Augusto777 commited on
Commit
c5386cd
·
verified ·
1 Parent(s): 78d9156

End of training

Browse files
README.md ADDED
@@ -0,0 +1,157 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: microsoft/beit-base-patch16-224
4
+ tags:
5
+ - generated_from_trainer
6
+ datasets:
7
+ - imagefolder
8
+ metrics:
9
+ - accuracy
10
+ model-index:
11
+ - name: BEiT-DMAE-13XDA-REVAL-80-32
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: 0.8478260869565217
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
+ # BEiT-DMAE-13XDA-REVAL-80-32
32
+
33
+ This model is a fine-tuned version of [microsoft/beit-base-patch16-224](https://huggingface.co/microsoft/beit-base-patch16-224) on the imagefolder dataset.
34
+ It achieves the following results on the evaluation set:
35
+ - Loss: 0.8105
36
+ - Accuracy: 0.8478
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: 3.5e-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
+ | 1.5474 | 1.0 | 60 | 1.2945 | 0.4565 |
71
+ | 1.3959 | 1.99 | 120 | 1.2745 | 0.4565 |
72
+ | 1.0517 | 2.99 | 180 | 0.9632 | 0.6087 |
73
+ | 0.7273 | 4.0 | 241 | 0.7709 | 0.6957 |
74
+ | 0.5246 | 5.0 | 301 | 0.7217 | 0.7391 |
75
+ | 0.3645 | 5.99 | 361 | 0.7142 | 0.8043 |
76
+ | 0.2211 | 6.99 | 421 | 0.6436 | 0.8043 |
77
+ | 0.266 | 8.0 | 482 | 1.1316 | 0.6087 |
78
+ | 0.1235 | 9.0 | 542 | 0.9257 | 0.7826 |
79
+ | 0.1613 | 9.99 | 602 | 0.8527 | 0.7826 |
80
+ | 0.0946 | 10.99 | 662 | 0.8274 | 0.8043 |
81
+ | 0.1392 | 12.0 | 723 | 0.8312 | 0.7609 |
82
+ | 0.1028 | 13.0 | 783 | 1.1959 | 0.7609 |
83
+ | 0.1072 | 13.99 | 843 | 1.0017 | 0.7391 |
84
+ | 0.0888 | 14.99 | 903 | 0.9214 | 0.8043 |
85
+ | 0.0951 | 16.0 | 964 | 0.9156 | 0.7609 |
86
+ | 0.0714 | 17.0 | 1024 | 1.3116 | 0.6957 |
87
+ | 0.0804 | 17.99 | 1084 | 1.1107 | 0.7826 |
88
+ | 0.08 | 18.99 | 1144 | 0.8105 | 0.8478 |
89
+ | 0.1619 | 20.0 | 1205 | 0.7581 | 0.8261 |
90
+ | 0.084 | 21.0 | 1265 | 1.0210 | 0.8261 |
91
+ | 0.072 | 21.99 | 1325 | 1.3092 | 0.7609 |
92
+ | 0.0303 | 22.99 | 1385 | 1.3367 | 0.7826 |
93
+ | 0.0228 | 24.0 | 1446 | 1.0277 | 0.8261 |
94
+ | 0.0755 | 25.0 | 1506 | 0.9436 | 0.8261 |
95
+ | 0.0756 | 25.99 | 1566 | 1.1588 | 0.7609 |
96
+ | 0.0875 | 26.99 | 1626 | 1.3280 | 0.7174 |
97
+ | 0.0771 | 28.0 | 1687 | 1.8558 | 0.6739 |
98
+ | 0.0467 | 29.0 | 1747 | 1.6476 | 0.7391 |
99
+ | 0.0382 | 29.99 | 1807 | 0.9374 | 0.8478 |
100
+ | 0.0511 | 30.99 | 1867 | 1.0847 | 0.8043 |
101
+ | 0.0161 | 32.0 | 1928 | 1.2028 | 0.7826 |
102
+ | 0.0301 | 33.0 | 1988 | 1.2971 | 0.7391 |
103
+ | 0.0443 | 33.99 | 2048 | 1.3993 | 0.7174 |
104
+ | 0.0782 | 34.99 | 2108 | 1.3359 | 0.8043 |
105
+ | 0.0287 | 36.0 | 2169 | 1.3011 | 0.7826 |
106
+ | 0.0347 | 37.0 | 2229 | 1.2450 | 0.7826 |
107
+ | 0.0538 | 37.99 | 2289 | 1.8216 | 0.7609 |
108
+ | 0.027 | 38.99 | 2349 | 1.1701 | 0.8043 |
109
+ | 0.038 | 40.0 | 2410 | 1.1025 | 0.8043 |
110
+ | 0.0244 | 41.0 | 2470 | 1.2912 | 0.7609 |
111
+ | 0.0122 | 41.99 | 2530 | 1.5699 | 0.7609 |
112
+ | 0.023 | 42.99 | 2590 | 1.5114 | 0.7826 |
113
+ | 0.0297 | 44.0 | 2651 | 1.2189 | 0.8478 |
114
+ | 0.0284 | 45.0 | 2711 | 1.3997 | 0.7826 |
115
+ | 0.0203 | 45.99 | 2771 | 1.4792 | 0.8043 |
116
+ | 0.03 | 46.99 | 2831 | 1.7487 | 0.7174 |
117
+ | 0.025 | 48.0 | 2892 | 1.6605 | 0.7609 |
118
+ | 0.0134 | 49.0 | 2952 | 1.4106 | 0.7826 |
119
+ | 0.026 | 49.99 | 3012 | 1.2972 | 0.7609 |
120
+ | 0.0507 | 50.99 | 3072 | 1.3303 | 0.7826 |
121
+ | 0.0394 | 52.0 | 3133 | 1.1954 | 0.8478 |
122
+ | 0.0271 | 53.0 | 3193 | 1.3125 | 0.8261 |
123
+ | 0.0115 | 53.99 | 3253 | 1.3444 | 0.8478 |
124
+ | 0.0138 | 54.99 | 3313 | 1.4689 | 0.8261 |
125
+ | 0.0184 | 56.0 | 3374 | 1.4959 | 0.8261 |
126
+ | 0.0163 | 57.0 | 3434 | 1.3490 | 0.7826 |
127
+ | 0.0112 | 57.99 | 3494 | 1.4749 | 0.7826 |
128
+ | 0.0185 | 58.99 | 3554 | 1.5823 | 0.7826 |
129
+ | 0.031 | 60.0 | 3615 | 1.5190 | 0.7826 |
130
+ | 0.0161 | 61.0 | 3675 | 1.5476 | 0.8043 |
131
+ | 0.0146 | 61.99 | 3735 | 1.3930 | 0.7826 |
132
+ | 0.005 | 62.99 | 3795 | 1.5454 | 0.8043 |
133
+ | 0.0093 | 64.0 | 3856 | 1.5959 | 0.7826 |
134
+ | 0.0224 | 65.0 | 3916 | 1.4554 | 0.8043 |
135
+ | 0.0154 | 65.99 | 3976 | 1.5327 | 0.8261 |
136
+ | 0.0116 | 66.99 | 4036 | 1.6030 | 0.8043 |
137
+ | 0.0037 | 68.0 | 4097 | 1.5046 | 0.8261 |
138
+ | 0.0023 | 69.0 | 4157 | 1.5222 | 0.8261 |
139
+ | 0.0068 | 69.99 | 4217 | 1.4339 | 0.8261 |
140
+ | 0.0342 | 70.99 | 4277 | 1.6964 | 0.8043 |
141
+ | 0.0077 | 72.0 | 4338 | 1.6102 | 0.8043 |
142
+ | 0.0043 | 73.0 | 4398 | 1.6687 | 0.8043 |
143
+ | 0.0131 | 73.99 | 4458 | 1.6847 | 0.8043 |
144
+ | 0.0031 | 74.99 | 4518 | 1.7195 | 0.8043 |
145
+ | 0.0087 | 76.0 | 4579 | 1.7209 | 0.7826 |
146
+ | 0.0219 | 77.0 | 4639 | 1.6715 | 0.8043 |
147
+ | 0.0229 | 77.99 | 4699 | 1.6823 | 0.8043 |
148
+ | 0.008 | 78.99 | 4759 | 1.6751 | 0.8043 |
149
+ | 0.0051 | 79.67 | 4800 | 1.6758 | 0.8043 |
150
+
151
+
152
+ ### Framework versions
153
+
154
+ - Transformers 4.36.2
155
+ - Pytorch 2.1.2+cu118
156
+ - Datasets 2.16.1
157
+ - Tokenizers 0.15.0
all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 79.67,
3
+ "eval_accuracy": 0.8478260869565217,
4
+ "eval_loss": 0.8104602694511414,
5
+ "eval_runtime": 0.8232,
6
+ "eval_samples_per_second": 55.881,
7
+ "eval_steps_per_second": 3.644,
8
+ "train_loss": 0.12199072747869649,
9
+ "train_runtime": 7091.3369,
10
+ "train_samples_per_second": 43.467,
11
+ "train_steps_per_second": 0.677
12
+ }
config.json ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/beit-base-patch16-224",
3
+ "add_fpn": false,
4
+ "architectures": [
5
+ "BeitForImageClassification"
6
+ ],
7
+ "attention_probs_dropout_prob": 0.0,
8
+ "auxiliary_channels": 256,
9
+ "auxiliary_concat_input": false,
10
+ "auxiliary_loss_weight": 0.4,
11
+ "auxiliary_num_convs": 1,
12
+ "drop_path_rate": 0.1,
13
+ "hidden_act": "gelu",
14
+ "hidden_dropout_prob": 0.0,
15
+ "hidden_size": 768,
16
+ "id2label": {
17
+ "0": "avanzada",
18
+ "1": "leve",
19
+ "2": "moderada",
20
+ "3": "no dmae"
21
+ },
22
+ "image_size": 224,
23
+ "initializer_range": 0.02,
24
+ "intermediate_size": 3072,
25
+ "label2id": {
26
+ "avanzada": 0,
27
+ "leve": 1,
28
+ "moderada": 2,
29
+ "no dmae": 3
30
+ },
31
+ "layer_norm_eps": 1e-12,
32
+ "layer_scale_init_value": 0.1,
33
+ "model_type": "beit",
34
+ "num_attention_heads": 12,
35
+ "num_channels": 3,
36
+ "num_hidden_layers": 12,
37
+ "out_features": [
38
+ "stage12"
39
+ ],
40
+ "out_indices": [
41
+ 12
42
+ ],
43
+ "patch_size": 16,
44
+ "pool_scales": [
45
+ 1,
46
+ 2,
47
+ 3,
48
+ 6
49
+ ],
50
+ "problem_type": "single_label_classification",
51
+ "reshape_hidden_states": true,
52
+ "semantic_loss_ignore_index": 255,
53
+ "stage_names": [
54
+ "stem",
55
+ "stage1",
56
+ "stage2",
57
+ "stage3",
58
+ "stage4",
59
+ "stage5",
60
+ "stage6",
61
+ "stage7",
62
+ "stage8",
63
+ "stage9",
64
+ "stage10",
65
+ "stage11",
66
+ "stage12"
67
+ ],
68
+ "torch_dtype": "float32",
69
+ "transformers_version": "4.36.2",
70
+ "use_absolute_position_embeddings": false,
71
+ "use_auxiliary_head": true,
72
+ "use_mask_token": false,
73
+ "use_mean_pooling": true,
74
+ "use_relative_position_bias": true,
75
+ "use_shared_relative_position_bias": false,
76
+ "vocab_size": 8192
77
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 79.67,
3
+ "eval_accuracy": 0.8478260869565217,
4
+ "eval_loss": 0.8104602694511414,
5
+ "eval_runtime": 0.8232,
6
+ "eval_samples_per_second": 55.881,
7
+ "eval_steps_per_second": 3.644
8
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2023469c5a56435584bf0fc942ffda18b112172b479eabb680782c5c2625e019
3
+ size 343086480
preprocessor_config.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": {
3
+ "height": 224,
4
+ "width": 224
5
+ },
6
+ "do_center_crop": false,
7
+ "do_normalize": true,
8
+ "do_reduce_labels": false,
9
+ "do_rescale": true,
10
+ "do_resize": true,
11
+ "image_mean": [
12
+ 0.5,
13
+ 0.5,
14
+ 0.5
15
+ ],
16
+ "image_processor_type": "BeitImageProcessor",
17
+ "image_std": [
18
+ 0.5,
19
+ 0.5,
20
+ 0.5
21
+ ],
22
+ "resample": 2,
23
+ "rescale_factor": 0.00392156862745098,
24
+ "size": {
25
+ "height": 224,
26
+ "width": 224
27
+ }
28
+ }
runs/Dec11_18-55-56_DESKTOP-SKBE9FB/events.out.tfevents.1733964957.DESKTOP-SKBE9FB.18380.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:50cff70628a210535d507d3d96c0c8790c006f84559a44b0d69c0c9820a9a36f
3
+ size 106820
runs/Dec11_18-55-56_DESKTOP-SKBE9FB/events.out.tfevents.1733972049.DESKTOP-SKBE9FB.18380.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f258c795d204e2a90b011b4c22eb8f2dd803d83d11d1b7d7e9292930e95bdac6
3
+ size 411
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 79.67,
3
+ "train_loss": 0.12199072747869649,
4
+ "train_runtime": 7091.3369,
5
+ "train_samples_per_second": 43.467,
6
+ "train_steps_per_second": 0.677
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,3630 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8478260869565217,
3
+ "best_model_checkpoint": "BEiT-DMAE-13XDA-REVAL-80-32\\checkpoint-1144",
4
+ "epoch": 79.66804979253112,
5
+ "eval_steps": 500,
6
+ "global_step": 4800,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.17,
13
+ "learning_rate": 7.291666666666665e-07,
14
+ "loss": 1.6222,
15
+ "step": 10
16
+ },
17
+ {
18
+ "epoch": 0.33,
19
+ "learning_rate": 1.458333333333333e-06,
20
+ "loss": 1.6488,
21
+ "step": 20
22
+ },
23
+ {
24
+ "epoch": 0.5,
25
+ "learning_rate": 2.1875e-06,
26
+ "loss": 1.611,
27
+ "step": 30
28
+ },
29
+ {
30
+ "epoch": 0.66,
31
+ "learning_rate": 2.916666666666666e-06,
32
+ "loss": 1.6356,
33
+ "step": 40
34
+ },
35
+ {
36
+ "epoch": 0.83,
37
+ "learning_rate": 3.6458333333333333e-06,
38
+ "loss": 1.602,
39
+ "step": 50
40
+ },
41
+ {
42
+ "epoch": 1.0,
43
+ "learning_rate": 4.375e-06,
44
+ "loss": 1.5474,
45
+ "step": 60
46
+ },
47
+ {
48
+ "epoch": 1.0,
49
+ "eval_accuracy": 0.45652173913043476,
50
+ "eval_loss": 1.2944936752319336,
51
+ "eval_runtime": 0.6632,
52
+ "eval_samples_per_second": 69.361,
53
+ "eval_steps_per_second": 4.524,
54
+ "step": 60
55
+ },
56
+ {
57
+ "epoch": 1.16,
58
+ "learning_rate": 5.104166666666667e-06,
59
+ "loss": 1.561,
60
+ "step": 70
61
+ },
62
+ {
63
+ "epoch": 1.33,
64
+ "learning_rate": 5.833333333333332e-06,
65
+ "loss": 1.4872,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 1.49,
70
+ "learning_rate": 6.5624999999999994e-06,
71
+ "loss": 1.458,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 1.66,
76
+ "learning_rate": 7.291666666666667e-06,
77
+ "loss": 1.4426,
78
+ "step": 100
79
+ },
80
+ {
81
+ "epoch": 1.83,
82
+ "learning_rate": 8.020833333333333e-06,
83
+ "loss": 1.4296,
84
+ "step": 110
85
+ },
86
+ {
87
+ "epoch": 1.99,
88
+ "learning_rate": 8.75e-06,
89
+ "loss": 1.3959,
90
+ "step": 120
91
+ },
92
+ {
93
+ "epoch": 1.99,
94
+ "eval_accuracy": 0.45652173913043476,
95
+ "eval_loss": 1.2745169401168823,
96
+ "eval_runtime": 0.6327,
97
+ "eval_samples_per_second": 72.708,
98
+ "eval_steps_per_second": 4.742,
99
+ "step": 120
100
+ },
101
+ {
102
+ "epoch": 2.16,
103
+ "learning_rate": 9.479166666666666e-06,
104
+ "loss": 1.3373,
105
+ "step": 130
106
+ },
107
+ {
108
+ "epoch": 2.32,
109
+ "learning_rate": 1.0208333333333334e-05,
110
+ "loss": 1.2946,
111
+ "step": 140
112
+ },
113
+ {
114
+ "epoch": 2.49,
115
+ "learning_rate": 1.0937499999999998e-05,
116
+ "loss": 1.236,
117
+ "step": 150
118
+ },
119
+ {
120
+ "epoch": 2.66,
121
+ "learning_rate": 1.1666666666666665e-05,
122
+ "loss": 1.1899,
123
+ "step": 160
124
+ },
125
+ {
126
+ "epoch": 2.82,
127
+ "learning_rate": 1.2395833333333333e-05,
128
+ "loss": 1.1186,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 2.99,
