,
837
+ "step": 660
838
+ },
839
+ {
840
+ "epoch": 48.73,
841
+ "learning_rate": 1.9764957264957266e-05,
842
+ "loss": 0.1808,
843
+ "step": 670
844
+ },
845
+ {
846
+ "epoch": 48.95,
847
+ "eval_accuracy": 0.6941176470588235,
848
+ "eval_loss": 1.3336892127990723,
849
+ "eval_runtime": 0.4211,
850
+ "eval_samples_per_second": 201.855,
851
+ "eval_steps_per_second": 14.249,
852
+ "step": 673
853
+ },
854
+ {
855
+ "epoch": 49.45,
856
+ "learning_rate": 1.923076923076923e-05,
857
+ "loss": 0.1872,
858
+ "step": 680
859
+ },
860
+ {
861
+ "epoch": 49.96,
862
+ "eval_accuracy": 0.6823529411764706,
863
+ "eval_loss": 1.3495038747787476,
864
+ "eval_runtime": 0.4221,
865
+ "eval_samples_per_second": 201.376,
866
+ "eval_steps_per_second": 14.215,
867
+ "step": 687
868
+ },
869
+ {
870
+ "epoch": 50.18,
871
+ "learning_rate": 1.8696581196581197e-05,
872
+ "loss": 0.1647,
873
+ "step": 690
874
+ },
875
+ {
876
+ "epoch": 50.91,
877
+ "learning_rate": 1.8162393162393162e-05,
878
+ "loss": 0.1584,
879
+ "step": 700
880
+ },
881
+ {
882
+ "epoch": 50.98,
883
+ "eval_accuracy": 0.6941176470588235,
884
+ "eval_loss": 1.3835796117782593,
885
+ "eval_runtime": 0.4181,
886
+ "eval_samples_per_second": 203.303,
887
+ "eval_steps_per_second": 14.351,
888
+ "step": 701
889
+ },
890
+ {
891
+ "epoch": 51.64,
892
+ "learning_rate": 1.762820512820513e-05,
893
+ "loss": 0.1237,
894
+ "step": 710
895
+ },
896
+ {
897
+ "epoch": 52.0,
898
+ "eval_accuracy": 0.7176470588235294,
899
+ "eval_loss": 1.3275655508041382,
900
+ "eval_runtime": 0.4631,
901
+ "eval_samples_per_second": 183.544,
902
+ "eval_steps_per_second": 12.956,
903
+ "step": 715
904
+ },
905
+ {
906
+ "epoch": 52.36,
907
+ "learning_rate": 1.7094017094017095e-05,
908
+ "loss": 0.2186,
909
+ "step": 720
910
+ },
911
+ {
912
+ "epoch": 52.95,
913
+ "eval_accuracy": 0.6823529411764706,
914
+ "eval_loss": 1.3548001050949097,
915
+ "eval_runtime": 0.4231,
916
+ "eval_samples_per_second": 200.901,
917
+ "eval_steps_per_second": 14.181,
918
+ "step": 728
919
+ },
920
+ {
921
+ "epoch": 53.09,
922
+ "learning_rate": 1.655982905982906e-05,
923
+ "loss": 0.2106,
924
+ "step": 730
925
+ },
926
+ {
927
+ "epoch": 53.82,
928
+ "learning_rate": 1.602564102564103e-05,
929
+ "loss": 0.2053,
930
+ "step": 740
931
+ },
932
+ {
933
+ "epoch": 53.96,
934
+ "eval_accuracy": 0.6941176470588235,
935
+ "eval_loss": 1.320010781288147,
936
+ "eval_runtime": 0.4181,
937
+ "eval_samples_per_second": 203.303,
938
+ "eval_steps_per_second": 14.351,
939
+ "step": 742
940
+ },
941
+ {
942
+ "epoch": 54.55,
943
+ "learning_rate": 1.5491452991452994e-05,
944
+ "loss": 0.1618,
945
+ "step": 750
946
+ },
947
+ {
948
+ "epoch": 54.98,
949
+ "eval_accuracy": 0.7058823529411765,
950
+ "eval_loss": 1.2398937940597534,
951
+ "eval_runtime": 0.4191,
952
+ "eval_samples_per_second": 202.818,
953
+ "eval_steps_per_second": 14.317,
954
+ "step": 756
955
+ },
956
+ {
957
+ "epoch": 55.27,
958
+ "learning_rate": 1.4957264957264958e-05,
959
+ "loss": 0.1739,
960
+ "step": 760
961
+ },
962
+ {
963
+ "epoch": 56.0,
964
+ "learning_rate": 1.4423076923076923e-05,
965
+ "loss": 0.1568,
966
+ "step": 770
967
+ },
968
+ {
969
+ "epoch": 56.0,
970
