spycoder commited on
Commit
13b148c
1 Parent(s): 590e6da
README.md ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: google/vit-base-patch16-224-in21k
4
+ tags:
5
+ - image-classification
6
+ - generated_from_trainer
7
+ metrics:
8
+ - accuracy
9
+ model-index:
10
+ - name: vit-base-beans
11
+ results: []
12
+ ---
13
+
14
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
+ should probably proofread and complete it, then remove this comment. -->
16
+
17
+ # vit-base-beans
18
+
19
+ This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the ahishamm/HAM_db_enhanced_balanced_reduced_50_20_20_50 dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 0.5305
22
+ - Accuracy: 0.8451
23
+
24
+ ## Model description
25
+
26
+ More information needed
27
+
28
+ ## Intended uses & limitations
29
+
30
+ More information needed
31
+
32
+ ## Training and evaluation data
33
+
34
+ More information needed
35
+
36
+ ## Training procedure
37
+
38
+ ### Training hyperparameters
39
+
40
+ The following hyperparameters were used during training:
41
+ - learning_rate: 0.0002
42
+ - train_batch_size: 16
43
+ - eval_batch_size: 8
44
+ - seed: 42
45
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
46
+ - lr_scheduler_type: linear
47
+ - num_epochs: 4
48
+ - mixed_precision_training: Native AMP
49
+
50
+ ### Training results
51
+
52
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy |
53
+ |:-------------:|:------:|:----:|:---------------:|:--------:|
54
+ | 1.0791 | 0.2304 | 100 | 1.0348 | 0.6335 |
55
+ | 0.9415 | 0.4608 | 200 | 0.9576 | 0.6449 |
56
+ | 0.7839 | 0.6912 | 300 | 0.8963 | 0.6662 |
57
+ | 0.7181 | 0.9217 | 400 | 0.8479 | 0.6963 |
58
+ | 0.3995 | 1.1521 | 500 | 0.7821 | 0.7170 |
59
+ | 0.5025 | 1.3825 | 600 | 0.6300 | 0.7837 |
60
+ | 0.4985 | 1.6129 | 700 | 0.7059 | 0.7490 |
61
+ | 0.4388 | 1.8433 | 800 | 0.5893 | 0.7857 |
62
+ | 0.2389 | 2.0737 | 900 | 0.5929 | 0.8077 |
63
+ | 0.2767 | 2.3041 | 1000 | 0.5795 | 0.8091 |
64
+ | 0.2387 | 2.5346 | 1100 | 0.6100 | 0.8091 |
65
+ | 0.1691 | 2.7650 | 1200 | 0.6175 | 0.8071 |
66
+ | 0.1738 | 2.9954 | 1300 | 0.5877 | 0.8198 |
67
+ | 0.0397 | 3.2258 | 1400 | 0.5766 | 0.8358 |
68
+ | 0.03 | 3.4562 | 1500 | 0.5681 | 0.8371 |
69
+ | 0.092 | 3.6866 | 1600 | 0.5305 | 0.8451 |
70
+ | 0.0416 | 3.9171 | 1700 | 0.5443 | 0.8471 |
71
+
72
+
73
+ ### Framework versions
74
+
75
+ - Transformers 4.41.2
76
+ - Pytorch 2.3.0+cu121
77
+ - Datasets 2.20.0
78
+ - Tokenizers 0.19.1
all_results.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.0,
3
+ "eval_accuracy": 0.8451268357810414,
4
+ "eval_loss": 0.5304816365242004,
5
+ "eval_runtime": 12.1921,
6
+ "eval_samples_per_second": 122.867,
7
+ "eval_steps_per_second": 15.42,
8
+ "total_flos": 2.1525139607212524e+18,
9
+ "train_loss": 0.4232822818690181,
10
+ "train_runtime": 559.5799,
11
+ "train_samples_per_second": 49.637,
12
+ "train_steps_per_second": 3.102
13
+ }
config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/vit-base-patch16-224-in21k",
3
+ "architectures": [
4
+ "ViTForImageClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "encoder_stride": 16,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.0,
10
+ "hidden_size": 768,
11
+ "id2label": {
12
+ "0": "akiec",
13
+ "1": "bcc",