133
+ "learning_rate": 1.3124999999999999e-05,
134
+ "loss": 1.0517,
135
+ "step": 180
136
+ },
137
+ {
138
+ "epoch": 2.99,
139
+ "eval_accuracy": 0.6086956521739131,
140
+ "eval_loss": 0.9631962180137634,
141
+ "eval_runtime": 0.6872,
142
+ "eval_samples_per_second": 66.943,
143
+ "eval_steps_per_second": 4.366,
144
+ "step": 180
145
+ },
146
+ {
147
+ "epoch": 3.15,
148
+ "learning_rate": 1.3854166666666665e-05,
149
+ "loss": 0.9828,
150
+ "step": 190
151
+ },
152
+ {
153
+ "epoch": 3.32,
154
+ "learning_rate": 1.4583333333333333e-05,
155
+ "loss": 0.9259,
156
+ "step": 200
157
+ },
158
+ {
159
+ "epoch": 3.49,
160
+ "learning_rate": 1.53125e-05,
161
+ "loss": 0.8746,
162
+ "step": 210
163
+ },
164
+ {
165
+ "epoch": 3.65,
166
+ "learning_rate": 1.6041666666666666e-05,
167
+ "loss": 0.8467,
168
+ "step": 220
169
+ },
170
+ {
171
+ "epoch": 3.82,
172
+ "learning_rate": 1.6770833333333332e-05,
173
+ "loss": 0.7669,
174
+ "step": 230
175
+ },
176
+ {
177
+ "epoch": 3.98,
178
+ "learning_rate": 1.75e-05,
179
+ "loss": 0.7273,
180
+ "step": 240
181
+ },
182
+ {
183
+ "epoch": 4.0,
184
+ "eval_accuracy": 0.6956521739130435,
185
+ "eval_loss": 0.7708569169044495,
186
+ "eval_runtime": 0.6952,
187
+ "eval_samples_per_second": 66.169,
188
+ "eval_steps_per_second": 4.315,
189
+ "step": 241
190
+ },
191
+ {
192
+ "epoch": 4.15,
193
+ "learning_rate": 1.8229166666666665e-05,
194
+ "loss": 0.6937,
195
+ "step": 250
196
+ },
197
+ {
198
+ "epoch": 4.32,
199
+ "learning_rate": 1.895833333333333e-05,
200
+ "loss": 0.6738,
201
+ "step": 260
202
+ },
203
+ {
204
+ "epoch": 4.48,
205
+ "learning_rate": 1.9687499999999997e-05,
206
+ "loss": 0.6577,
207
+ "step": 270
208
+ },
209
+ {
210
+ "epoch": 4.65,
211
+ "learning_rate": 2.0416666666666667e-05,
212
+ "loss": 0.6679,
213
+ "step": 280
214
+ },
215
+ {
216
+ "epoch": 4.81,
217
+ "learning_rate": 2.114583333333333e-05,
218
+ "loss": 0.5962,
219
+ "step": 290
220
+ },
221
+ {
222
+ "epoch": 4.98,
223
+ "learning_rate": 2.1874999999999996e-05,
224
+ "loss": 0.5246,
225
+ "step": 300
226
+ },
227
+ {
228
+ "epoch": 5.0,
229
+ "eval_accuracy": 0.7391304347826086,
230
+ "eval_loss": 0.7217094302177429,
231
+ "eval_runtime": 0.6875,
232
+ "eval_samples_per_second": 66.909,
233
+ "eval_steps_per_second": 4.364,
234
+ "step": 301
235
+ },
236
+ {
237
+ "epoch": 5.15,
238
+ "learning_rate": 2.2604166666666666e-05,
239
+ "loss": 0.4905,
240
+ "step": 310
241
+ },
242
+ {
243
+ "epoch": 5.31,
244
+ "learning_rate": 2.333333333333333e-05,
245
+ "loss": 0.4651,
246
+ "step": 320
247
+ },
248
+ {
249
+ "epoch": 5.48,
250
+ "learning_rate": 2.40625e-05,
251
+ "loss": 0.3947,
252
+ "step": 330
253
+ },
254
+ {
255
+ "epoch": 5.64,
256
+ "learning_rate": 2.4791666666666665e-05,
257
+ "loss": 0.3819,
258
+ "step": 340
259
+ },
260
+ {
261
+ "epoch": 5.81,
262
+ "learning_rate": 2.552083333333333e-05,
263
+ "loss": 0.3977,
264
+ "step": 350
265
+ },
266
+ {
267
+ "epoch": 5.98,
268
+ "learning_rate": 2.6249999999999998e-05,
269
+ "loss": 0.3645,
270
+ "step": 360
271
+ },
272
+ {
273
+ "epoch": 5.99,
274
+ "eval_accuracy": 0.8043478260869565,
275
+ "eval_loss": 0.7142456769943237,
276
+ "eval_runtime": 0.6651,
277
+ "eval_samples_per_second": 69.157,
278
+ "eval_steps_per_second": 4.51,
279
+ "step": 361
280
+ },
281
+ {
282
+ "epoch": 6.14,
283
+ "learning_rate": 2.6979166666666664e-05,
284
+ "loss": 0.3382,
285
+ "step": 370
286
+ },
287
+ {
288
+ "epoch": 6.31,
289
+ "learning_rate": 2.770833333333333e-05,
290
+ "loss": 0.2709,
291
+ "step": 380
292
+ },
293
+ {
294
+ "epoch": 6.47,
295
+ "learning_rate": 2.8437499999999997e-05,
296
+ "loss": 0.2641,
297
+ "step": 390
298
+ },
299
+ {
300
+ "epoch": 6.64,
301
+ "learning_rate": 2.9166666666666666e-05,
302
+ "loss": 0.2518,
303
+ "step": 400
304
+ },
305
+ {
306
+ "epoch": 6.8,
307
+ "learning_rate": 2.989583333333333e-05,
308
+ "loss": 0.2631,
309
+ "step": 410
310
+ },
311
+ {
312
+ "epoch": 6.97,
313
+ "learning_rate": 3.0625e-05,
314
+ "loss": 0.2211,
315
+ "step": 420
316
+ },
317
+ {
318
+ "epoch": 6.99,
319
+ "eval_accuracy": 0.8043478260869565,
320
+ "eval_loss": 0.6436068415641785,
321
+ "eval_runtime": 0.6647,
322
+ "eval_samples_per_second": 69.209,
323
+ "eval_steps_per_second": 4.514,
324
+ "step": 421
325
+ },
326
+ {
327
+ "epoch": 7.14,
328
+ "learning_rate": 3.1354166666666665e-05,
329
+ "loss": 0.225,
330
+ "step": 430
331
+ },
332
+ {
333
+ "epoch": 7.3,
334
+ "learning_rate": 3.208333333333333e-05,
335
+ "loss": 0.2625,
336
+ "step": 440
337
+ },
338
+ {
339
+ "epoch": 7.47,
340
+ "learning_rate": 3.28125e-05,
341
+ "loss": 0.183,
342
+ "step": 450
343
+ },
344
+ {
345
+ "epoch": 7.63,
346
+ "learning_rate": 3.3541666666666664e-05,
347
+ "loss": 0.1866,
348
+ "step": 460
349
+ },
350
+ {
351
+ "epoch": 7.8,
352
+ "learning_rate": 3.427083333333333e-05,
353
+ "loss": 0.2547,
354
+ "step": 470
355
+ },
356
+ {
357
+ "epoch": 7.97,
358
+ "learning_rate": 3.5e-05,
359
+ "loss": 0.266,
360
+ "step": 480
361
+ },
362
+ {
363
+ "epoch": 8.0,
364
+ "eval_accuracy": 0.6086956521739131,
365
+ "eval_loss": 1.1315726041793823,
366
+ "eval_runtime": 0.6621,
367
+ "eval_samples_per_second": 69.472,
368
+ "eval_steps_per_second": 4.531,
369
+ "step": 482
370
+ },
371
+ {
372
+ "epoch": 8.13,
373
+ "learning_rate": 3.491898148148148e-05,
374
+ "loss": 0.2918,
375
+ "step": 490
376
+ },
377
+ {
378
+ "epoch": 8.3,
379
+ "learning_rate": 3.483796296296296e-05,
380
+ "loss": 0.209,
381
+ "step": 500
382
+ },
383
+ {
384
+ "epoch": 8.46,
385
+ "learning_rate": 3.475694444444444e-05,
386
+ "loss": 0.1858,
387
+ "step": 510
388
+ },
389
+ {
390
+ "epoch": 8.63,
391
+ "learning_rate": 3.467592592592592e-05,
392
+ "loss": 0.1967,
393
+ "step": 520
394
+ },
395
+ {
396
+ "epoch": 8.8,
397
+ "learning_rate": 3.459490740740741e-05,
398
+ "loss": 0.1914,
399
+ "step": 530
400
+ },
401
+ {
402
+ "epoch": 8.96,
403
+ "learning_rate": 3.4513888888888886e-05,
404
+ "loss": 0.1235,
405
+ "step": 540
406
+ },
407
+ {
408
+ "epoch": 9.0,
409
+ "eval_accuracy": 0.782608695652174,
410
+ "eval_loss": 0.9256875514984131,
411
+ "eval_runtime": 0.6634,
412
+ "eval_samples_per_second": 69.344,
413
+ "eval_steps_per_second": 4.522,
414
+ "step": 542
415
+ },
416
+ {
417
+ "epoch": 9.13,
418
+ "learning_rate": 3.4432870370370365e-05,
419
+ "loss": 0.2088,
420
+ "step": 550
421
+ },
422
+ {
423
+ "epoch": 9.29,
424
+ "learning_rate": 3.435185185185185e-05,
425
+ "loss": 0.2034,
426
+ "step": 560
427
+ },
428
+ {
429
+ "epoch": 9.46,
430
+ "learning_rate": 3.427083333333333e-05,
431
+ "loss": 0.1289,
432
+ "step": 570
433
+ },
434
+ {
435
+ "epoch": 9.63,
436
+ "learning_rate": 3.418981481481481e-05,
437
+ "loss": 0.162,
438
+ "step": 580
439
+ },
440
+ {
441
+ "epoch": 9.79,
442
+ "learning_rate": 3.4108796296296296e-05,
443
+ "loss": 0.1831,
444
+ "step": 590
445
+ },
446
+ {
447
+ "epoch": 9.96,
448
+ "learning_rate": 3.4027777777777775e-05,
449
+ "loss": 0.1613,
450
+ "step": 600
451
+ },
452
+ {
453
+ "epoch": 9.99,
454
+ "eval_accuracy": 0.782608695652174,
455
+ "eval_loss": 0.8526802062988281,
456
+ "eval_runtime": 0.6424,
457
+ "eval_samples_per_second": 71.601,
458
+ "eval_steps_per_second": 4.67,
459
+ "step": 602
460
+ },
461
+ {
462
+ "epoch": 10.12,
463
+ "learning_rate": 3.3946759259259254e-05,
464
+ "loss": 0.0869,
465
+ "step": 610
466
+ },
467
+ {
468
+ "epoch": 10.29,
469
+ "learning_rate": 3.3865740740740734e-05,
470
+ "loss": 0.1814,
471
+ "step": 620
472
+ },
473
+ {
474
+ "epoch": 10.46,
475
+ "learning_rate": 3.378472222222222e-05,
476
+ "loss": 0.1488,
477
+ "step": 630
478
+ },
479
+ {
480
+ "epoch": 10.62,
481
+ "learning_rate": 3.37037037037037e-05,
482
+ "loss": 0.1249,
483
+ "step": 640
484
+ },
485
+ {
486
+ "epoch": 10.79,
487
+ "learning_rate": 3.3622685185185185e-05,
488
+ "loss": 0.1397,
489
+ "step": 650
490
+ },
491
+ {
492
+ "epoch": 10.95,
493
+ "learning_rate": 3.3541666666666664e-05,
494
+ "loss": 0.0946,
495
+ "step": 660
496
+ },
497
+ {
498
+ "epoch": 10.99,
499
+ "eval_accuracy": 0.8043478260869565,
500
+ "eval_loss": 0.8274199366569519,
501
+ "eval_runtime": 0.8621,
502
+ "eval_samples_per_second": 53.358,
503
+ "eval_steps_per_second": 3.48,
504
+ "step": 662
505
+ },
506
+ {
507
+ "epoch": 11.12,
508
+ "learning_rate": 3.3460648148148144e-05,
509
+ "loss": 0.1215,
510
+ "step": 670
511
+ },
512
+ {
513
+ "epoch": 11.29,
514
+ "learning_rate": 3.337962962962963e-05,
515
+ "loss": 0.1146,
516
+ "step": 680
517
+ },
518
+ {
519
+ "epoch": 11.45,
520
+ "learning_rate": 3.329861111111111e-05,
521
+ "loss": 0.0991,
522
+ "step": 690
523
+ },
524
+ {
525
+ "epoch": 11.62,
526
+ "learning_rate": 3.321759259259259e-05,
527
+ "loss": 0.1463,
528
+ "step": 700
529
+ },
530
+ {
531
+ "epoch": 11.78,
532
+ "learning_rate": 3.3136574074074074e-05,
533
+ "loss": 0.1049,
534
+ "step": 710
535
+ },
536
+ {
537
+ "epoch": 11.95,
538
+ "learning_rate": 3.3055555555555553e-05,
539
+ "loss": 0.1392,
540
+ "step": 720
541
+ },
542
+ {
543
+ "epoch": 12.0,
544
+ "eval_accuracy": 0.7608695652173914,
545
+ "eval_loss": 0.8311833739280701,
546
+ "eval_runtime": 0.6987,
547
+ "eval_samples_per_second": 65.839,
548
+ "eval_steps_per_second": 4.294,
549
+ "step": 723
550
+ },
551
+ {
552
+ "epoch": 12.12,
553
+ "learning_rate": 3.297453703703703e-05,
554
+ "loss": 0.1116,
555
+ "step": 730
556
+ },
557
+ {
558
+ "epoch": 12.28,
559
+ "learning_rate": 3.289351851851851e-05,
560
+ "loss": 0.1217,
561
+ "step": 740
562
+ },
563
+ {
564
+ "epoch": 12.45,
565
+ "learning_rate": 3.28125e-05,
566
+ "loss": 0.0952,
567
+ "step": 750
568
+ },
569
+ {
570
+ "epoch": 12.61,
571
+ "learning_rate": 3.273148148148148e-05,
572
+ "loss": 0.0972,
573
+ "step": 760
574
+ },
575
+ {
576
+ "epoch": 12.78,
577
+ "learning_rate": 3.2650462962962956e-05,
578
+ "loss": 0.1193,
579
+ "step": 770
580
+ },
581
+ {
582
+ "epoch": 12.95,
583
+ "learning_rate": 3.256944444444444e-05,
584
+ "loss": 0.1028,
585
+ "step": 780
586
+ },
587
+ {
588
+ "epoch": 13.0,
589
+ "eval_accuracy": 0.7608695652173914,
590
+ "eval_loss": 1.1959415674209595,
591
+ "eval_runtime": 0.6697,
592
+ "eval_samples_per_second": 68.691,
593
+ "eval_steps_per_second": 4.48,
594
+ "step": 783
595
+ },
596
+ {
597
+ "epoch": 13.11,
598
+ "learning_rate": 3.248842592592592e-05,
599
+ "loss": 0.1014,
600
+ "step": 790
601
+ },
602
+ {
603
+ "epoch": 13.28,
604
+ "learning_rate": 3.240740740740741e-05,
605
+ "loss": 0.1055,
606
+ "step": 800
607
+ },
608
+ {
609
+ "epoch": 13.44,
610
+ "learning_rate": 3.232638888888889e-05,
611
+ "loss": 0.1317,
612
+ "step": 810
613
+ },
614
+ {
615
+ "epoch": 13.61,
616
+ "learning_rate": 3.2245370370370366e-05,
617
+ "loss": 0.0899,
618
+ "step": 820
619
+ },
620
+ {
621
+ "epoch": 13.78,
622
+ "learning_rate": 3.216435185185185e-05,
623
+ "loss": 0.1158,
624
+ "step": 830
625
+ },
626
+ {
627
+ "epoch": 13.94,
628
+ "learning_rate": 3.208333333333333e-05,
629
+ "loss": 0.1072,
630
+ "step": 840
631
+ },
632
+ {
633
+ "epoch": 13.99,
634
+ "eval_accuracy": 0.7391304347826086,
635
+ "eval_loss": 1.0017004013061523,
636
+ "eval_runtime": 0.6881,
637
+ "eval_samples_per_second": 66.851,
638
+ "eval_steps_per_second": 4.36,
639
+ "step": 843
640
+ },
641
+ {
642
+ "epoch": 14.11,
643
+ "learning_rate": 3.200231481481481e-05,
644
+ "loss": 0.0812,
645
+ "step": 850
646
+ },
647
+ {
648
+ "epoch": 14.27,
649
+ "learning_rate": 3.19212962962963e-05,
650
+ "loss": 0.1025,
651
+ "step": 860
652
+ },
653
+ {
654
+ "epoch": 14.44,
655
+ "learning_rate": 3.1840277777777776e-05,