+ "eval_accuracy": 0.7058823529411765,
971
+ "eval_loss": 1.3509585857391357,
972
+ "eval_runtime": 0.4221,
973
+ "eval_samples_per_second": 201.377,
974
+ "eval_steps_per_second": 14.215,
975
+ "step": 770
976
+ },
977
+ {
978
+ "epoch": 56.73,
979
+ "learning_rate": 1.388888888888889e-05,
980
+ "loss": 0.175,
981
+ "step": 780
982
+ },
983
+ {
984
+ "epoch": 56.95,
985
+ "eval_accuracy": 0.7058823529411765,
986
+ "eval_loss": 1.385174036026001,
987
+ "eval_runtime": 0.4231,
988
+ "eval_samples_per_second": 200.901,
989
+ "eval_steps_per_second": 14.181,
990
+ "step": 783
991
+ },
992
+ {
993
+ "epoch": 57.45,
994
+ "learning_rate": 1.3354700854700855e-05,
995
+ "loss": 0.15,
996
+ "step": 790
997
+ },
998
+ {
999
+ "epoch": 57.96,
1000
+ "eval_accuracy": 0.7176470588235294,
1001
+ "eval_loss": 1.369346261024475,
1002
+ "eval_runtime": 0.4161,
1003
+ "eval_samples_per_second": 204.281,
1004
+ "eval_steps_per_second": 14.42,
1005
+ "step": 797
1006
+ },
1007
+ {
1008
+ "epoch": 58.18,
1009
+ "learning_rate": 1.282051282051282e-05,
1010
+ "loss": 0.1607,
1011
+ "step": 800
1012
+ },
1013
+ {
1014
+ "epoch": 58.91,
1015
+ "learning_rate": 1.2286324786324787e-05,
1016
+ "loss": 0.2088,
1017
+ "step": 810
1018
+ },
1019
+ {
1020
+ "epoch": 58.98,
1021
+ "eval_accuracy": 0.6823529411764706,
1022
+ "eval_loss": 1.4165531396865845,
1023
+ "eval_runtime": 0.4171,
1024
+ "eval_samples_per_second": 203.791,
1025
+ "eval_steps_per_second": 14.385,
1026
+ "step": 811
1027
+ },
1028
+ {
1029
+ "epoch": 59.64,
1030
+ "learning_rate": 1.1752136752136752e-05,
1031
+ "loss": 0.1635,
1032
+ "step": 820
1033
+ },
1034
+ {
1035
+ "epoch": 60.0,
1036
+ "eval_accuracy": 0.7058823529411765,
1037
+ "eval_loss": 1.2656704187393188,
1038
+ "eval_runtime": 0.4231,
1039
+ "eval_samples_per_second": 200.9,
1040
+ "eval_steps_per_second": 14.181,
1041
+ "step": 825
1042
+ },
1043
+ {
1044
+ "epoch": 60.36,
1045
+ "learning_rate": 1.1217948717948719e-05,
1046
+ "loss": 0.1681,
1047
+ "step": 830
1048
+ },
1049
+ {
1050
+ "epoch": 60.95,
1051
+ "eval_accuracy": 0.7176470588235294,
1052
+ "eval_loss": 1.3033798933029175,
1053
+ "eval_runtime": 0.4341,
1054
+ "eval_samples_per_second": 195.809,
1055
+ "eval_steps_per_second": 13.822,
1056
+ "step": 838
1057
+ },
1058
+ {
1059
+ "epoch": 61.09,
1060
+ "learning_rate": 1.0683760683760684e-05,
1061
+ "loss": 0.1875,
1062
+ "step": 840
1063
+ },
1064
+ {
1065
+ "epoch": 61.82,
1066
+ "learning_rate": 1.014957264957265e-05,
1067
+ "loss": 0.166,
1068
+ "step": 850
1069
+ },
1070
+ {
1071
+ "epoch": 61.96,
1072
+ "eval_accuracy": 0.7294117647058823,
1073
+ "eval_loss": 1.2799781560897827,
1074
+ "eval_runtime": 0.4171,
1075
+ "eval_samples_per_second": 203.791,
1076
+ "eval_steps_per_second": 14.385,
1077
+ "step": 852
1078
+ },
1079
+ {
1080
+ "epoch": 62.55,
1081
+ "learning_rate": 9.615384615384616e-06,
1082
+ "loss": 0.1585,
1083
+ "step": 860
1084
+ },
1085
+ {
1086
+ "epoch": 62.98,
1087
+ "eval_accuracy": 0.7294117647058823,
1088
+ "eval_loss": 1.3041337728500366,
1089
+ "eval_runtime": 0.4201,
1090
+ "eval_samples_per_second": 202.335,
1091
+ "eval_steps_per_second": 14.283,
1092
+ "step": 866
1093
+ },
1094
+ {