14
+ "2": "bkl",
15
+ "3": "df",
16
+ "4": "mel",
17
+ "5": "nv",
18
+ "6": "vasc"
19
+ },
20
+ "image_size": 224,
21
+ "initializer_range": 0.02,
22
+ "intermediate_size": 3072,
23
+ "label2id": {
24
+ "akiec": "0",
25
+ "bcc": "1",
26
+ "bkl": "2",
27
+ "df": "3",
28
+ "mel": "4",
29
+ "nv": "5",
30
+ "vasc": "6"
31
+ },
32
+ "layer_norm_eps": 1e-12,
33
+ "model_type": "vit",
34
+ "num_attention_heads": 12,
35
+ "num_channels": 3,
36
+ "num_hidden_layers": 12,
37
+ "patch_size": 16,
38
+ "problem_type": "single_label_classification",
39
+ "qkv_bias": true,
40
+ "torch_dtype": "float32",
41
+ "transformers_version": "4.41.2"
42
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.0,
3
+ "eval_accuracy": 0.8451268357810414,
4
+ "eval_loss": 0.5304816365242004,
5
+ "eval_runtime": 12.1921,
6
+ "eval_samples_per_second": 122.867,
7
+ "eval_steps_per_second": 15.42
8
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c50d54887b112813f65a3f6a29d4f8ab8283e6ebdc2346ae5bfd8a698f082669
3
+ size 343239356
preprocessor_config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_valid_processor_keys": [
3
+ "images",
4
+ "do_resize",
5
+ "size",
6
+ "resample",
7
+ "do_rescale",
8
+ "rescale_factor",
9
+ "do_normalize",
10
+ "image_mean",
11
+ "image_std",
12
+ "return_tensors",
13
+ "data_format",
14
+ "input_data_format"
15
+ ],
16
+ "do_normalize": true,
17
+ "do_rescale": true,
18
+ "do_resize": true,
19
+ "image_mean": [
20
+ 0.5,
21
+ 0.5,
22
+ 0.5
23
+ ],
24
+ "image_processor_type": "ViTImageProcessor",
25
+ "image_std": [
26
+ 0.5,
27
+ 0.5,
28
+ 0.5
29
+ ],
30
+ "resample": 2,
31
+ "rescale_factor": 0.00392156862745098,
32
+ "size": {
33
+ "height": 224,
34
+ "width": 224
35
+ }
36
+ }
runs/Jun17_09-47-11_d38a0241bb1c/events.out.tfevents.1718617637.d38a0241bb1c.6107.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a3fd03c2aa467f2e20416689acc759f9b9af78d90fc73186bca8b1771f738345
3
+ size 47243
runs/Jun17_09-47-11_d38a0241bb1c/events.out.tfevents.1718618209.d38a0241bb1c.6107.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4eb402f867490e011e329ea5ab7a2503a8cbf8bb69eea1d1b36ccb71c38e540b
3
+ size 734
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.0,
3
+ "total_flos": 2.1525139607212524e+18,
4
+ "train_loss": 0.4232822818690181,
5
+ "train_runtime": 559.5799,
6
+ "train_samples_per_second": 49.637,
7
+ "train_steps_per_second": 3.102
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1406 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.5304816365242004,
3
+ "best_model_checkpoint": "./vit-base-beans/checkpoint-1600",
4
+ "epoch": 4.0,
5
+ "eval_steps": 100,
6
+ "global_step": 1736,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.02304147465437788,
13
+ "grad_norm": 2.396202564239502,
14
+ "learning_rate": 0.00019884792626728113,
15
+ "loss": 1.8485,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.04608294930875576,
20
+ "grad_norm": 1.289166808128357,
21
+ "learning_rate": 0.00019769585253456222,
22
+ "loss": 1.5911,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.06912442396313365,
27
+ "grad_norm": 2.512033462524414,
28
+ "learning_rate": 0.00019654377880184333,
29
+ "loss": 1.4806,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.09216589861751152,