656
+ "loss": 0.1296,
657
+ "step": 870
658
+ },
659
+ {
660
+ "epoch": 14.61,
661
+ "learning_rate": 3.1759259259259255e-05,
662
+ "loss": 0.146,
663
+ "step": 880
664
+ },
665
+ {
666
+ "epoch": 14.77,
667
+ "learning_rate": 3.1678240740740735e-05,
668
+ "loss": 0.0782,
669
+ "step": 890
670
+ },
671
+ {
672
+ "epoch": 14.94,
673
+ "learning_rate": 3.159722222222222e-05,
674
+ "loss": 0.0888,
675
+ "step": 900
676
+ },
677
+ {
678
+ "epoch": 14.99,
679
+ "eval_accuracy": 0.8043478260869565,
680
+ "eval_loss": 0.921440064907074,
681
+ "eval_runtime": 0.714,
682
+ "eval_samples_per_second": 64.425,
683
+ "eval_steps_per_second": 4.202,
684
+ "step": 903
685
+ },
686
+ {
687
+ "epoch": 15.1,
688
+ "learning_rate": 3.15162037037037e-05,
689
+ "loss": 0.077,
690
+ "step": 910
691
+ },
692
+ {
693
+ "epoch": 15.27,
694
+ "learning_rate": 3.143518518518518e-05,
695
+ "loss": 0.0821,
696
+ "step": 920
697
+ },
698
+ {
699
+ "epoch": 15.44,
700
+ "learning_rate": 3.1354166666666665e-05,
701
+ "loss": 0.0692,
702
+ "step": 930
703
+ },
704
+ {
705
+ "epoch": 15.6,
706
+ "learning_rate": 3.1273148148148145e-05,
707
+ "loss": 0.0801,
708
+ "step": 940
709
+ },
710
+ {
711
+ "epoch": 15.77,
712
+ "learning_rate": 3.119212962962963e-05,
713
+ "loss": 0.1015,
714
+ "step": 950
715
+ },
716
+ {
717
+ "epoch": 15.93,
718
+ "learning_rate": 3.111111111111111e-05,
719
+ "loss": 0.0951,
720
+ "step": 960
721
+ },
722
+ {
723
+ "epoch": 16.0,
724
+ "eval_accuracy": 0.7608695652173914,
725
+ "eval_loss": 0.9156137704849243,
726
+ "eval_runtime": 0.6762,
727
+ "eval_samples_per_second": 68.028,
728
+ "eval_steps_per_second": 4.437,
729
+ "step": 964
730
+ },
731
+ {
732
+ "epoch": 16.1,
733
+ "learning_rate": 3.103009259259259e-05,
734
+ "loss": 0.0768,
735
+ "step": 970
736
+ },
737
+ {
738
+ "epoch": 16.27,
739
+ "learning_rate": 3.0949074074074075e-05,
740
+ "loss": 0.0494,
741
+ "step": 980
742
+ },
743
+ {
744
+ "epoch": 16.43,
745
+ "learning_rate": 3.0868055555555554e-05,
746
+ "loss": 0.1444,
747
+ "step": 990
748
+ },
749
+ {
750
+ "epoch": 16.6,
751
+ "learning_rate": 3.0787037037037034e-05,
752
+ "loss": 0.086,
753
+ "step": 1000
754
+ },
755
+ {
756
+ "epoch": 16.76,
757
+ "learning_rate": 3.070601851851851e-05,
758
+ "loss": 0.0883,
759
+ "step": 1010
760
+ },
761
+ {
762
+ "epoch": 16.93,
763
+ "learning_rate": 3.0625e-05,
764
+ "loss": 0.0714,
765
+ "step": 1020
766
+ },
767
+ {
768
+ "epoch": 17.0,
769
+ "eval_accuracy": 0.6956521739130435,
770
+ "eval_loss": 1.3116263151168823,
771
+ "eval_runtime": 0.6431,
772
+ "eval_samples_per_second": 71.525,
773
+ "eval_steps_per_second": 4.665,
774
+ "step": 1024
775
+ },
776
+ {
777
+ "epoch": 17.1,
778
+ "learning_rate": 3.054398148148148e-05,
779
+ "loss": 0.0637,
780
+ "step": 1030
781
+ },
782
+ {
783
+ "epoch": 17.26,
784
+ "learning_rate": 3.046296296296296e-05,
785
+ "loss": 0.0581,
786
+ "step": 1040
787
+ },
788
+ {
789
+ "epoch": 17.43,
790
+ "learning_rate": 3.0381944444444444e-05,
791
+ "loss": 0.1138,
792
+ "step": 1050
793
+ },
794
+ {
795
+ "epoch": 17.59,
796
+ "learning_rate": 3.0300925925925923e-05,
797
+ "loss": 0.0894,
798
+ "step": 1060
799
+ },
800
+ {
801
+ "epoch": 17.76,
802
+ "learning_rate": 3.0219907407407405e-05,
803
+ "loss": 0.065,
804
+ "step": 1070
805
+ },
806
+ {
807
+ "epoch": 17.93,
808
+ "learning_rate": 3.0138888888888888e-05,
809
+ "loss": 0.0804,
810
+ "step": 1080
811
+ },
812
+ {
813
+ "epoch": 17.99,
814
+ "eval_accuracy": 0.782608695652174,
815
+ "eval_loss": 1.1107388734817505,
816
+ "eval_runtime": 0.7361,
817
+ "eval_samples_per_second": 62.487,
818
+ "eval_steps_per_second": 4.075,
819
+ "step": 1084
820
+ },
821
+ {
822
+ "epoch": 18.09,
823
+ "learning_rate": 3.0057870370370367e-05,
824
+ "loss": 0.1195,
825
+ "step": 1090
826
+ },
827
+ {
828
+ "epoch": 18.26,
829
+ "learning_rate": 2.997685185185185e-05,
830
+ "loss": 0.117,
831
+ "step": 1100
832
+ },
833
+ {
834
+ "epoch": 18.42,
835
+ "learning_rate": 2.989583333333333e-05,
836
+ "loss": 0.0535,
837
+ "step": 1110
838
+ },
839
+ {
840
+ "epoch": 18.59,
841
+ "learning_rate": 2.9814814814814812e-05,
842
+ "loss": 0.0778,
843
+ "step": 1120
844
+ },
845
+ {
846
+ "epoch": 18.76,
847
+ "learning_rate": 2.9733796296296295e-05,
848
+ "loss": 0.1094,
849
+ "step": 1130
850
+ },
851
+ {
852
+ "epoch": 18.92,
853
+ "learning_rate": 2.9652777777777774e-05,
854
+ "loss": 0.08,
855
+ "step": 1140
856
+ },
857
+ {
858
+ "epoch": 18.99,
859
+ "eval_accuracy": 0.8478260869565217,
860
+ "eval_loss": 0.8104602694511414,
861
+ "eval_runtime": 0.9009,
862
+ "eval_samples_per_second": 51.063,
863
+ "eval_steps_per_second": 3.33,
864
+ "step": 1144
865
+ },
866
+ {
867
+ "epoch": 19.09,
868
+ "learning_rate": 2.9571759259259257e-05,
869
+ "loss": 0.058,
870
+ "step": 1150
871
+ },
872
+ {
873
+ "epoch": 19.25,
874
+ "learning_rate": 2.9490740740740736e-05,
875
+ "loss": 0.0811,
876
+ "step": 1160
877
+ },
878
+ {
879
+ "epoch": 19.42,
880
+ "learning_rate": 2.940972222222222e-05,
881
+ "loss": 0.0626,
882
+ "step": 1170
883
+ },
884
+ {
885
+ "epoch": 19.59,
886
+ "learning_rate": 2.9328703703703698e-05,
887
+ "loss": 0.0502,
888
+ "step": 1180
889
+ },
890
+ {
891
+ "epoch": 19.75,
892
+ "learning_rate": 2.9247685185185184e-05,
893
+ "loss": 0.0799,
894
+ "step": 1190
895
+ },
896
+ {
897
+ "epoch": 19.92,
898
+ "learning_rate": 2.9166666666666666e-05,
899
+ "loss": 0.1619,
900
+ "step": 1200
901
+ },
902
+ {
903
+ "epoch": 20.0,
904
+ "eval_accuracy": 0.8260869565217391,
905
+ "eval_loss": 0.7581048011779785,
906
+ "eval_runtime": 0.6451,
907
+ "eval_samples_per_second": 71.304,
908
+ "eval_steps_per_second": 4.65,
909
+ "step": 1205
910
+ },
911
+ {
912
+ "epoch": 20.08,
913
+ "learning_rate": 2.9085648148148146e-05,
914
+ "loss": 0.0787,
915
+ "step": 1210
916
+ },
917
+ {
918
+ "epoch": 20.25,
919
+ "learning_rate": 2.9004629629629628e-05,
920
+ "loss": 0.0764,
921
+ "step": 1220
922
+ },
923
+ {
924
+ "epoch": 20.41,
925
+ "learning_rate": 2.8923611111111108e-05,
926
+ "loss": 0.0618,
927
+ "step": 1230
928
+ },
929
+ {
930
+ "epoch": 20.58,
931
+ "learning_rate": 2.884259259259259e-05,
932
+ "loss": 0.0612,
933
+ "step": 1240
934
+ },
935
+ {
936
+ "epoch": 20.75,
937
+ "learning_rate": 2.8761574074074073e-05,
938
+ "loss": 0.0666,
939
+ "step": 1250
940
+ },
941
+ {
942
+ "epoch": 20.91,
943
+ "learning_rate": 2.8680555555555552e-05,
944
+ "loss": 0.084,
945
+ "step": 1260
946
+ },
947
+ {
948
+ "epoch": 21.0,
949
+ "eval_accuracy": 0.8260869565217391,
950
+ "eval_loss": 1.0210282802581787,
951
+ "eval_runtime": 0.6611,
952
+ "eval_samples_per_second": 69.584,
953
+ "eval_steps_per_second": 4.538,
954
+ "step": 1265
955
+ },
956
+ {
957
+ "epoch": 21.08,
958
+ "learning_rate": 2.8599537037037035e-05,
959
+ "loss": 0.068,
960
+ "step": 1270
961
+ },
962
+ {
963
+ "epoch": 21.24,
964
+ "learning_rate": 2.8518518518518514e-05,
965
+ "loss": 0.0447,
966
+ "step": 1280
967
+ },
968
+ {
969
+ "epoch": 21.41,
970
+ "learning_rate": 2.8437499999999997e-05,
971
+ "loss": 0.0568,
972
+ "step": 1290
973
+ },
974
+ {
975
+ "epoch": 21.58,
976
+ "learning_rate": 2.835648148148148e-05,
977
+ "loss": 0.0823,
978
+ "step": 1300
979
+ },
980
+ {
981
+ "epoch": 21.74,
982
+ "learning_rate": 2.827546296296296e-05,
983
+ "loss": 0.0473,
984
+ "step": 1310
985
+ },
986
+ {
987
+ "epoch": 21.91,
988
+ "learning_rate": 2.819444444444444e-05,
989
+ "loss": 0.072,
990
+ "step": 1320
991
+ },
992
+ {
993
+ "epoch": 21.99,
994
+ "eval_accuracy": 0.7608695652173914,
995
+ "eval_loss": 1.3092247247695923,
996
+ "eval_runtime": 0.6707,
997
+ "eval_samples_per_second": 68.585,
998
+ "eval_steps_per_second": 4.473,
999
+ "step": 1325
1000
+ },
1001
+ {
1002
+ "epoch": 22.07,
1003
+ "learning_rate": 2.811342592592592e-05,
1004
+ "loss": 0.0629,
1005
+ "step": 1330
1006
+ },
1007
+ {
1008
+ "epoch": 22.24,
1009
+ "learning_rate": 2.8032407407407407e-05,
1010
+ "loss": 0.0908,
1011
+ "step": 1340
1012
+ },
1013
+ {
1014
+ "epoch": 22.41,
1015
+ "learning_rate": 2.795138888888889e-05,
1016
+ "loss": 0.0744,
1017
+ "step": 1350
1018
+ },
1019
+ {
1020
+ "epoch": 22.57,
1021
+ "learning_rate": 2.787037037037037e-05,
1022
+ "loss": 0.055,
1023
+ "step": 1360
1024
+ },
1025
+ {
1026
+ "epoch": 22.74,
1027
+ "learning_rate": 2.778935185185185e-05,
1028
+ "loss": 0.0572,
1029
+ "step": 1370
1030
+ },
1031
+ {
1032
+ "epoch": 22.9,
1033
+ "learning_rate": 2.770833333333333e-05,
1034
+ "loss": 0.0303,
1035
+ "step": 1380
1036
+ },
1037
+ {
1038
+ "epoch": 22.99,
1039
+ "eval_accuracy": 0.782608695652174,
1040
+ "eval_loss": 1.3367185592651367,
1041
+ "eval_runtime": 0.665,
1042
+ "eval_samples_per_second": 69.17,
1043
+ "eval_steps_per_second": 4.511,
1044
+ "step": 1385
1045
+ },
1046
+ {
1047
+ "epoch": 23.07,
1048
+ "learning_rate": 2.7627314814814813e-05,
1049
+ "loss": 0.0785,
1050
+ "step": 1390
1051
+ },
1052
+ {
1053
+ "epoch": 23.24,
1054
+ "learning_rate": 2.7546296296296296e-05,
1055
+ "loss": 0.0877,
1056
+ "step": 1400
1057
+ },
1058
+ {
1059
+ "epoch": 23.4,
1060
+ "learning_rate": 2.7465277777777775e-05,
1061
+ "loss": 0.0482,
1062
+ "step": 1410
1063
+ },
1064
+ {
1065
+ "epoch": 23.57,
1066
+ "learning_rate": 2.7384259259259258e-05,
1067
+ "loss": 0.0712,
1068
+ "step": 1420
1069
+ },
1070
+ {
1071
+ "epoch": 23.73,
1072
+ "learning_rate": 2.7303240740740737e-05,
1073
+ "loss": 0.0381,
1074
+ "step": 1430
1075
+ },
1076
+ {
1077
+ "epoch": 23.9,
1078
+ "learning_rate": 2.722222222222222e-05,
1079
+ "loss": 0.0228,
1080
+ "step": 1440
1081
+ },
1082
+ {
1083
+ "epoch": 24.0,
1084
+ "eval_accuracy": 0.8260869565217391,
1085
+ "eval_loss": 1.0276718139648438,
1086
+ "eval_runtime": 0.7532,
1087
+ "eval_samples_per_second": 61.074,
1088
+ "eval_steps_per_second": 3.983,
1089
+ "step": 1446
1090
+ },
1091
+ {
1092
+ "epoch": 24.07,
1093
+ "learning_rate": 2.71412037037037e-05,
1094
+ "loss": 0.0696,
1095
+ "step": 1450
1096
+ },
1097
+ {
1098
+ "epoch": 24.23,
1099
+ "learning_rate": 2.706018518518518e-05,
1100
+ "loss": 0.0612,
1101
+ "step": 1460
1102
+ },
1103
+ {
1104
+ "epoch": 24.4,
1105
+ "learning_rate": 2.6979166666666664e-05,
1106
+ "loss": 0.0549,
1107
+ "step": 1470
1108
+ },
1109
+ {
1110
+ "epoch": 24.56,
1111
+ "learning_rate": 2.6898148148148143e-05,
1112
+ "loss": 0.0253,
1113
+ "step": 1480
1114
+ },
1115
+ {
1116
+ "epoch": 24.73,
1117
+ "learning_rate": 2.681712962962963e-05,
1118
+ "loss": 0.0472,
1119
+ "step": 1490
1120
+ },
1121
+ {
1122
+ "epoch": 24.9,
1123
+ "learning_rate": 2.673611111111111e-05,
1124
+ "loss": 0.0755,
1125
+ "step": 1500
1126
+ },
1127
+ {
1128
+ "epoch": 25.0,
1129
+ "eval_accuracy": 0.8260869565217391,
1130
+ "eval_loss": 0.9436376094818115,
1131
+ "eval_runtime": 0.6656,
1132
+ "eval_samples_per_second": 69.111,
1133
+ "eval_steps_per_second": 4.507,
1134
+ "step": 1506
1135
+ },
1136
+ {
1137
+ "epoch": 25.06,
1138
+ "learning_rate": 2.665509259259259e-05,
1139
+ "loss": 0.0743,
1140
+ "step": 1510
1141
+ },
1142
+ {
1143
+ "epoch": 25.23,
1144
+ "learning_rate": 2.6574074074074074e-05,
1145
+ "loss": 0.045,
1146
+ "step": 1520
1147
+ },
1148
+ {
1149
+ "epoch": 25.39,
1150
+ "learning_rate": 2.6493055555555553e-05,
1151
+ "loss": 0.0669,
1152
+ "step": 1530
1153
+ },
1154
+ {
1155
+ "epoch": 25.56,
1156
+ "learning_rate": 2.6412037037037036e-05,
1157
+ "loss": 0.0551,
1158
+ "step": 1540
1159
+ },
1160
+ {
1161
+ "epoch": 25.73,
1162
+ "learning_rate": 2.6331018518518515e-05,
1163
+ "loss": 0.0256,
1164
+ "step": 1550
1165
+ },
1166
+ {
1167