1095
+ "epoch": 63.27,
1096
+ "learning_rate": 9.081196581196581e-06,
1097
+ "loss": 0.1498,
1098
+ "step": 870
1099
+ },
1100
+ {
1101
+ "epoch": 64.0,
1102
+ "learning_rate": 8.547008547008548e-06,
1103
+ "loss": 0.152,
1104
+ "step": 880
1105
+ },
1106
+ {
1107
+ "epoch": 64.0,
1108
+ "eval_accuracy": 0.6823529411764706,
1109
+ "eval_loss": 1.415377140045166,
1110
+ "eval_runtime": 0.4221,
1111
+ "eval_samples_per_second": 201.377,
1112
+ "eval_steps_per_second": 14.215,
1113
+ "step": 880
1114
+ },
1115
+ {
1116
+ "epoch": 64.73,
1117
+ "learning_rate": 8.012820512820515e-06,
1118
+ "loss": 0.155,
1119
+ "step": 890
1120
+ },
1121
+ {
1122
+ "epoch": 64.95,
1123
+ "eval_accuracy": 0.6823529411764706,
1124
+ "eval_loss": 1.3989214897155762,
1125
+ "eval_runtime": 0.4201,
1126
+ "eval_samples_per_second": 202.336,
1127
+ "eval_steps_per_second": 14.283,
1128
+ "step": 893
1129
+ },
1130
+ {
1131
+ "epoch": 65.45,
1132
+ "learning_rate": 7.478632478632479e-06,
1133
+ "loss": 0.1547,
1134
+ "step": 900
1135
+ },
1136
+ {
1137
+ "epoch": 65.96,
1138
+ "eval_accuracy": 0.7176470588235294,
1139
+ "eval_loss": 1.3432484865188599,
1140
+ "eval_runtime": 0.4241,
1141
+ "eval_samples_per_second": 200.426,
1142
+ "eval_steps_per_second": 14.148,
1143
+ "step": 907
1144
+ },
1145
+ {
1146
+ "epoch": 66.18,
1147
+ "learning_rate": 6.944444444444445e-06,
1148
+ "loss": 0.1666,
1149
+ "step": 910
1150
+ },
1151
+ {
1152
+ "epoch": 66.91,
1153
+ "learning_rate": 6.41025641025641e-06,
1154
+ "loss": 0.1113,
1155
+ "step": 920
1156
+ },
1157
+ {
1158
+ "epoch": 66.98,
1159
+ "eval_accuracy": 0.7294117647058823,
1160
+ "eval_loss": 1.3281364440917969,
1161
+ "eval_runtime": 0.4191,
1162
+ "eval_samples_per_second": 202.819,
1163
+ "eval_steps_per_second": 14.317,
1164
+ "step": 921
1165
+ },
1166
+ {
1167
+ "epoch": 67.64,
1168
+ "learning_rate": 5.876068376068376e-06,
1169
+ "loss": 0.159,
1170
+ "step": 930
1171
+ },
1172
+ {
1173
+ "epoch": 68.0,
1174
+ "eval_accuracy": 0.7294117647058823,
1175
+ "eval_loss": 1.3717232942581177,
1176
+ "eval_runtime": 0.4141,
1177
+ "eval_samples_per_second": 205.267,
1178
+ "eval_steps_per_second": 14.489,
1179
+ "step": 935
1180
+ },
1181
+ {
1182
+ "epoch": 68.36,
1183
+ "learning_rate": 5.341880341880342e-06,
1184
+ "loss": 0.1149,
1185
+ "step": 940
1186
+ },
1187
+ {
1188
+ "epoch": 68.95,
1189
+ "eval_accuracy": 0.7294117647058823,
1190
+ "eval_loss": 1.37764310836792,
1191
+ "eval_runtime": 0.4141,
1192
+ "eval_samples_per_second": 205.268,
1193
+ "eval_steps_per_second": 14.489,
1194
+ "step": 948
1195
+ },
1196
+ {
1197
+ "epoch": 69.09,
1198
+ "learning_rate": 4.807692307692308e-06,
1199
+ "loss": 0.1622,
1200
+ "step": 950
1201
+ },
1202
+ {
1203
+ "epoch": 69.82,
1204
+ "learning_rate": 4.273504273504274e-06,
1205
+ "loss": 0.1478,
1206
+ "step": 960
1207
+ },
1208
+ {
1209
+ "epoch": 69.96,
1210
+ "eval_accuracy": 0.7294117647058823,
1211
+ "eval_loss": 1.3459463119506836,
1212
+ "eval_runtime": 0.4091,
1213
+ "eval_samples_per_second": 207.777,
1214
+ "eval_steps_per_second": 14.667,
1215
+ "step": 962
1216
+ },
1217
+ {
1218
+ "epoch": 70.55,
1219
+ "learning_rate": 3.7393162393162394e-06,
1220