34
+ "grad_norm": 2.6234657764434814,
35
+ "learning_rate": 0.00019539170506912442,
36
+ "loss": 1.3684,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.1152073732718894,
41
+ "grad_norm": 2.335149049758911,
42
+ "learning_rate": 0.00019423963133640554,
43
+ "loss": 1.4012,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.1382488479262673,
48
+ "grad_norm": 3.386568546295166,
49
+ "learning_rate": 0.00019308755760368663,
50
+ "loss": 1.2248,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.16129032258064516,
55
+ "grad_norm": 1.9273797273635864,
56
+ "learning_rate": 0.00019193548387096775,
57
+ "loss": 1.144,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.18433179723502305,
62
+ "grad_norm": 2.2117414474487305,
63
+ "learning_rate": 0.00019078341013824886,
64
+ "loss": 1.0101,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.2073732718894009,
69
+ "grad_norm": 3.1132171154022217,
70
+ "learning_rate": 0.00018963133640552998,
71
+ "loss": 1.1411,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 0.2304147465437788,
76
+ "grad_norm": 3.0585570335388184,
77
+ "learning_rate": 0.00018847926267281107,
78
+ "loss": 1.0791,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 0.2304147465437788,
83
+ "eval_accuracy": 0.6335113484646195,
84
+ "eval_loss": 1.0347875356674194,
85
+ "eval_runtime": 11.9052,
86
+ "eval_samples_per_second": 125.828,
87
+ "eval_steps_per_second": 15.791,
88
+ "step": 100
89
+ },
90
+ {
91
+ "epoch": 0.2534562211981567,
92
+ "grad_norm": 2.400747299194336,
93
+ "learning_rate": 0.00018732718894009219,
94
+ "loss": 1.04,
95
+ "step": 110
96
+ },
97
+ {
98
+ "epoch": 0.2764976958525346,
99
+ "grad_norm": 2.432607412338257,
100
+ "learning_rate": 0.00018617511520737328,
101
+ "loss": 1.0396,
102
+ "step": 120
103
+ },
104
+ {
105
+ "epoch": 0.2995391705069124,
106
+ "grad_norm": 2.5169568061828613,
107
+ "learning_rate": 0.0001850230414746544,
108
+ "loss": 0.9925,
109
+ "step": 130
110
+ },
111
+ {
112
+ "epoch": 0.3225806451612903,
113
+ "grad_norm": 2.450554847717285,
114
+ "learning_rate": 0.00018387096774193548,
115
+ "loss": 1.0361,
116
+ "step": 140
117
+ },
118
+ {
119
+ "epoch": 0.3456221198156682,
120
+ "grad_norm": 1.5931885242462158,
121
+ "learning_rate": 0.0001827188940092166,
122
+ "loss": 0.9851,
123
+ "step": 150
124
+ },
125
+ {
126
+ "epoch": 0.3686635944700461,
127
+ "grad_norm": 1.8019052743911743,
128
+ "learning_rate": 0.0001815668202764977,
129
+ "loss": 0.8847,
130
+ "step": 160
131
+ },
132
+ {
133
+ "epoch": 0.391705069124424,
134
+ "grad_norm": 2.283034086227417,
135
+ "learning_rate": 0.0001804147465437788,
136
+ "loss": 0.8507,
137
+ "step": 170
138
+ },
139
+ {
140
+ "epoch": 0.4147465437788018,
141
+ "grad_norm": 2.5878796577453613,
142
+ "learning_rate": 0.0001792626728110599,
143
+ "loss": 0.9579,
144
+ "step": 180
145
+ },
146
+ {
147
+ "epoch": 0.4377880184331797,
148
+ "grad_norm": 3.469618558883667,
149
+ "learning_rate": 0.000178110599078341,
150
+ "loss": 0.9453,
151
+ "step": 190
152
+ },
153
+ {
154
+ "epoch": 0.4608294930875576,
155
+ "grad_norm": 1.9743025302886963,
156
+ "learning_rate": 0.00017695852534562213,
157
+ "loss": 0.9415,
158
+ "step": 200
159
+ },
160
+ {
161