+ "epoch": 25.89,
1168
+ "learning_rate": 2.6249999999999998e-05,
1169
+ "loss": 0.0756,
1170
+ "step": 1560
1171
+ },
1172
+ {
1173
+ "epoch": 25.99,
1174
+ "eval_accuracy": 0.7608695652173914,
1175
+ "eval_loss": 1.1588078737258911,
1176
+ "eval_runtime": 0.6572,
1177
+ "eval_samples_per_second": 69.997,
1178
+ "eval_steps_per_second": 4.565,
1179
+ "step": 1566
1180
+ },
1181
+ {
1182
+ "epoch": 26.06,
1183
+ "learning_rate": 2.616898148148148e-05,
1184
+ "loss": 0.0403,
1185
+ "step": 1570
1186
+ },
1187
+ {
1188
+ "epoch": 26.22,
1189
+ "learning_rate": 2.608796296296296e-05,
1190
+ "loss": 0.0486,
1191
+ "step": 1580
1192
+ },
1193
+ {
1194
+ "epoch": 26.39,
1195
+ "learning_rate": 2.6006944444444442e-05,
1196
+ "loss": 0.0489,
1197
+ "step": 1590
1198
+ },
1199
+ {
1200
+ "epoch": 26.56,
1201
+ "learning_rate": 2.592592592592592e-05,
1202
+ "loss": 0.0646,
1203
+ "step": 1600
1204
+ },
1205
+ {
1206
+ "epoch": 26.72,
1207
+ "learning_rate": 2.5844907407407404e-05,
1208
+ "loss": 0.0713,
1209
+ "step": 1610
1210
+ },
1211
+ {
1212
+ "epoch": 26.89,
1213
+ "learning_rate": 2.5763888888888887e-05,
1214
+ "loss": 0.0875,
1215
+ "step": 1620
1216
+ },
1217
+ {
1218
+ "epoch": 26.99,
1219
+ "eval_accuracy": 0.717391304347826,
1220
+ "eval_loss": 1.3280256986618042,
1221
+ "eval_runtime": 0.7112,
1222
+ "eval_samples_per_second": 64.681,
1223
+ "eval_steps_per_second": 4.218,
1224
+ "step": 1626
1225
+ },
1226
+ {
1227
+ "epoch": 27.05,
1228
+ "learning_rate": 2.5682870370370366e-05,
1229
+ "loss": 0.0704,
1230
+ "step": 1630
1231
+ },
1232
+ {
1233
+ "epoch": 27.22,
1234
+ "learning_rate": 2.5601851851851852e-05,
1235
+ "loss": 0.0606,
1236
+ "step": 1640
1237
+ },
1238
+ {
1239
+ "epoch": 27.39,
1240
+ "learning_rate": 2.552083333333333e-05,
1241
+ "loss": 0.0385,
1242
+ "step": 1650
1243
+ },
1244
+ {
1245
+ "epoch": 27.55,
1246
+ "learning_rate": 2.5439814814814814e-05,
1247
+ "loss": 0.0494,
1248
+ "step": 1660
1249
+ },
1250
+ {
1251
+ "epoch": 27.72,
1252
+ "learning_rate": 2.5358796296296297e-05,
1253
+ "loss": 0.0171,
1254
+ "step": 1670
1255
+ },
1256
+ {
1257
+ "epoch": 27.88,
1258
+ "learning_rate": 2.5277777777777776e-05,
1259
+ "loss": 0.0771,
1260
+ "step": 1680
1261
+ },
1262
+ {
1263
+ "epoch": 28.0,
1264
+ "eval_accuracy": 0.6739130434782609,
1265
+ "eval_loss": 1.8558253049850464,
1266
+ "eval_runtime": 0.6528,
1267
+ "eval_samples_per_second": 70.468,
1268
+ "eval_steps_per_second": 4.596,
1269
+ "step": 1687
1270
+ },
1271
+ {
1272
+ "epoch": 28.05,
1273
+ "learning_rate": 2.519675925925926e-05,
1274
+ "loss": 0.0447,
1275
+ "step": 1690
1276
+ },
1277
+ {
1278
+ "epoch": 28.22,
1279
+ "learning_rate": 2.5115740740740738e-05,
1280
+ "loss": 0.0365,
1281
+ "step": 1700
1282
+ },
1283
+ {
1284
+ "epoch": 28.38,
1285
+ "learning_rate": 2.503472222222222e-05,
1286
+ "loss": 0.0557,
1287
+ "step": 1710
1288
+ },
1289
+ {
1290
+ "epoch": 28.55,
1291
+ "learning_rate": 2.49537037037037e-05,
1292
+ "loss": 0.0447,
1293
+ "step": 1720
1294
+ },
1295
+ {
1296
+ "epoch": 28.71,
1297
+ "learning_rate": 2.4872685185185182e-05,
1298
+ "loss": 0.0412,
1299
+ "step": 1730
1300
+ },
1301
+ {
1302
+ "epoch": 28.88,
1303
+ "learning_rate": 2.4791666666666665e-05,
1304
+ "loss": 0.0467,
1305
+ "step": 1740
1306
+ },
1307
+ {
1308
+ "epoch": 29.0,
1309
+ "eval_accuracy": 0.7391304347826086,
1310
+ "eval_loss": 1.6476134061813354,
1311
+ "eval_runtime": 0.6892,
1312
+ "eval_samples_per_second": 66.748,
1313
+ "eval_steps_per_second": 4.353,
1314
+ "step": 1747
1315
+ },
1316
+ {
1317
+ "epoch": 29.05,
1318
+ "learning_rate": 2.4710648148148144e-05,
1319
+ "loss": 0.0483,
1320
+ "step": 1750
1321
+ },
1322
+ {
1323
+ "epoch": 29.21,
1324
+ "learning_rate": 2.4629629629629627e-05,
1325
+ "loss": 0.1003,
1326
+ "step": 1760
1327
+ },
1328
+ {
1329
+ "epoch": 29.38,
1330
+ "learning_rate": 2.4548611111111106e-05,
1331
+ "loss": 0.0879,
1332
+ "step": 1770
1333
+ },
1334
+ {
1335
+ "epoch": 29.54,
1336
+ "learning_rate": 2.446759259259259e-05,
1337
+ "loss": 0.0333,
1338
+ "step": 1780
1339
+ },
1340
+ {
1341
+ "epoch": 29.71,
1342
+ "learning_rate": 2.4386574074074075e-05,
1343
+ "loss": 0.0849,
1344
+ "step": 1790
1345
+ },
1346
+ {
1347
+ "epoch": 29.88,
1348
+ "learning_rate": 2.4305555555555554e-05,
1349
+ "loss": 0.0382,
1350
+ "step": 1800
1351
+ },
1352
+ {
1353
+ "epoch": 29.99,
1354
+ "eval_accuracy": 0.8478260869565217,
1355
+ "eval_loss": 0.9374191164970398,
1356
+ "eval_runtime": 0.7092,
1357
+ "eval_samples_per_second": 64.864,
1358
+ "eval_steps_per_second": 4.23,
1359
+ "step": 1807
1360
+ },
1361
+ {
1362
+ "epoch": 30.04,
1363
+ "learning_rate": 2.4224537037037037e-05,
1364
+ "loss": 0.0482,
1365
+ "step": 1810
1366
+ },
1367
+ {
1368
+ "epoch": 30.21,
1369
+ "learning_rate": 2.4143518518518516e-05,
1370
+ "loss": 0.0232,
1371
+ "step": 1820
1372
+ },
1373
+ {
1374
+ "epoch": 30.37,
1375
+ "learning_rate": 2.40625e-05,
1376
+ "loss": 0.0342,
1377
+ "step": 1830
1378
+ },
1379
+ {
1380
+ "epoch": 30.54,
1381
+ "learning_rate": 2.398148148148148e-05,
1382
+ "loss": 0.0344,
1383
+ "step": 1840
1384
+ },
1385
+ {
1386
+ "epoch": 30.71,
1387
+ "learning_rate": 2.390046296296296e-05,
1388
+ "loss": 0.0325,
1389
+ "step": 1850
1390
+ },
1391
+ {
1392
+ "epoch": 30.87,
1393
+ "learning_rate": 2.3819444444444443e-05,
1394
+ "loss": 0.0511,
1395
+ "step": 1860
1396
+ },
1397
+ {
1398
+ "epoch": 30.99,
1399
+ "eval_accuracy": 0.8043478260869565,
1400
+ "eval_loss": 1.0847010612487793,
1401
+ "eval_runtime": 0.7072,
1402
+ "eval_samples_per_second": 65.049,
1403
+ "eval_steps_per_second": 4.242,
1404
+ "step": 1867
1405
+ },
1406
+ {
1407
+ "epoch": 31.04,
1408
+ "learning_rate": 2.3738425925925923e-05,
1409
+ "loss": 0.0243,
1410
+ "step": 1870
1411
+ },
1412
+ {
1413
+ "epoch": 31.2,
1414
+ "learning_rate": 2.3657407407407405e-05,
1415
+ "loss": 0.0466,
1416
+ "step": 1880
1417
+ },
1418
+ {
1419
+ "epoch": 31.37,
1420
+ "learning_rate": 2.3576388888888888e-05,
1421
+ "loss": 0.0543,
1422
+ "step": 1890
1423
+ },
1424
+ {
1425
+ "epoch": 31.54,
1426
+ "learning_rate": 2.3495370370370367e-05,
1427
+ "loss": 0.0415,
1428
+ "step": 1900
1429
+ },
1430
+ {
1431
+ "epoch": 31.7,
1432
+ "learning_rate": 2.341435185185185e-05,
1433
+ "loss": 0.0327,
1434
+ "step": 1910
1435
+ },
1436
+ {
1437
+ "epoch": 31.87,
1438
+ "learning_rate": 2.333333333333333e-05,
1439
+ "loss": 0.0161,
1440
+ "step": 1920
1441
+ },
1442
+ {
1443
+ "epoch": 32.0,
1444
+ "eval_accuracy": 0.782608695652174,
1445
+ "eval_loss": 1.2028273344039917,
1446
+ "eval_runtime": 0.6812,
1447
+ "eval_samples_per_second": 67.531,
1448
+ "eval_steps_per_second": 4.404,
1449
+ "step": 1928
1450
+ },
1451
+ {
1452
+ "epoch": 32.03,
1453
+ "learning_rate": 2.325231481481481e-05,
1454
+ "loss": 0.0372,
1455
+ "step": 1930
1456
+ },
1457
+ {
1458
+ "epoch": 32.2,
1459
+ "learning_rate": 2.3171296296296298e-05,
1460
+ "loss": 0.0433,
1461
+ "step": 1940
1462
+ },
1463
+ {
1464
+ "epoch": 32.37,
1465
+ "learning_rate": 2.3090277777777777e-05,
1466
+ "loss": 0.0536,
1467
+ "step": 1950
1468
+ },
1469
+ {
1470
+ "epoch": 32.53,
1471
+ "learning_rate": 2.300925925925926e-05,
1472
+ "loss": 0.0497,
1473
+ "step": 1960
1474
+ },
1475
+ {
1476
+ "epoch": 32.7,
1477
+ "learning_rate": 2.292824074074074e-05,
1478
+ "loss": 0.0317,
1479
+ "step": 1970
1480
+ },
1481
+ {
1482
+ "epoch": 32.86,
1483
+ "learning_rate": 2.284722222222222e-05,
1484
+ "loss": 0.0301,
1485
+ "step": 1980
1486
+ },
1487
+ {
1488
+ "epoch": 33.0,
1489
+ "eval_accuracy": 0.7391304347826086,
1490
+ "eval_loss": 1.297067403793335,
1491
+ "eval_runtime": 0.6982,
1492
+ "eval_samples_per_second": 65.888,
1493
+ "eval_steps_per_second": 4.297,
1494
+ "step": 1988
1495
+ },
1496
+ {
1497
+ "epoch": 33.03,
1498
+ "learning_rate": 2.27662037037037e-05,
1499
+ "loss": 0.0488,
1500
+ "step": 1990
1501
+ },
1502
+ {
1503
+ "epoch": 33.2,
1504
+ "learning_rate": 2.2685185185185183e-05,
1505
+ "loss": 0.0382,
1506
+ "step": 2000
1507
+ },
1508
+ {
1509
+ "epoch": 33.36,
1510
+ "learning_rate": 2.2604166666666666e-05,
1511
+ "loss": 0.0576,
1512
+ "step": 2010
1513
+ },
1514
+ {
1515
+ "epoch": 33.53,
1516
+ "learning_rate": 2.2523148148148145e-05,
1517
+ "loss": 0.0111,
1518
+ "step": 2020
1519
+ },
1520
+ {
1521
+ "epoch": 33.69,
1522
+ "learning_rate": 2.2442129629629628e-05,
1523
+ "loss": 0.0685,
1524
+ "step": 2030
1525
+ },
1526
+ {
1527
+ "epoch": 33.86,
1528
+ "learning_rate": 2.2361111111111107e-05,
1529
+ "loss": 0.0443,
1530
+ "step": 2040
1531
+ },
1532
+ {
1533
+ "epoch": 33.99,
1534
+ "eval_accuracy": 0.717391304347826,
1535
+ "eval_loss": 1.3993148803710938,
1536
+ "eval_runtime": 0.6712,
1537
+ "eval_samples_per_second": 68.537,
1538
+ "eval_steps_per_second": 4.47,
1539
+ "step": 2048
1540
+ },
1541
+ {
1542
+ "epoch": 34.02,
1543
+ "learning_rate": 2.228009259259259e-05,
1544
+ "loss": 0.0488,
1545
+ "step": 2050
1546
+ },
1547
+ {
1548
+ "epoch": 34.19,
1549
+ "learning_rate": 2.2199074074074073e-05,
1550
+ "loss": 0.0322,
1551
+ "step": 2060
1552
+ },
1553
+ {
1554
+ "epoch": 34.36,
1555
+ "learning_rate": 2.2118055555555552e-05,
1556
+ "loss": 0.0353,
1557
+ "step": 2070
1558
+ },
1559
+ {
1560
+ "epoch": 34.52,
1561
+ "learning_rate": 2.2037037037037034e-05,
1562
+ "loss": 0.0387,
1563
+ "step": 2080
1564
+ },
1565
+ {
1566
+ "epoch": 34.69,
1567
+ "learning_rate": 2.1956018518518514e-05,
1568
+ "loss": 0.086,
1569
+ "step": 2090
1570
+ },
1571
+ {
1572
+ "epoch": 34.85,
1573
+ "learning_rate": 2.1874999999999996e-05,
1574
+ "loss": 0.0782,
1575
+ "step": 2100
1576
+ },
1577
+ {
1578
+ "epoch": 34.99,
1579
+ "eval_accuracy": 0.8043478260869565,
1580
+ "eval_loss": 1.3358551263809204,
1581
+ "eval_runtime": 0.6692,
1582
+ "eval_samples_per_second": 68.742,
1583
+ "eval_steps_per_second": 4.483,
1584
+ "step": 2108
1585
+ },
1586
+ {
1587
+ "epoch": 35.02,
1588
+ "learning_rate": 2.1793981481481482e-05,
1589
+ "loss": 0.0232,
1590
+ "step": 2110
1591
+ },
1592
+ {
1593
+ "epoch": 35.19,
1594
+ "learning_rate": 2.171296296296296e-05,
1595
+ "loss": 0.0295,
1596
+ "step": 2120
1597
+ },
1598
+ {
1599
+ "epoch": 35.35,
1600
+ "learning_rate": 2.1631944444444444e-05,
1601
+ "loss": 0.0456,
1602
+ "step": 2130
1603
+ },
1604
+ {
1605
+ "epoch": 35.52,
1606
+ "learning_rate": 2.1550925925925924e-05,
1607
+ "loss": 0.0266,
1608
+ "step": 2140
1609
+ },
1610
+ {
1611
+ "epoch": 35.68,
1612
+ "learning_rate": 2.1469907407407406e-05,
1613
+ "loss": 0.0462,
1614
+ "step": 2150
1615
+ },
1616
+ {
1617
+ "epoch": 35.85,
1618
+ "learning_rate": 2.138888888888889e-05,
1619
+ "loss": 0.0287,
1620
+ "step": 2160
1621
+ },
1622
+ {
1623
+ "epoch": 36.0,
1624
+ "eval_accuracy": 0.782608695652174,
1625
+ "eval_loss": 1.3011459112167358,
1626
+ "eval_runtime": 0.7002,
1627
+ "eval_samples_per_second": 65.698,
1628
+ "eval_steps_per_second": 4.285,
1629
+ "step": 2169
1630
+ },
1631
+ {
1632
+ "epoch": 36.02,
1633
+ "learning_rate": 2.1307870370370368e-05,
1634
+ "loss": 0.0502,
1635
+ "step": 2170
1636
+ },
1637
+ {
1638
+ "epoch": 36.18,
1639
+ "learning_rate": 2.122685185185185e-05,
1640
+ "loss": 0.0499,
1641
+ "step": 2180
1642
+ },
1643
+ {
1644
+ "epoch": 36.35,
1645
+ "learning_rate": 2.114583333333333e-05,
1646
+ "loss": 0.0313,
1647
+ "step": 2190
1648
+ },
1649
+ {
1650
+ "epoch": 36.51,
1651
+ "learning_rate": 2.1064814814814813e-05,
1652
+ "loss": 0.0467,
1653
+ "step": 2200
1654
+ },
1655
+ {
1656
+ "epoch": 36.68,
1657
+ "learning_rate": 2.0983796296296295e-05,
1658
+ "loss": 0.0496,
1659
+ "step": 2210