+ "loss": 0.1675,
1221
+ "step": 970
1222
+ },
1223
+ {
1224
+ "epoch": 70.98,
1225
+ "eval_accuracy": 0.7294117647058823,
1226
+ "eval_loss": 1.3941361904144287,
1227
+ "eval_runtime": 0.4201,
1228
+ "eval_samples_per_second": 202.335,
1229
+ "eval_steps_per_second": 14.282,
1230
+ "step": 976
1231
+ },
1232
+ {
1233
+ "epoch": 71.27,
1234
+ "learning_rate": 3.205128205128205e-06,
1235
+ "loss": 0.1617,
1236
+ "step": 980
1237
+ },
1238
+ {
1239
+ "epoch": 72.0,
1240
+ "learning_rate": 2.670940170940171e-06,
1241
+ "loss": 0.1476,
1242
+ "step": 990
1243
+ },
1244
+ {
1245
+ "epoch": 72.0,
1246
+ "eval_accuracy": 0.7058823529411765,
1247
+ "eval_loss": 1.4110150337219238,
1248
+ "eval_runtime": 0.4201,
1249
+ "eval_samples_per_second": 202.335,
1250
+ "eval_steps_per_second": 14.282,
1251
+ "step": 990
1252
+ },
1253
+ {
1254
+ "epoch": 72.73,
1255
+ "learning_rate": 2.136752136752137e-06,
1256
+ "loss": 0.1515,
1257
+ "step": 1000
1258
+ },
1259
+ {
1260
+ "epoch": 72.95,
1261
+ "eval_accuracy": 0.7058823529411765,
1262
+ "eval_loss": 1.370444893836975,
1263
+ "eval_runtime": 0.4171,
1264
+ "eval_samples_per_second": 203.791,
1265
+ "eval_steps_per_second": 14.385,
1266
+ "step": 1003
1267
+ },
1268
+ {
1269
+ "epoch": 73.45,
1270
+ "learning_rate": 1.6025641025641025e-06,
1271
+ "loss": 0.1635,
1272
+ "step": 1010
1273
+ },
1274
+ {
1275
+ "epoch": 73.96,
1276
+ "eval_accuracy": 0.7176470588235294,
1277
+ "eval_loss": 1.3646893501281738,
1278
+ "eval_runtime": 0.4201,
1279
+ "eval_samples_per_second": 202.335,
1280
+ "eval_steps_per_second": 14.282,
1281
+ "step": 1017
1282
+ },
1283
+ {
1284
+ "epoch": 74.18,
1285
+ "learning_rate": 1.0683760683760685e-06,
1286
+ "loss": 0.1281,
1287
+ "step": 1020
1288
+ },
1289
+ {
1290
+ "epoch": 74.91,
1291
+ "learning_rate": 5.341880341880342e-07,
1292
+ "loss": 0.1445,
1293
+ "step": 1030
1294
+ },
1295
+ {
1296
+ "epoch": 74.98,
1297
+ "eval_accuracy": 0.7176470588235294,
1298
+ "eval_loss": 1.3672693967819214,
1299
+ "eval_runtime": 0.4071,
1300
+ "eval_samples_per_second": 208.798,
1301
+ "eval_steps_per_second": 14.739,
1302
+ "step": 1031
1303
+ },
1304
+ {
1305
+ "epoch": 75.64,
1306
+ "learning_rate": 0.0,
1307
+ "loss": 0.1398,
1308
+ "step": 1040
1309
+ },
1310
+ {
1311
+ "epoch": 75.64,
1312
+ "eval_accuracy": 0.7176470588235294,
1313
+ "eval_loss": 1.3689277172088623,
1314
+ "eval_runtime": 0.4071,
1315
+ "eval_samples_per_second": 208.798,
1316
+ "eval_steps_per_second": 14.739,
1317
+ "step": 1040
1318
+ },
1319
+ {
1320
+ "epoch": 75.64,
1321
+ "step": 1040,
1322
+ "total_flos": 2.15834520885461e+18,
1323
+ "train_loss": 0.41205361909591237,
1324
+ "train_runtime": 895.6069,
1325
+ "train_samples_per_second": 78.338,
1326
+ "train_steps_per_second": 1.161
1327
+ }
1328
+ ],
1329
+ "logging_steps": 10,
1330
+ "max_steps": 1040,
1331
+ "num_input_tokens_seen": 0,
1332
+ "num_train_epochs": 80,
1333
+ "save_steps": 500,
1334
+ "total_flos": 2.15834520885461e+18,
1335
+ "train_batch_size": 16,
1336
+ "trial_name": null,
1337
+ "trial_params": null
1338
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eaaed6432f8abbc6dc4e038dd2ef0ea9f136ec698c8ffa2b08a649af4d60304e
3
+ size 4728