1660
+ },
1661
+ {
1662
+ "epoch": 36.85,
1663
+ "learning_rate": 2.0902777777777775e-05,
1664
+ "loss": 0.0347,
1665
+ "step": 2220
1666
+ },
1667
+ {
1668
+ "epoch": 37.0,
1669
+ "eval_accuracy": 0.782608695652174,
1670
+ "eval_loss": 1.2450307607650757,
1671
+ "eval_runtime": 0.7057,
1672
+ "eval_samples_per_second": 65.186,
1673
+ "eval_steps_per_second": 4.251,
1674
+ "step": 2229
1675
+ },
1676
+ {
1677
+ "epoch": 37.01,
1678
+ "learning_rate": 2.0821759259259257e-05,
1679
+ "loss": 0.0322,
1680
+ "step": 2230
1681
+ },
1682
+ {
1683
+ "epoch": 37.18,
1684
+ "learning_rate": 2.0740740740740737e-05,
1685
+ "loss": 0.0213,
1686
+ "step": 2240
1687
+ },
1688
+ {
1689
+ "epoch": 37.34,
1690
+ "learning_rate": 2.065972222222222e-05,
1691
+ "loss": 0.0215,
1692
+ "step": 2250
1693
+ },
1694
+ {
1695
+ "epoch": 37.51,
1696
+ "learning_rate": 2.05787037037037e-05,
1697
+ "loss": 0.0172,
1698
+ "step": 2260
1699
+ },
1700
+ {
1701
+ "epoch": 37.68,
1702
+ "learning_rate": 2.0497685185185184e-05,
1703
+ "loss": 0.0264,
1704
+ "step": 2270
1705
+ },
1706
+ {
1707
+ "epoch": 37.84,
1708
+ "learning_rate": 2.0416666666666667e-05,
1709
+ "loss": 0.0538,
1710
+ "step": 2280
1711
+ },
1712
+ {
1713
+ "epoch": 37.99,
1714
+ "eval_accuracy": 0.7608695652173914,
1715
+ "eval_loss": 1.8215843439102173,
1716
+ "eval_runtime": 0.6952,
1717
+ "eval_samples_per_second": 66.171,
1718
+ "eval_steps_per_second": 4.315,
1719
+ "step": 2289
1720
+ },
1721
+ {
1722
+ "epoch": 38.01,
1723
+ "learning_rate": 2.0335648148148146e-05,
1724
+ "loss": 0.0541,
1725
+ "step": 2290
1726
+ },
1727
+ {
1728
+ "epoch": 38.17,
1729
+ "learning_rate": 2.025462962962963e-05,
1730
+ "loss": 0.0261,
1731
+ "step": 2300
1732
+ },
1733
+ {
1734
+ "epoch": 38.34,
1735
+ "learning_rate": 2.017361111111111e-05,
1736
+ "loss": 0.0371,
1737
+ "step": 2310
1738
+ },
1739
+ {
1740
+ "epoch": 38.51,
1741
+ "learning_rate": 2.009259259259259e-05,
1742
+ "loss": 0.028,
1743
+ "step": 2320
1744
+ },
1745
+ {
1746
+ "epoch": 38.67,
1747
+ "learning_rate": 2.0011574074074074e-05,
1748
+ "loss": 0.0167,
1749
+ "step": 2330
1750
+ },
1751
+ {
1752
+ "epoch": 38.84,
1753
+ "learning_rate": 1.9930555555555553e-05,
1754
+ "loss": 0.027,
1755
+ "step": 2340
1756
+ },
1757
+ {
1758
+ "epoch": 38.99,
1759
+ "eval_accuracy": 0.8043478260869565,
1760
+ "eval_loss": 1.170117735862732,
1761
+ "eval_runtime": 0.7309,
1762
+ "eval_samples_per_second": 62.933,
1763
+ "eval_steps_per_second": 4.104,
1764
+ "step": 2349
1765
+ },
1766
+ {
1767
+ "epoch": 39.0,
1768
+ "learning_rate": 1.9849537037037036e-05,
1769
+ "loss": 0.0375,
1770
+ "step": 2350
1771
+ },
1772
+ {
1773
+ "epoch": 39.17,
1774
+ "learning_rate": 1.9768518518518515e-05,
1775
+ "loss": 0.0467,
1776
+ "step": 2360
1777
+ },
1778
+ {
1779
+ "epoch": 39.34,
1780
+ "learning_rate": 1.9687499999999997e-05,
1781
+ "loss": 0.03,
1782
+ "step": 2370
1783
+ },
1784
+ {
1785
+ "epoch": 39.5,
1786
+ "learning_rate": 1.960648148148148e-05,
1787
+ "loss": 0.0243,
1788
+ "step": 2380
1789
+ },
1790
+ {
1791
+ "epoch": 39.67,
1792
+ "learning_rate": 1.952546296296296e-05,
1793
+ "loss": 0.0393,
1794
+ "step": 2390
1795
+ },
1796
+ {
1797
+ "epoch": 39.83,
1798
+ "learning_rate": 1.9444444444444442e-05,
1799
+ "loss": 0.0284,
1800
+ "step": 2400
1801
+ },
1802
+ {
1803
+ "epoch": 40.0,
1804
+ "learning_rate": 1.936342592592592e-05,
1805
+ "loss": 0.038,
1806
+ "step": 2410
1807
+ },
1808
+ {
1809
+ "epoch": 40.0,
1810
+ "eval_accuracy": 0.8043478260869565,
1811
+ "eval_loss": 1.1025199890136719,
1812
+ "eval_runtime": 0.7243,
1813
+ "eval_samples_per_second": 63.514,
1814
+ "eval_steps_per_second": 4.142,
1815
+ "step": 2410
1816
+ },
1817
+ {
1818
+ "epoch": 40.17,
1819
+ "learning_rate": 1.9282407407407407e-05,
1820
+ "loss": 0.0243,
1821
+ "step": 2420
1822
+ },
1823
+ {
1824
+ "epoch": 40.33,
1825
+ "learning_rate": 1.920138888888889e-05,
1826
+ "loss": 0.049,
1827
+ "step": 2430
1828
+ },
1829
+ {
1830
+ "epoch": 40.5,
1831
+ "learning_rate": 1.912037037037037e-05,
1832
+ "loss": 0.0319,
1833
+ "step": 2440
1834
+ },
1835
+ {
1836
+ "epoch": 40.66,
1837
+ "learning_rate": 1.9039351851851852e-05,
1838
+ "loss": 0.0413,
1839
+ "step": 2450
1840
+ },
1841
+ {
1842
+ "epoch": 40.83,
1843
+ "learning_rate": 1.895833333333333e-05,
1844
+ "loss": 0.0369,
1845
+ "step": 2460
1846
+ },
1847
+ {
1848
+ "epoch": 41.0,
1849
+ "learning_rate": 1.8877314814814814e-05,
1850
+ "loss": 0.0244,
1851
+ "step": 2470
1852
+ },
1853
+ {
1854
+ "epoch": 41.0,
1855
+ "eval_accuracy": 0.7608695652173914,
1856
+ "eval_loss": 1.2912131547927856,
1857
+ "eval_runtime": 0.6786,
1858
+ "eval_samples_per_second": 67.783,
1859
+ "eval_steps_per_second": 4.421,
1860
+ "step": 2470
1861
+ },
1862
+ {
1863
+ "epoch": 41.16,
1864
+ "learning_rate": 1.8796296296296296e-05,
1865
+ "loss": 0.0467,
1866
+ "step": 2480
1867
+ },
1868
+ {
1869
+ "epoch": 41.33,
1870
+ "learning_rate": 1.8715277777777776e-05,
1871
+ "loss": 0.0295,
1872
+ "step": 2490
1873
+ },
1874
+ {
1875
+ "epoch": 41.49,
1876
+ "learning_rate": 1.863425925925926e-05,
1877
+ "loss": 0.0241,
1878
+ "step": 2500
1879
+ },
1880
+ {
1881
+ "epoch": 41.66,
1882
+ "learning_rate": 1.8553240740740738e-05,
1883
+ "loss": 0.0238,
1884
+ "step": 2510
1885
+ },
1886
+ {
1887
+ "epoch": 41.83,
1888
+ "learning_rate": 1.847222222222222e-05,
1889
+ "loss": 0.0201,
1890
+ "step": 2520
1891
+ },
1892
+ {
1893
+ "epoch": 41.99,
1894
+ "learning_rate": 1.83912037037037e-05,
1895
+ "loss": 0.0122,
1896
+ "step": 2530
1897
+ },
1898
+ {
1899
+ "epoch": 41.99,
1900
+ "eval_accuracy": 0.7608695652173914,
1901
+ "eval_loss": 1.5698707103729248,
1902
+ "eval_runtime": 0.7096,
1903
+ "eval_samples_per_second": 64.826,
1904
+ "eval_steps_per_second": 4.228,
1905
+ "step": 2530
1906
+ },
1907
+ {
1908
+ "epoch": 42.16,
1909
+ "learning_rate": 1.8310185185185182e-05,
1910
+ "loss": 0.0321,
1911
+ "step": 2540
1912
+ },
1913
+ {
1914
+ "epoch": 42.32,
1915
+ "learning_rate": 1.8229166666666665e-05,
1916
+ "loss": 0.0336,
1917
+ "step": 2550
1918
+ },
1919
+ {
1920
+ "epoch": 42.49,
1921
+ "learning_rate": 1.8148148148148144e-05,
1922
+ "loss": 0.0143,
1923
+ "step": 2560
1924
+ },
1925
+ {
1926
+ "epoch": 42.66,
1927
+ "learning_rate": 1.806712962962963e-05,
1928
+ "loss": 0.0205,
1929
+ "step": 2570
1930
+ },
1931
+ {
1932
+ "epoch": 42.82,
1933
+ "learning_rate": 1.798611111111111e-05,
1934
+ "loss": 0.0347,
1935
+ "step": 2580
1936
+ },
1937
+ {
1938
+ "epoch": 42.99,
1939
+ "learning_rate": 1.7905092592592592e-05,
1940
+ "loss": 0.023,
1941
+ "step": 2590
1942
+ },
1943
+ {
1944
+ "epoch": 42.99,
1945
+ "eval_accuracy": 0.782608695652174,
1946
+ "eval_loss": 1.51142156124115,
1947
+ "eval_runtime": 0.7048,
1948
+ "eval_samples_per_second": 65.263,
1949
+ "eval_steps_per_second": 4.256,
1950
+ "step": 2590
1951
+ },
1952
+ {
1953
+ "epoch": 43.15,
1954
+ "learning_rate": 1.7824074074074075e-05,
1955
+ "loss": 0.025,
1956
+ "step": 2600
1957
+ },
1958
+ {
1959
+ "epoch": 43.32,
1960
+ "learning_rate": 1.7743055555555554e-05,
1961
+ "loss": 0.0213,
1962
+ "step": 2610
1963
+ },
1964
+ {
1965
+ "epoch": 43.49,
1966
+ "learning_rate": 1.7662037037037037e-05,
1967
+ "loss": 0.0491,
1968
+ "step": 2620
1969
+ },
1970
+ {
1971
+ "epoch": 43.65,
1972
+ "learning_rate": 1.7581018518518516e-05,
1973
+ "loss": 0.0148,
1974
+ "step": 2630
1975
+ },
1976
+ {
1977
+ "epoch": 43.82,
1978
+ "learning_rate": 1.75e-05,
1979
+ "loss": 0.0267,
1980
+ "step": 2640
1981
+ },
1982
+ {
1983
+ "epoch": 43.98,
1984
+ "learning_rate": 1.741898148148148e-05,
1985
+ "loss": 0.0297,
1986
+ "step": 2650
1987
+ },
1988
+ {
1989
+ "epoch": 44.0,
1990
+ "eval_accuracy": 0.8478260869565217,
1991
+ "eval_loss": 1.2188609838485718,
1992
+ "eval_runtime": 0.7237,
1993
+ "eval_samples_per_second": 63.561,
1994
+ "eval_steps_per_second": 4.145,
1995
+ "step": 2651
1996
+ },
1997
+ {
1998
+ "epoch": 44.15,
1999
+ "learning_rate": 1.733796296296296e-05,
2000
+ "loss": 0.0183,
2001
+ "step": 2660
2002
+ },
2003
+ {
2004
+ "epoch": 44.32,
2005
+ "learning_rate": 1.7256944444444443e-05,
2006
+ "loss": 0.0352,
2007
+ "step": 2670
2008
+ },
2009
+ {
2010
+ "epoch": 44.48,
2011
+ "learning_rate": 1.7175925925925926e-05,
2012
+ "loss": 0.0243,
2013
+ "step": 2680
2014
+ },
2015
+ {
2016
+ "epoch": 44.65,
2017
+ "learning_rate": 1.7094907407407405e-05,
2018
+ "loss": 0.104,
2019
+ "step": 2690
2020
+ },
2021
+ {
2022
+ "epoch": 44.81,
2023
+ "learning_rate": 1.7013888888888888e-05,
2024
+ "loss": 0.015,
2025
+ "step": 2700
2026
+ },
2027
+ {
2028
+ "epoch": 44.98,
2029
+ "learning_rate": 1.6932870370370367e-05,
2030
+ "loss": 0.0284,
2031
+ "step": 2710
2032
+ },
2033
+ {
2034
+ "epoch": 45.0,
2035
+ "eval_accuracy": 0.782608695652174,
2036
+ "eval_loss": 1.3997039794921875,
2037
+ "eval_runtime": 0.7121,
2038
+ "eval_samples_per_second": 64.595,
2039
+ "eval_steps_per_second": 4.213,
2040
+ "step": 2711
2041
+ },
2042
+ {
2043
+ "epoch": 45.15,
2044
+ "learning_rate": 1.685185185185185e-05,
2045
+ "loss": 0.0133,
2046
+ "step": 2720
2047
+ },
2048
+ {
2049
+ "epoch": 45.31,
2050
+ "learning_rate": 1.6770833333333332e-05,
2051
+ "loss": 0.0159,
2052
+ "step": 2730
2053
+ },
2054
+ {
2055
+ "epoch": 45.48,
2056
+ "learning_rate": 1.6689814814814815e-05,
2057
+ "loss": 0.0183,
2058
+ "step": 2740
2059
+ },
2060
+ {
2061
+ "epoch": 45.64,
2062
+ "learning_rate": 1.6608796296296294e-05,
2063
+ "loss": 0.0302,
2064
+ "step": 2750
2065
+ },
2066
+ {
2067
+ "epoch": 45.81,
2068
+ "learning_rate": 1.6527777777777777e-05,
2069
+ "loss": 0.0232,
2070
+ "step": 2760
2071
+ },
2072
+ {
2073
+ "epoch": 45.98,
2074
+ "learning_rate": 1.6446759259259256e-05,
2075
+ "loss": 0.0203,
2076
+ "step": 2770
2077
+ },
2078
+ {
2079
+ "epoch": 45.99,
2080
+ "eval_accuracy": 0.8043478260869565,
2081
+ "eval_loss": 1.4792205095291138,
2082
+ "eval_runtime": 0.6983,
2083
+ "eval_samples_per_second": 65.879,
2084
+ "eval_steps_per_second": 4.296,
2085
+ "step": 2771
2086
+ },
2087
+ {
2088
+ "epoch": 46.14,
2089
+ "learning_rate": 1.636574074074074e-05,
2090
+ "loss": 0.037,
2091
+ "step": 2780
2092
+ },
2093
+ {
2094
+ "epoch": 46.31,
2095
+ "learning_rate": 1.628472222222222e-05,
2096
+ "loss": 0.0337,
2097
+ "step": 2790
2098
+ },
2099
+ {
2100
+ "epoch": 46.47,
2101
+ "learning_rate": 1.6203703703703704e-05,
2102
+ "loss": 0.012,
2103
+ "step": 2800
2104
+ },
2105
+ {
2106
+ "epoch": 46.64,
2107
+ "learning_rate": 1.6122685185185183e-05,
2108
+ "loss": 0.0568,
2109
+ "step": 2810
2110
+ },
2111
+ {
2112
+ "epoch": 46.8,
2113
+ "learning_rate": 1.6041666666666666e-05,
2114
+ "loss": 0.018,
2115
+ "step": 2820
2116
+ },
2117
+ {
2118
+ "epoch": 46.97,
2119
+ "learning_rate": 1.596064814814815e-05,
2120
+ "loss": 0.03,
2121
+ "step": 2830
2122
+ },
2123
+ {
2124
+ "epoch": 46.99,
2125
+ "eval_accuracy": 0.717391304347826,
2126
+ "eval_loss": 1.7487258911132812,
2127
+ "eval_runtime": 0.7112,
2128
+ "eval_samples_per_second": 64.68,
2129
+ "eval_steps_per_second": 4.218,
2130
+ "step": 2831
2131
+ },
2132
+ {
2133
+ "epoch": 47.14,
2134
+ "learning_rate": 1.5879629629629628e-05,
2135
+ "loss": 0.0083,
2136
+ "step": 2840
2137
+ },
2138
+ {
2139
+ "epoch": 47.3,
2140
+ "learning_rate": 1.579861111111111e-05,
2141
+ "loss": 0.0431,
2142
+ "step": 2850
2143
+ },
2144
+ {
2145
+ "epoch": 47.47,
2146
+ "learning_rate": 1.571759259259259e-05,
2147
+ "loss": 0.0104,
2148
+ "step": 2860
2149
+ },
2150
+ {
2151
+ "epoch": 47.63,
2152
+ "learning_rate": 1.5636574074074072e-05,
2153
+ "loss": 0.0322,
2154
+ "step": 2870
2155
+ },
2156
+ {
2157
+ "epoch": 47.8,
2158
+ "learning_rate": 1.5555555555555555e-05,
2159
+ "loss": 0.0185,
2160
+ "step": 2880
2161
+ },
2162
+ {
2163
+ "epoch": 47.97,
2164
+ "learning_rate": 1.5474537037037038e-05,
2165
+ "loss": 0.025,
2166
+ "step": 2890
2167
+ },
2168
+ {
2169
+ "epoch": 48.0,
2170
+ "eval_accuracy": 0.7608695652173914,
2171
+ "eval_loss": 1.6605217456817627,
2172
+ "eval_runtime": 0.6827,
2173
+ "eval_samples_per_second": 67.383,
2174
+ "eval_steps_per_second": 4.395,
2175
+ "step": 2892
2176
+ },
2177
+ {
2178
+ "epoch": 48.13,
2179
+ "learning_rate": 1.5393518518518517e-05,
2180
+ "loss": 0.0269,
2181
+ "step": 2900
2182
+ },
2183
+ {
2184
+ "epoch": 48.3,
2185
+ "learning_rate": 1.53125e-05,
2186
+ "loss": 0.0274,
2187
+ "step": 2910
2188
+ },
2189
+ {
2190
+ "epoch": 48.46,
2191
+ "learning_rate": 1.523148148148148e-05,
2192
+ "loss": 0.033,
2193
+ "step": 2920
2194
+ },
2195
+ {
2196
+ "epoch": 48.63,
2197
+ "learning_rate": 1.5150462962962961e-05,
2198
+ "loss": 0.0241,
2199
+ "step": 2930
2200
+ },
2201
+ {
2202
+ "epoch": 48.8,
2203
+ "learning_rate": 1.5069444444444444e-05,
2204
+ "loss": 0.0445,
2205
+ "step": 2940
2206
+ },
2207
+ {
2208
+ "epoch": 48.96,
2209
+ "learning_rate": 1.4988425925925925e-05,
2210
+ "loss": 0.0134,
2211
+ "step": 2950
2212
+ },
2213
+ {
2214
+ "epoch": 49.0,
2215
+ "eval_accuracy": 0.782608695652174,
2216
+ "eval_loss": 1.410630226135254,
2217
+ "eval_runtime": 0.7027,
2218
+ "eval_samples_per_second": 65.459,
2219
+ "eval_steps_per_second": 4.269,
2220
+ "step": 2952
2221
+ },
2222
+ {
2223
+ "epoch": 49.13,
2224
+ "learning_rate": 1.4907407407407406e-05,
2225
+ "loss": 0.0366,
2226
+ "step": 2960
2227
+ },
2228
+ {
2229
+ "epoch": 49.29,
2230
+ "learning_rate": 1.4826388888888887e-05,
2231
+ "loss": 0.0219,
2232
+ "step": 2970
2233
+ },
2234
+ {
2235
+ "epoch": 49.46,
2236
+ "learning_rate": 1.4745370370370368e-05,
2237
+ "loss": 0.0092,
2238
+ "step": 2980
2239
+ },
2240
+ {
2241
+ "epoch": 49.63,
2242
+ "learning_rate": 1.4664351851851849e-05,
2243
+ "loss": 0.0295,
2244
+ "step": 2990
2245
+ },
2246
+ {
2247
+ "epoch": 49.79,
2248
+ "learning_rate": 1.4583333333333333e-05,
2249
+ "loss": 0.0183,
2250
+ "step": 3000
2251
+ },
2252
+ {
2253
+ "epoch": 49.96,
2254
+ "learning_rate": 1.4502314814814814e-05,
2255
+ "loss": 0.026,
2256
+ "step": 3010
2257
+ },
2258
+ {
2259
+ "epoch": 49.99,
2260
+ "eval_accuracy": 0.7608695652173914,
2261
+ "eval_loss": 1.2972004413604736,
2262
+ "eval_runtime": 0.6842,
2263
+ "eval_samples_per_second": 67.236,
2264
+ "eval_steps_per_second": 4.385,
2265
+ "step": 3012
2266
+ },
2267
+ {
2268
+ "epoch": 50.12,
2269
+ "learning_rate": 1.4421296296296295e-05,
2270
+ "loss": 0.0207,
2271
+ "step": 3020
2272
+ },
2273
+ {
2274
+ "epoch": 50.29,
2275
+ "learning_rate": 1.4340277777777776e-05,
2276
+ "loss": 0.0111,
2277
+ "step": 3030
2278
+ },
2279
+ {
2280
+ "epoch": 50.46,
2281
+ "learning_rate": 1.4259259259259257e-05,
2282
+ "loss": 0.0364,
2283
+ "step": 3040
2284
+ },
2285
+ {
2286
+ "epoch": 50.62,
2287
+ "learning_rate": 1.417824074074074e-05,
2288
+ "loss": 0.0148,
2289
+ "step": 3050
2290
+ },
2291
+ {
2292
+ "epoch": 50.79,
2293
+ "learning_rate": 1.409722222222222e-05,
2294
+ "loss": 0.0271,
2295
+ "step": 3060
2296
+ },
2297
+ {
2298
+ "epoch": 50.95,
2299
+ "learning_rate": 1.4016203703703703e-05,
2300
+ "loss": 0.0507,
2301
+ "step": 3070
2302
+ },
2303
+ {
2304
+ "epoch": 50.99,
2305
+ "eval_accuracy": 0.782608695652174,
2306
+ "eval_loss": 1.3302582502365112,
2307
+ "eval_runtime": 0.694,
2308
+ "eval_samples_per_second": 66.283,
2309
+ "eval_steps_per_second": 4.323,
2310
+ "step": 3072
2311
+ },
2312
+ {
2313
+ "epoch": 51.12,
2314
+ "learning_rate": 1.3935185185185184e-05,
2315
+ "loss": 0.0164,
2316
+ "step": 3080
2317
+ },
2318
+ {
2319
+ "epoch": 51.29,
2320
+ "learning_rate": 1.3854166666666665e-05,
2321
+ "loss": 0.0151,
2322
+ "step": 3090
2323
+ },
2324
+ {
2325
+ "epoch": 51.45,
2326
+ "learning_rate": 1.3773148148148148e-05,
2327
+ "loss": 0.0495,
2328
+ "step": 3100
2329
+ },
2330
+ {
2331
+ "epoch": 51.62,
2332
+ "learning_rate": 1.3692129629629629e-05,
2333
+ "loss": 0.0056,
2334
+ "step": 3110
2335
+ },
2336
+ {
2337
+ "epoch": 51.78,
2338
+ "learning_rate": 1.361111111111111e-05,
2339
+ "loss": 0.0323,
2340
+ "step": 3120
2341
+ },
2342
+ {
2343
+ "epoch": 51.95,
2344
+ "learning_rate": 1.353009259259259e-05,
2345
+ "loss": 0.0394,
2346
+ "step": 3130
2347
+ },
2348
+ {
2349
+ "epoch": 52.0,
2350
+ "eval_accuracy": 0.8478260869565217,
2351
+ "eval_loss": 1.1954048871994019,
2352
+ "eval_runtime": 0.7081,
2353
+ "eval_samples_per_second": 64.959,
2354
+ "eval_steps_per_second": 4.236,
2355
+ "step": 3133
2356
+ },
2357
+ {
2358
+ "epoch": 52.12,
2359
+ "learning_rate": 1.3449074074074072e-05,
2360
+ "loss": 0.0266,
2361
+ "step": 3140
2362
+ },
2363
+ {
2364
+ "epoch": 52.28,
2365
+ "learning_rate": 1.3368055555555554e-05,
2366
+ "loss": 0.0354,
2367
+ "step": 3150
2368
+ },
2369
+ {
2370
+ "epoch": 52.45,
2371
+ "learning_rate": 1.3287037037037037e-05,
2372
+ "loss": 0.0168,
2373
+ "step": 3160
2374
+ },
2375
+ {
2376
+ "epoch": 52.61,
2377
+ "learning_rate": 1.3206018518518518e-05,
2378
+ "loss": 0.0259,
2379
+ "step": 3170
2380
+ },
2381
+ {
2382
+ "epoch": 52.78,
2383
+ "learning_rate": 1.3124999999999999e-05,
2384
+ "loss": 0.027,
2385
+ "step": 3180
2386
+ },
2387
+ {
2388
+ "epoch": 52.95,
2389
+ "learning_rate": 1.304398148148148e-05,
2390
+ "loss": 0.0271,
2391
+ "step": 3190
2392
+ },
2393
+ {
2394
+ "epoch": 53.0,
2395
+ "eval_accuracy": 0.8260869565217391,
2396
+ "eval_loss": 1.312525987625122,
2397
+ "eval_runtime": 0.6832,
2398
+ "eval_samples_per_second": 67.329,
2399
+ "eval_steps_per_second": 4.391,
2400
+ "step": 3193
2401
+ },
2402
+ {
2403
+ "epoch": 53.11,
2404
+ "learning_rate": 1.296296296296296e-05,
2405
+ "loss": 0.0101,
2406
+ "step": 3200
2407
+ },
2408
+ {
2409
+ "epoch": 53.28,
2410
+ "learning_rate": 1.2881944444444443e-05,
2411
+ "loss": 0.0212,
2412
+ "step": 3210
2413
+ },
2414
+ {
2415
+ "epoch": 53.44,
2416
+ "learning_rate": 1.2800925925925926e-05,
2417
+ "loss": 0.0057,
2418
+ "step": 3220
2419
+ },
2420
+ {
2421
+ "epoch": 53.61,
2422
+ "learning_rate": 1.2719907407407407e-05,
2423
+ "loss": 0.0132,
2424
+ "step": 3230
2425
+ },
2426
+ {
2427
+ "epoch": 53.78,
2428
+ "learning_rate": 1.2638888888888888e-05,
2429
+ "loss": 0.0173,
2430
+ "step": 3240
2431
+ },
2432
+ {
2433
+ "epoch": 53.94,
2434
+ "learning_rate": 1.2557870370370369e-05,
2435
+ "loss": 0.0115,
2436
+ "step": 3250
2437
+ },
2438
+ {
2439
+ "epoch": 53.99,
2440
+ "eval_accuracy": 0.8478260869565217,
2441
+ "eval_loss": 1.3443913459777832,
2442
+ "eval_runtime": 0.7167,
2443
+ "eval_samples_per_second": 64.185,
2444
+ "eval_steps_per_second": 4.186,
2445
+ "step": 3253
2446
+ },
2447
+ {
2448
+ "epoch": 54.11,
2449
+ "learning_rate": 1.247685185185185e-05,
2450
+ "loss": 0.0028,
2451
+ "step": 3260
2452
+ },
2453
+ {
2454
+ "epoch": 54.27,
2455
+ "learning_rate": 1.2395833333333333e-05,
2456
+ "loss": 0.0184,
2457
+ "step": 3270
2458
+ },
2459
+ {
2460
+ "epoch": 54.44,
2461
+ "learning_rate": 1.2314814814814813e-05,
2462
+ "loss": 0.0007,
2463
+ "step": 3280
2464
+ },
2465
+ {
2466
+ "epoch": 54.61,
2467
+ "learning_rate": 1.2233796296296294e-05,
2468
+ "loss": 0.0233,
2469
+ "step": 3290
2470
+ },
2471
+ {
2472
+ "epoch": 54.77,
2473
+ "learning_rate": 1.2152777777777777e-05,
2474
+ "loss": 0.0078,
2475
+ "step": 3300
2476
+ },
2477
+ {
2478
+ "epoch": 54.94,
2479
+ "learning_rate": 1.2071759259259258e-05,
2480
+ "loss": 0.0138,
2481
+ "step": 3310
2482
+ },
2483
+ {
2484
+ "epoch": 54.99,
2485
+ "eval_accuracy": 0.8260869565217391,
2486
+ "eval_loss": 1.4688717126846313,
2487
+ "eval_runtime": 0.6497,
2488
+ "eval_samples_per_second": 70.806,
2489
+ "eval_steps_per_second": 4.618,
2490
+ "step": 3313
2491
+ },
2492
+ {
2493
+ "epoch": 55.1,
2494
+ "learning_rate": 1.199074074074074e-05,
2495
+ "loss": 0.0081,
2496
+ "step": 3320
2497
+ },
2498
+ {
2499
+ "epoch": 55.27,
2500
+ "learning_rate": 1.1909722222222222e-05,
2501
+ "loss": 0.0218,
2502
+ "step": 3330
2503
+ },
2504
+ {
2505
+ "epoch": 55.44,
2506
+ "learning_rate": 1.1828703703703703e-05,
2507
+ "loss": 0.0143,
2508
+ "step": 3340
2509
+ },
2510
+ {
2511
+ "epoch": 55.6,
2512
+ "learning_rate": 1.1747685185185184e-05,
2513
+ "loss": 0.0216,
2514
+ "step": 3350
2515
+ },
2516
+ {
2517
+ "epoch": 55.77,
2518
+ "learning_rate": 1.1666666666666665e-05,
2519
+ "loss": 0.0091,
2520
+ "step": 3360
2521
+ },
2522
+ {
2523
+ "epoch": 55.93,
2524
+ "learning_rate": 1.1585648148148149e-05,
2525
+ "loss": 0.0184,
2526
+ "step": 3370
2527
+ },
2528
+ {
2529
+ "epoch": 56.0,
2530
+ "eval_accuracy": 0.8260869565217391,
2531
+ "eval_loss": 1.4959295988082886,
2532
+ "eval_runtime": 0.6842,
2533
+ "eval_samples_per_second": 67.228,
2534
+ "eval_steps_per_second": 4.384,
2535
+ "step": 3374
2536
+ },
2537
+ {
2538
+ "epoch": 56.1,
2539
+ "learning_rate": 1.150462962962963e-05,
2540
+ "loss": 0.0134,
2541
+ "step": 3380
2542
+ },
2543
+ {
2544
+ "epoch": 56.27,
2545
+ "learning_rate": 1.142361111111111e-05,
2546
+ "loss": 0.0193,
2547
+ "step": 3390
2548
+ },
2549
+ {
2550
+ "epoch": 56.43,
2551
+ "learning_rate": 1.1342592592592592e-05,
2552
+ "loss": 0.0049,
2553
+ "step": 3400
2554
+ },
2555
+ {
2556
+ "epoch": 56.6,
2557
+ "learning_rate": 1.1261574074074073e-05,
2558
+ "loss": 0.0105,
2559
+ "step": 3410
2560
+ },
2561
+ {
2562
+ "epoch": 56.76,
2563
+ "learning_rate": 1.1180555555555554e-05,
2564
+ "loss": 0.0245,
2565
+ "step": 3420
2566
+ },
2567
+ {
2568
+ "epoch": 56.93,
2569
+ "learning_rate": 1.1099537037037036e-05,
2570
+ "loss": 0.0163,
2571
+ "step": 3430
2572
+ },
2573
+ {
2574
+ "epoch": 57.0,
2575
+ "eval_accuracy": 0.782608695652174,
2576
+ "eval_loss": 1.349004864692688,
2577
+ "eval_runtime": 0.6951,
2578
+ "eval_samples_per_second": 66.174,
2579
+ "eval_steps_per_second": 4.316,
2580
+ "step": 3434
2581
+ },
2582
+ {
2583
+ "epoch": 57.1,
2584
+ "learning_rate": 1.1018518518518517e-05,
2585
+ "loss": 0.0116,
2586
+ "step": 3440
2587
+ },
2588
+ {
2589
+ "epoch": 57.26,
2590
+ "learning_rate": 1.0937499999999998e-05,
2591
+ "loss": 0.0352,
2592
+ "step": 3450
2593
+ },
2594
+ {
2595
+ "epoch": 57.43,
2596
+ "learning_rate": 1.085648148148148e-05,
2597
+ "loss": 0.0316,
2598
+ "step": 3460
2599
+ },
2600
+ {
2601
+ "epoch": 57.59,
2602
+ "learning_rate": 1.0775462962962962e-05,
2603
+ "loss": 0.0291,
2604
+ "step": 3470
2605
+ },
2606
+ {
2607
+ "epoch": 57.76,
2608
+ "learning_rate": 1.0694444444444444e-05,
2609
+ "loss": 0.0388,
2610
+ "step": 3480
2611
+ },
2612
+ {
2613
+ "epoch": 57.93,
2614
+ "learning_rate": 1.0613425925925925e-05,
2615
+ "loss": 0.0112,
2616
+ "step": 3490
2617
+ },
2618
+ {
2619
+ "epoch": 57.99,
2620
+ "eval_accuracy": 0.782608695652174,
2621
+ "eval_loss": 1.4749095439910889,
2622
+ "eval_runtime": 0.6781,
2623
+ "eval_samples_per_second": 67.833,
2624
+ "eval_steps_per_second": 4.424,
2625
+ "step": 3494
2626
+ },
2627
+ {
2628
+ "epoch": 58.09,
2629
+ "learning_rate": 1.0532407407407406e-05,
2630
+ "loss": 0.025,
2631
+ "step": 3500
2632
+ },
2633
+ {
2634
+ "epoch": 58.26,
2635
+ "learning_rate": 1.0451388888888887e-05,
2636
+ "loss": 0.0354,
2637
+ "step": 3510
2638
+ },
2639
+ {
2640
+ "epoch": 58.42,
2641
+ "learning_rate": 1.0370370370370368e-05,
2642
+ "loss": 0.0064,
2643
+ "step": 3520
2644
+ },
2645
+ {
2646
+ "epoch": 58.59,
2647
+ "learning_rate": 1.028935185185185e-05,
2648
+ "loss": 0.0058,
2649
+ "step": 3530
2650
+ },
2651
+ {
2652
+ "epoch": 58.76,
2653
+ "learning_rate": 1.0208333333333334e-05,
2654
+ "loss": 0.0241,
2655
+ "step": 3540
2656
+ },
2657
+ {
2658
+ "epoch": 58.92,
2659
+ "learning_rate": 1.0127314814814815e-05,
2660
+ "loss": 0.0185,
2661
+ "step": 3550
2662
+ },
2663
+ {
2664
+ "epoch": 58.99,
2665
+ "eval_accuracy": 0.782608695652174,
2666
+ "eval_loss": 1.5823291540145874,
2667
+ "eval_runtime": 0.6695,
2668
+ "eval_samples_per_second": 68.712,
2669
+ "eval_steps_per_second": 4.481,
2670
+ "step": 3554
2671
+ },
2672
+ {
2673
+ "epoch": 59.09,
2674
+ "learning_rate": 1.0046296296296295e-05,
2675
+ "loss": 0.0089,
2676
+ "step": 3560
2677
+ },
2678
+ {
2679
+ "epoch": 59.25,
2680
+ "learning_rate": 9.965277777777776e-06,
2681
+ "loss": 0.0291,
2682
+ "step": 3570
2683
+ },
2684
+ {
2685
+ "epoch": 59.42,
2686
+ "learning_rate": 9.884259259259257e-06,
2687
+ "loss": 0.0118,
2688
+ "step": 3580
2689
+ },
2690
+ {
2691
+ "epoch": 59.59,
2692
+ "learning_rate": 9.80324074074074e-06,
2693
+ "loss": 0.0318,
2694
+ "step": 3590
2695
+ },
2696
+ {
2697
+ "epoch": 59.75,
2698
+ "learning_rate": 9.722222222222221e-06,
2699
+ "loss": 0.0395,
2700
+ "step": 3600
2701
+ },
2702
+ {
2703
+ "epoch": 59.92,
2704
+ "learning_rate": 9.641203703703704e-06,
2705
+ "loss": 0.031,
2706
+ "step": 3610
2707
+ },
2708
+ {
2709
+ "epoch": 60.0,
2710
+ "eval_accuracy": 0.782608695652174,
2711
+ "eval_loss": 1.5190129280090332,
2712
+ "eval_runtime": 0.7102,
2713
+ "eval_samples_per_second": 64.766,
2714
+ "eval_steps_per_second": 4.224,
2715
+ "step": 3615
2716
+ },
2717
+ {
2718
+ "epoch": 60.08,
2719
+ "learning_rate": 9.560185185185185e-06,
2720
+ "loss": 0.0127,
2721
+ "step": 3620
2722
+ },
2723
+ {
2724
+ "epoch": 60.25,
2725
+ "learning_rate": 9.479166666666666e-06,
2726
+ "loss": 0.0065,
2727
+ "step": 3630
2728
+ },
2729
+ {
2730
+ "epoch": 60.41,
2731
+ "learning_rate": 9.398148148148148e-06,
2732
+ "loss": 0.0129,
2733
+ "step": 3640
2734
+ },
2735
+ {
2736
+ "epoch": 60.58,
2737
+ "learning_rate": 9.31712962962963e-06,
2738
+ "loss": 0.0157,
2739
+ "step": 3650
2740
+ },
2741
+ {
2742
+ "epoch": 60.75,
2743
+ "learning_rate": 9.23611111111111e-06,
2744
+ "loss": 0.0134,
2745
+ "step": 3660
2746
+ },
2747
+ {
2748
+ "epoch": 60.91,
2749
+ "learning_rate": 9.155092592592591e-06,
2750
+ "loss": 0.0161,
2751
+ "step": 3670
2752
+ },
2753
+ {
2754
+ "epoch": 61.0,
2755
+ "eval_accuracy": 0.8043478260869565,
2756
+ "eval_loss": 1.5475845336914062,
2757
+ "eval_runtime": 0.7113,
2758
+ "eval_samples_per_second": 64.674,
2759
+ "eval_steps_per_second": 4.218,
2760
+ "step": 3675
2761
+ },
2762
+ {
2763
+ "epoch": 61.08,
2764
+ "learning_rate": 9.074074074074072e-06,
2765
+ "loss": 0.0117,
2766
+ "step": 3680
2767
+ },
2768
+ {
2769
+ "epoch": 61.24,
2770
+ "learning_rate": 8.993055555555555e-06,
2771
+ "loss": 0.0135,
2772
+ "step": 3690
2773
+ },
2774
+ {
2775
+ "epoch": 61.41,
2776
+ "learning_rate": 8.912037037037037e-06,
2777
+ "loss": 0.0179,
2778
+ "step": 3700
2779
+ },
2780
+ {
2781
+ "epoch": 61.58,
2782
+ "learning_rate": 8.831018518518518e-06,
2783
+ "loss": 0.0253,
2784
+ "step": 3710
2785
+ },
2786
+ {
2787
+ "epoch": 61.74,
2788
+ "learning_rate": 8.75e-06,
2789
+ "loss": 0.0267,
2790
+ "step": 3720
2791
+ },
2792
+ {
2793
+ "epoch": 61.91,
2794
+ "learning_rate": 8.66898148148148e-06,
2795
+ "loss": 0.0146,
2796
+ "step": 3730
2797
+ },
2798
+ {
2799
+ "epoch": 61.99,
2800
+ "eval_accuracy": 0.782608695652174,
2801
+ "eval_loss": 1.3930317163467407,
2802
+ "eval_runtime": 0.6749,
2803
+ "eval_samples_per_second": 68.162,
2804
+ "eval_steps_per_second": 4.445,
2805
+ "step": 3735
2806
+ },
2807
+ {
2808
+ "epoch": 62.07,
2809
+ "learning_rate": 8.587962962962963e-06,
2810
+ "loss": 0.0202,
2811
+ "step": 3740
2812
+ },
2813
+ {
2814
+ "epoch": 62.24,
2815
+ "learning_rate": 8.506944444444444e-06,
2816
+ "loss": 0.0162,
2817
+ "step": 3750
2818
+ },
2819
+ {
2820
+ "epoch": 62.41,
2821
+ "learning_rate": 8.425925925925925e-06,
2822
+ "loss": 0.0199,
2823
+ "step": 3760
2824
+ },
2825
+ {
2826
+ "epoch": 62.57,
2827
+ "learning_rate": 8.344907407407407e-06,
2828
+ "loss": 0.0135,
2829
+ "step": 3770
2830
+ },
2831
+ {
2832
+ "epoch": 62.74,
2833
+ "learning_rate": 8.263888888888888e-06,
2834
+ "loss": 0.0132,
2835
+ "step": 3780
2836
+ },
2837
+ {
2838
+ "epoch": 62.9,
2839
+ "learning_rate": 8.18287037037037e-06,
2840
+ "loss": 0.005,
2841
+ "step": 3790
2842
+ },
2843
+ {
2844
+ "epoch": 62.99,
2845
+ "eval_accuracy": 0.8043478260869565,
2846
+ "eval_loss": 1.5454456806182861,
2847
+ "eval_runtime": 0.7582,
2848
+ "eval_samples_per_second": 60.67,
2849
+ "eval_steps_per_second": 3.957,
2850
+ "step": 3795
2851
+ },
2852
+ {
2853
+ "epoch": 63.07,
2854
+ "learning_rate": 8.101851851851852e-06,
2855
+ "loss": 0.03,
2856
+ "step": 3800
2857
+ },
2858
+ {
2859
+ "epoch": 63.24,
2860
+ "learning_rate": 8.020833333333333e-06,
2861
+ "loss": 0.0326,
2862
+ "step": 3810
2863
+ },
2864
+ {
2865
+ "epoch": 63.4,
2866
+ "learning_rate": 7.939814814814814e-06,
2867
+ "loss": 0.0128,
2868
+ "step": 3820
2869
+ },
2870
+ {
2871
+ "epoch": 63.57,
2872
+ "learning_rate": 7.858796296296295e-06,
2873
+ "loss": 0.0108,
2874
+ "step": 3830
2875
+ },
2876
+ {
2877
+ "epoch": 63.73,
2878
+ "learning_rate": 7.777777777777777e-06,
2879
+ "loss": 0.0011,
2880
+ "step": 3840
2881
+ },
2882
+ {
2883
+ "epoch": 63.9,
2884
+ "learning_rate": 7.696759259259258e-06,
2885
+ "loss": 0.0093,
2886
+ "step": 3850
2887
+ },
2888
+ {
2889
+ "epoch": 64.0,
2890
+ "eval_accuracy": 0.782608695652174,
2891
+ "eval_loss": 1.5958888530731201,
2892
+ "eval_runtime": 0.6768,
2893
+ "eval_samples_per_second": 67.965,
2894
+ "eval_steps_per_second": 4.433,
2895
+ "step": 3856
2896
+ },
2897
+ {
2898
+ "epoch": 64.07,
2899
+ "learning_rate": 7.61574074074074e-06,
2900
+ "loss": 0.0132,
2901
+ "step": 3860
2902
+ },
2903
+ {
2904
+ "epoch": 64.23,
2905
+ "learning_rate": 7.534722222222222e-06,
2906
+ "loss": 0.0085,
2907
+ "step": 3870
2908
+ },
2909
+ {
2910
+ "epoch": 64.4,
2911
+ "learning_rate": 7.453703703703703e-06,
2912
+ "loss": 0.0326,
2913
+ "step": 3880
2914
+ },
2915
+ {
2916
+ "epoch": 64.56,
2917
+ "learning_rate": 7.372685185185184e-06,
2918
+ "loss": 0.0295,
2919
+ "step": 3890
2920
+ },
2921
+ {
2922
+ "epoch": 64.73,
2923
+ "learning_rate": 7.291666666666667e-06,
2924
+ "loss": 0.0172,
2925
+ "step": 3900
2926
+ },
2927
+ {
2928
+ "epoch": 64.9,
2929
+ "learning_rate": 7.2106481481481475e-06,
2930
+ "loss": 0.0224,
2931
+ "step": 3910
2932
+ },
2933
+ {
2934
+ "epoch": 65.0,
2935
+ "eval_accuracy": 0.8043478260869565,
2936
+ "eval_loss": 1.4553816318511963,
2937
+ "eval_runtime": 0.7364,
2938
+ "eval_samples_per_second": 62.467,
2939
+ "eval_steps_per_second": 4.074,
2940
+ "step": 3916
2941
+ },
2942
+ {
2943
+ "epoch": 65.06,
2944
+ "learning_rate": 7.1296296296296285e-06,
2945
+ "loss": 0.0208,
2946
+ "step": 3920
2947
+ },
2948
+ {
2949
+ "epoch": 65.23,
2950
+ "learning_rate": 7.04861111111111e-06,
2951
+ "loss": 0.023,
2952
+ "step": 3930
2953
+ },
2954
+ {
2955
+ "epoch": 65.39,
2956
+ "learning_rate": 6.967592592592592e-06,
2957
+ "loss": 0.0205,
2958
+ "step": 3940
2959
+ },
2960
+ {
2961
+ "epoch": 65.56,
2962
+ "learning_rate": 6.886574074074074e-06,
2963
+ "loss": 0.0088,
2964
+ "step": 3950
2965
+ },
2966
+ {
2967
+ "epoch": 65.73,
2968
+ "learning_rate": 6.805555555555555e-06,
2969
+ "loss": 0.0223,
2970
+ "step": 3960
2971
+ },
2972
+ {
2973
+ "epoch": 65.89,
2974
+ "learning_rate": 6.724537037037036e-06,
2975
+ "loss": 0.0154,
2976
+ "step": 3970
2977
+ },
2978
+ {
2979
+ "epoch": 65.99,
2980
+ "eval_accuracy": 0.8260869565217391,
2981
+ "eval_loss": 1.5326697826385498,
2982
+ "eval_runtime": 0.7162,
2983
+ "eval_samples_per_second": 64.228,
2984
+ "eval_steps_per_second": 4.189,
2985
+ "step": 3976
2986
+ },
2987
+ {
2988
+ "epoch": 66.06,
2989
+ "learning_rate": 6.6435185185185185e-06,
2990
+ "loss": 0.0093,
2991
+ "step": 3980
2992
+ },
2993
+ {
2994
+ "epoch": 66.22,
2995
+ "learning_rate": 6.5624999999999994e-06,
2996
+ "loss": 0.0138,
2997
+ "step": 3990
2998
+ },
2999
+ {
3000
+ "epoch": 66.39,
3001
+ "learning_rate": 6.48148148148148e-06,
3002
+ "loss": 0.0059,
3003
+ "step": 4000
3004
+ },
3005
+ {
3006
+ "epoch": 66.56,
3007
+ "learning_rate": 6.400462962962963e-06,
3008
+ "loss": 0.0033,
3009
+ "step": 4010
3010
+ },
3011
+ {
3012
+ "epoch": 66.72,
3013
+ "learning_rate": 6.319444444444444e-06,
3014
+ "loss": 0.0237,
3015
+ "step": 4020
3016
+ },
3017
+ {
3018
+ "epoch": 66.89,
3019
+ "learning_rate": 6.238425925925925e-06,
3020
+ "loss": 0.0116,
3021
+ "step": 4030
3022
+ },
3023
+ {
3024
+ "epoch": 66.99,
3025
+ "eval_accuracy": 0.8043478260869565,
3026
+ "eval_loss": 1.6029945611953735,
3027
+ "eval_runtime": 0.7141,
3028
+ "eval_samples_per_second": 64.418,
3029
+ "eval_steps_per_second": 4.201,
3030
+ "step": 4036
3031
+ },
3032
+ {
3033
+ "epoch": 67.05,
3034
+ "learning_rate": 6.157407407407407e-06,
3035
+ "loss": 0.0064,
3036
+ "step": 4040
3037
+ },
3038
+ {
3039
+ "epoch": 67.22,
3040
+ "learning_rate": 6.0763888888888885e-06,
3041
+ "loss": 0.0129,
3042
+ "step": 4050
3043
+ },
3044
+ {
3045
+ "epoch": 67.39,
3046
+ "learning_rate": 5.99537037037037e-06,
3047
+ "loss": 0.0032,
3048
+ "step": 4060
3049
+ },
3050
+ {
3051
+ "epoch": 67.55,
3052
+ "learning_rate": 5.914351851851851e-06,
3053
+ "loss": 0.0101,
3054
+ "step": 4070
3055
+ },
3056
+ {
3057
+ "epoch": 67.72,
3058
+ "learning_rate": 5.833333333333332e-06,
3059
+ "loss": 0.0051,
3060
+ "step": 4080
3061
+ },
3062
+ {
3063
+ "epoch": 67.88,
3064
+ "learning_rate": 5.752314814814815e-06,
3065
+ "loss": 0.0037,
3066
+ "step": 4090
3067
+ },
3068
+ {
3069
+ "epoch": 68.0,
3070
+ "eval_accuracy": 0.8260869565217391,
3071
+ "eval_loss": 1.5046288967132568,
3072
+ "eval_runtime": 0.7027,
3073
+ "eval_samples_per_second": 65.464,
3074
+ "eval_steps_per_second": 4.269,
3075
+ "step": 4097
3076
+ },
3077
+ {
3078
+ "epoch": 68.05,
3079
+ "learning_rate": 5.671296296296296e-06,
3080
+ "loss": 0.0151,
3081
+ "step": 4100
3082
+ },
3083
+ {
3084
+ "epoch": 68.22,
3085
+ "learning_rate": 5.590277777777777e-06,
3086
+ "loss": 0.0007,
3087
+ "step": 4110
3088
+ },
3089
+ {
3090
+ "epoch": 68.38,
3091
+ "learning_rate": 5.509259259259259e-06,
3092
+ "loss": 0.0212,
3093
+ "step": 4120
3094
+ },
3095
+ {
3096
+ "epoch": 68.55,
3097
+ "learning_rate": 5.42824074074074e-06,
3098
+ "loss": 0.0038,
3099
+ "step": 4130
3100
+ },
3101
+ {
3102
+ "epoch": 68.71,
3103
+ "learning_rate": 5.347222222222222e-06,
3104
+ "loss": 0.0211,
3105
+ "step": 4140
3106
+ },
3107
+ {
3108
+ "epoch": 68.88,
3109
+ "learning_rate": 5.266203703703703e-06,
3110
+ "loss": 0.0023,
3111
+ "step": 4150
3112
+ },
3113
+ {
3114
+ "epoch": 69.0,
3115
+ "eval_accuracy": 0.8260869565217391,
3116
+ "eval_loss": 1.5222116708755493,
3117
+ "eval_runtime": 0.6773,
3118
+ "eval_samples_per_second": 67.922,
3119
+ "eval_steps_per_second": 4.43,
3120
+ "step": 4157
3121
+ },
3122
+ {
3123
+ "epoch": 69.05,
3124
+ "learning_rate": 5.185185185185184e-06,
3125
+ "loss": 0.012,
3126
+ "step": 4160
3127
+ },
3128
+ {
3129
+ "epoch": 69.21,
3130
+ "learning_rate": 5.104166666666667e-06,
3131
+ "loss": 0.0086,
3132
+ "step": 4170
3133
+ },
3134
+ {
3135
+ "epoch": 69.38,
3136
+ "learning_rate": 5.023148148148148e-06,
3137
+ "loss": 0.0203,
3138
+ "step": 4180
3139
+ },
3140
+ {
3141
+ "epoch": 69.54,
3142
+ "learning_rate": 4.942129629629629e-06,
3143
+ "loss": 0.0074,
3144
+ "step": 4190
3145
+ },
3146
+ {
3147
+ "epoch": 69.71,
3148
+ "learning_rate": 4.8611111111111105e-06,
3149
+ "loss": 0.0071,
3150
+ "step": 4200
3151
+ },
3152
+ {
3153
+ "epoch": 69.88,
3154
+ "learning_rate": 4.780092592592592e-06,
3155
+ "loss": 0.0068,
3156
+ "step": 4210
3157
+ },
3158
+ {
3159
+ "epoch": 69.99,
3160
+ "eval_accuracy": 0.8260869565217391,
3161
+ "eval_loss": 1.4339091777801514,
3162
+ "eval_runtime": 0.686,
3163
+ "eval_samples_per_second": 67.057,
3164
+ "eval_steps_per_second": 4.373,
3165
+ "step": 4217
3166
+ },
3167
+ {
3168
+ "epoch": 70.04,
3169
+ "learning_rate": 4.699074074074074e-06,
3170
+ "loss": 0.0015,
3171
+ "step": 4220
3172
+ },
3173
+ {
3174
+ "epoch": 70.21,
3175
+ "learning_rate": 4.618055555555555e-06,
3176
+ "loss": 0.0096,
3177
+ "step": 4230
3178
+ },
3179
+ {
3180
+ "epoch": 70.37,
3181
+ "learning_rate": 4.537037037037036e-06,
3182
+ "loss": 0.0074,
3183
+ "step": 4240
3184
+ },
3185
+ {
3186
+ "epoch": 70.54,
3187
+ "learning_rate": 4.456018518518519e-06,
3188
+ "loss": 0.0172,
3189
+ "step": 4250
3190
+ },
3191
+ {
3192
+ "epoch": 70.71,
3193
+ "learning_rate": 4.375e-06,
3194
+ "loss": 0.0024,
3195
+ "step": 4260
3196
+ },
3197
+ {
3198
+ "epoch": 70.87,
3199
+ "learning_rate": 4.293981481481481e-06,
3200
+ "loss": 0.0342,
3201
+ "step": 4270
3202
+ },
3203
+ {
3204
+ "epoch": 70.99,
3205
+ "eval_accuracy": 0.8043478260869565,
3206
+ "eval_loss": 1.6964349746704102,
3207
+ "eval_runtime": 0.7042,
3208
+ "eval_samples_per_second": 65.326,
3209
+ "eval_steps_per_second": 4.26,
3210
+ "step": 4277
3211
+ },
3212
+ {
3213
+ "epoch": 71.04,
3214
+ "learning_rate": 4.212962962962962e-06,
3215
+ "loss": 0.0028,
3216
+ "step": 4280
3217
+ },
3218
+ {
3219
+ "epoch": 71.2,
3220
+ "learning_rate": 4.131944444444444e-06,
3221
+ "loss": 0.0185,
3222
+ "step": 4290
3223
+ },
3224
+ {
3225
+ "epoch": 71.37,
3226
+ "learning_rate": 4.050925925925926e-06,
3227
+ "loss": 0.0172,
3228
+ "step": 4300
3229
+ },
3230
+ {
3231
+ "epoch": 71.54,
3232
+ "learning_rate": 3.969907407407407e-06,
3233
+ "loss": 0.0051,
3234
+ "step": 4310
3235
+ },
3236
+ {
3237
+ "epoch": 71.7,
3238
+ "learning_rate": 3.888888888888889e-06,
3239
+ "loss": 0.0049,
3240
+ "step": 4320
3241
+ },
3242
+ {
3243
+ "epoch": 71.87,
3244
+ "learning_rate": 3.80787037037037e-06,
3245
+ "loss": 0.0077,
3246
+ "step": 4330
3247
+ },
3248
+ {
3249
+ "epoch": 72.0,
3250
+ "eval_accuracy": 0.8043478260869565,
3251
+ "eval_loss": 1.6102274656295776,
3252
+ "eval_runtime": 0.691,
3253
+ "eval_samples_per_second": 66.569,
3254
+ "eval_steps_per_second": 4.341,
3255
+ "step": 4338
3256
+ },
3257
+ {
3258
+ "epoch": 72.03,
3259
+ "learning_rate": 3.7268518518518515e-06,
3260
+ "loss": 0.0207,
3261
+ "step": 4340
3262
+ },
3263
+ {
3264
+ "epoch": 72.2,
3265
+ "learning_rate": 3.6458333333333333e-06,
3266
+ "loss": 0.0135,
3267
+ "step": 4350
3268
+ },
3269
+ {
3270
+ "epoch": 72.37,
3271
+ "learning_rate": 3.5648148148148143e-06,
3272
+ "loss": 0.0098,
3273
+ "step": 4360
3274
+ },
3275
+ {
3276
+ "epoch": 72.53,
3277
+ "learning_rate": 3.483796296296296e-06,
3278
+ "loss": 0.0207,
3279
+ "step": 4370
3280
+ },
3281
+ {
3282
+ "epoch": 72.7,
3283
+ "learning_rate": 3.4027777777777774e-06,
3284
+ "loss": 0.0007,
3285
+ "step": 4380
3286
+ },
3287
+ {
3288
+ "epoch": 72.86,
3289
+ "learning_rate": 3.3217592592592592e-06,
3290
+ "loss": 0.0043,
3291
+ "step": 4390
3292
+ },
3293
+ {
3294
+ "epoch": 73.0,
3295
+ "eval_accuracy": 0.8043478260869565,
3296
+ "eval_loss": 1.668748140335083,
3297
+ "eval_runtime": 0.7135,
3298
+ "eval_samples_per_second": 64.472,
3299
+ "eval_steps_per_second": 4.205,
3300
+ "step": 4398
3301
+ },
3302
+ {
3303
+ "epoch": 73.03,
3304
+ "learning_rate": 3.24074074074074e-06,
3305
+ "loss": 0.0045,
3306
+ "step": 4400
3307
+ },
3308
+ {
3309
+ "epoch": 73.2,
3310
+ "learning_rate": 3.159722222222222e-06,
3311
+ "loss": 0.0121,
3312
+ "step": 4410
3313
+ },
3314
+ {
3315
+ "epoch": 73.36,
3316
+ "learning_rate": 3.0787037037037034e-06,
3317
+ "loss": 0.0131,
3318
+ "step": 4420
3319
+ },
3320
+ {
3321
+ "epoch": 73.53,
3322
+ "learning_rate": 2.997685185185185e-06,
3323
+ "loss": 0.0134,
3324
+ "step": 4430
3325
+ },
3326
+ {
3327
+ "epoch": 73.69,
3328
+ "learning_rate": 2.916666666666666e-06,
3329
+ "loss": 0.0233,
3330
+ "step": 4440
3331
+ },
3332
+ {
3333
+ "epoch": 73.86,
3334
+ "learning_rate": 2.835648148148148e-06,
3335
+ "loss": 0.0131,
3336
+ "step": 4450
3337
+ },
3338
+ {
3339
+ "epoch": 73.99,
3340
+ "eval_accuracy": 0.8043478260869565,
3341
+ "eval_loss": 1.6847338676452637,
3342
+ "eval_runtime": 0.702,
3343
+ "eval_samples_per_second": 65.531,
3344
+ "eval_steps_per_second": 4.274,
3345
+ "step": 4458
3346
+ },
3347
+ {
3348
+ "epoch": 74.02,
3349
+ "learning_rate": 2.7546296296296293e-06,
3350
+ "loss": 0.0151,
3351
+ "step": 4460
3352
+ },
3353
+ {
3354
+ "epoch": 74.19,
3355
+ "learning_rate": 2.673611111111111e-06,
3356
+ "loss": 0.0031,
3357
+ "step": 4470
3358
+ },
3359
+ {
3360
+ "epoch": 74.36,
3361
+ "learning_rate": 2.592592592592592e-06,
3362
+ "loss": 0.0135,
3363
+ "step": 4480
3364
+ },
3365
+ {
3366
+ "epoch": 74.52,
3367
+ "learning_rate": 2.511574074074074e-06,
3368
+ "loss": 0.0121,
3369
+ "step": 4490
3370
+ },
3371
+ {
3372
+ "epoch": 74.69,
3373
+ "learning_rate": 2.4305555555555552e-06,
3374
+ "loss": 0.0173,
3375
+ "step": 4500
3376
+ },
3377
+ {
3378
+ "epoch": 74.85,
3379
+ "learning_rate": 2.349537037037037e-06,
3380
+ "loss": 0.0031,
3381
+ "step": 4510
3382
+ },
3383
+ {
3384
+ "epoch": 74.99,
3385
+ "eval_accuracy": 0.8043478260869565,
3386
+ "eval_loss": 1.7195295095443726,
3387
+ "eval_runtime": 0.7118,
3388
+ "eval_samples_per_second": 64.626,
3389
+ "eval_steps_per_second": 4.215,
3390
+ "step": 4518
3391
+ },
3392
+ {
3393
+ "epoch": 75.02,
3394
+ "learning_rate": 2.268518518518518e-06,
3395
+ "loss": 0.0069,
3396
+ "step": 4520
3397
+ },
3398
+ {
3399
+ "epoch": 75.19,
3400
+ "learning_rate": 2.1875e-06,
3401
+ "loss": 0.0063,
3402
+ "step": 4530
3403
+ },
3404
+ {
3405
+ "epoch": 75.35,
3406
+ "learning_rate": 2.106481481481481e-06,
3407
+ "loss": 0.0288,
3408
+ "step": 4540
3409
+ },
3410
+ {
3411
+ "epoch": 75.52,
3412
+ "learning_rate": 2.025462962962963e-06,
3413
+ "loss": 0.0162,
3414
+ "step": 4550
3415
+ },
3416
+ {
3417
+ "epoch": 75.68,
3418
+ "learning_rate": 1.9444444444444444e-06,
3419
+ "loss": 0.0041,
3420
+ "step": 4560
3421
+ },
3422
+ {
3423
+ "epoch": 75.85,
3424
+ "learning_rate": 1.8634259259259257e-06,
3425
+ "loss": 0.0087,
3426
+ "step": 4570
3427
+ },
3428
+ {
3429
+ "epoch": 76.0,
3430
+ "eval_accuracy": 0.782608695652174,
3431
+ "eval_loss": 1.720942735671997,
3432
+ "eval_runtime": 0.666,
3433
+ "eval_samples_per_second": 69.067,
3434
+ "eval_steps_per_second": 4.504,
3435
+ "step": 4579
3436
+ },
3437
+ {
3438
+ "epoch": 76.02,
3439
+ "learning_rate": 1.7824074074074071e-06,
3440
+ "loss": 0.0145,
3441
+ "step": 4580
3442
+ },
3443
+ {
3444
+ "epoch": 76.18,
3445
+ "learning_rate": 1.7013888888888887e-06,
3446
+ "loss": 0.0056,
3447
+ "step": 4590
3448
+ },
3449
+ {
3450
+ "epoch": 76.35,
3451
+ "learning_rate": 1.62037037037037e-06,
3452
+ "loss": 0.0108,
3453
+ "step": 4600
3454
+ },
3455
+ {
3456
+ "epoch": 76.51,
3457
+ "learning_rate": 1.5393518518518517e-06,
3458
+ "loss": 0.0152,
3459
+ "step": 4610
3460
+ },
3461
+ {
3462
+ "epoch": 76.68,
3463
+ "learning_rate": 1.458333333333333e-06,
3464
+ "loss": 0.0126,
3465
+ "step": 4620
3466
+ },
3467
+ {
3468
+ "epoch": 76.85,
3469
+ "learning_rate": 1.3773148148148147e-06,
3470
+ "loss": 0.0219,
3471
+ "step": 4630
3472
+ },
3473
+ {
3474
+ "epoch": 77.0,
3475
+ "eval_accuracy": 0.8043478260869565,
3476
+ "eval_loss": 1.6714533567428589,
3477
+ "eval_runtime": 0.7121,
3478
+ "eval_samples_per_second": 64.594,
3479
+ "eval_steps_per_second": 4.213,
3480
+ "step": 4639
3481
+ },
3482
+ {
3483
+ "epoch": 77.01,
3484
+ "learning_rate": 1.296296296296296e-06,
3485
+ "loss": 0.0078,
3486
+ "step": 4640
3487
+ },
3488
+ {
3489
+ "epoch": 77.18,
3490
+ "learning_rate": 1.2152777777777776e-06,
3491
+ "loss": 0.0186,
3492
+ "step": 4650
3493
+ },
3494
+ {
3495
+ "epoch": 77.34,
3496
+ "learning_rate": 1.134259259259259e-06,
3497
+ "loss": 0.0096,
3498
+ "step": 4660
3499
+ },
3500
+ {
3501
+ "epoch": 77.51,
3502
+ "learning_rate": 1.0532407407407406e-06,
3503
+ "loss": 0.0015,
3504
+ "step": 4670
3505
+ },
3506
+ {
3507
+ "epoch": 77.68,
3508
+ "learning_rate": 9.722222222222222e-07,
3509
+ "loss": 0.005,
3510
+ "step": 4680
3511
+ },
3512
+ {
3513
+ "epoch": 77.84,
3514
+ "learning_rate": 8.912037037037036e-07,
3515
+ "loss": 0.0229,
3516
+ "step": 4690
3517
+ },
3518
+ {
3519
+ "epoch": 77.99,
3520
+ "eval_accuracy": 0.8043478260869565,
3521
+ "eval_loss": 1.6823025941848755,
3522
+ "eval_runtime": 0.6989,
3523
+ "eval_samples_per_second": 65.82,
3524
+ "eval_steps_per_second": 4.293,
3525
+ "step": 4699
3526
+ },
3527
+ {
3528
+ "epoch": 78.01,
3529
+ "learning_rate": 8.10185185185185e-07,
3530
+ "loss": 0.0117,
3531
+ "step": 4700
3532
+ },
3533
+ {
3534
+ "epoch": 78.17,
3535
+ "learning_rate": 7.291666666666665e-07,
3536
+ "loss": 0.0014,
3537
+ "step": 4710
3538
+ },
3539
+ {
3540
+ "epoch": 78.34,
3541
+ "learning_rate": 6.48148148148148e-07,
3542
+ "loss": 0.0039,
3543
+ "step": 4720
3544
+ },
3545
+ {
3546
+ "epoch": 78.51,
3547
+ "learning_rate": 5.671296296296295e-07,
3548
+ "loss": 0.0111,
3549
+ "step": 4730
3550
+ },
3551
+ {
3552
+ "epoch": 78.67,
3553
+ "learning_rate": 4.861111111111111e-07,
3554
+ "loss": 0.0129,
3555
+ "step": 4740
3556
+ },
3557
+ {
3558
+ "epoch": 78.84,
3559
+ "learning_rate": 4.050925925925925e-07,
3560
+ "loss": 0.008,
3561
+ "step": 4750
3562
+ },
3563
+ {
3564
+ "epoch": 78.99,
3565
+ "eval_accuracy": 0.8043478260869565,
3566
+ "eval_loss": 1.6750718355178833,
3567
+ "eval_runtime": 0.6921,
3568
+ "eval_samples_per_second": 66.467,
3569
+ "eval_steps_per_second": 4.335,
3570
+ "step": 4759
3571
+ },
3572
+ {
3573
+ "epoch": 79.0,
3574
+ "learning_rate": 3.24074074074074e-07,
3575
+ "loss": 0.0032,
3576
+ "step": 4760
3577
+ },
3578
+ {
3579
+ "epoch": 79.17,
3580
+ "learning_rate": 2.4305555555555555e-07,
3581
+ "loss": 0.0079,
3582
+ "step": 4770
3583
+ },
3584
+ {
3585
+ "epoch": 79.34,
3586
+ "learning_rate": 1.62037037037037e-07,
3587
+ "loss": 0.008,
3588
+ "step": 4780
3589
+ },
3590
+ {
3591
+ "epoch": 79.5,
3592
+ "learning_rate": 8.10185185185185e-08,
3593
+ "loss": 0.0057,
3594
+ "step": 4790
3595
+ },
3596
+ {
3597
+ "epoch": 79.67,
3598
+ "learning_rate": 0.0,
3599
+ "loss": 0.0051,
3600
+ "step": 4800
3601
+ },
3602
+ {
3603
+ "epoch": 79.67,
3604
+ "eval_accuracy": 0.8043478260869565,
3605
+ "eval_loss": 1.675818681716919,
3606
+ "eval_runtime": 0.6446,
3607
+ "eval_samples_per_second": 71.357,
3608
+ "eval_steps_per_second": 4.654,
3609
+ "step": 4800
3610
+ },
3611
+ {
3612
+ "epoch": 79.67,
3613
+ "step": 4800,
3614
+ "total_flos": 2.3777433087459287e+19,
3615
+ "train_loss": 0.12199072747869649,
3616
+ "train_runtime": 7091.3369,
3617
+ "train_samples_per_second": 43.467,
3618
+ "train_steps_per_second": 0.677
3619
+ }
3620
+ ],
3621
+ "logging_steps": 10,
3622
+ "max_steps": 4800,
3623
+ "num_input_tokens_seen": 0,
3624
+ "num_train_epochs": 80,
3625
+ "save_steps": 500,
3626
+ "total_flos": 2.3777433087459287e+19,
3627
+ "train_batch_size": 16,
3628
+ "trial_name": null,
3629
+ "trial_params": null
3630
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1cd3fa58ea61f8a33971ebc2d2aec0e08e3cf3a5cfd744a4341c50305fff40a9
3
+ size 4728