+ "epoch": 0.4608294930875576,
162
+ "eval_accuracy": 0.6448598130841121,
163
+ "eval_loss": 0.9576324820518494,
164
+ "eval_runtime": 11.862,
165
+ "eval_samples_per_second": 126.285,
166
+ "eval_steps_per_second": 15.849,
167
+ "step": 200
168
+ },
169
+ {
170
+ "epoch": 0.4838709677419355,
171
+ "grad_norm": 3.031723976135254,
172
+ "learning_rate": 0.00017580645161290325,
173
+ "loss": 0.7819,
174
+ "step": 210
175
+ },
176
+ {
177
+ "epoch": 0.5069124423963134,
178
+ "grad_norm": 2.2470805644989014,
179
+ "learning_rate": 0.00017465437788018436,
180
+ "loss": 0.8163,
181
+ "step": 220
182
+ },
183
+ {
184
+ "epoch": 0.5299539170506913,
185
+ "grad_norm": 1.723471760749817,
186
+ "learning_rate": 0.00017350230414746545,
187
+ "loss": 0.6728,
188
+ "step": 230
189
+ },
190
+ {
191
+ "epoch": 0.5529953917050692,
192
+ "grad_norm": 3.93212628364563,
193
+ "learning_rate": 0.00017235023041474657,
194
+ "loss": 0.684,
195
+ "step": 240
196
+ },
197
+ {
198
+ "epoch": 0.576036866359447,
199
+ "grad_norm": 1.4867981672286987,
200
+ "learning_rate": 0.00017119815668202766,
201
+ "loss": 0.8527,
202
+ "step": 250
203
+ },
204
+ {
205
+ "epoch": 0.5990783410138248,
206
+ "grad_norm": 2.4340641498565674,
207
+ "learning_rate": 0.00017004608294930878,
208
+ "loss": 1.0102,
209
+ "step": 260
210
+ },
211
+ {
212
+ "epoch": 0.6221198156682027,
213
+ "grad_norm": 2.8441660404205322,
214
+ "learning_rate": 0.00016889400921658987,
215
+ "loss": 0.7739,
216
+ "step": 270
217
+ },
218
+ {
219
+ "epoch": 0.6451612903225806,
220
+ "grad_norm": 1.6598294973373413,
221
+ "learning_rate": 0.00016774193548387098,
222
+ "loss": 0.7442,
223
+ "step": 280
224
+ },
225
+ {
226
+ "epoch": 0.6682027649769585,
227
+ "grad_norm": 3.455202102661133,
228
+ "learning_rate": 0.00016658986175115207,
229
+ "loss": 0.7643,
230
+ "step": 290
231
+ },
232
+ {
233
+ "epoch": 0.6912442396313364,
234
+ "grad_norm": 2.480116367340088,
235
+ "learning_rate": 0.0001654377880184332,
236
+ "loss": 0.7839,
237
+ "step": 300
238
+ },
239
+ {
240
+ "epoch": 0.6912442396313364,
241
+ "eval_accuracy": 0.6662216288384513,
242
+ "eval_loss": 0.89629727602005,
243
+ "eval_runtime": 11.7103,
244
+ "eval_samples_per_second": 127.921,
245
+ "eval_steps_per_second": 16.054,
246
+ "step": 300
247
+ },
248
+ {
249
+ "epoch": 0.7142857142857143,
250
+ "grad_norm": 3.3055620193481445,
251
+ "learning_rate": 0.00016428571428571428,
252
+ "loss": 0.639,
253
+ "step": 310
254
+ },
255
+ {
256
+ "epoch": 0.7373271889400922,
257
+ "grad_norm": 1.8542070388793945,
258
+ "learning_rate": 0.0001631336405529954,
259
+ "loss": 0.8931,
260
+ "step": 320
261
+ },
262
+ {
263
+ "epoch": 0.7603686635944701,
264
+ "grad_norm": 1.6089766025543213,
265
+ "learning_rate": 0.00016198156682027649,
266
+ "loss": 0.9023,
267
+ "step": 330
268
+ },
269
+ {
270
+ "epoch": 0.783410138248848,
271
+ "grad_norm": 1.5780836343765259,
272
+ "learning_rate": 0.0001608294930875576,
273
+ "loss": 0.7285,
274
+ "step": 340
275
+ },
276
+ {
277
+ "epoch": 0.8064516129032258,
278
+ "grad_norm": 3.153092384338379,
279
+ "learning_rate": 0.00015967741935483872,
280
+ "loss": 0.8702,
281
+ "step": 350
282
+ },
283
+ {
284
+ "epoch": 0.8294930875576036,
285
+ "grad_norm": 2.3161656856536865,
286
+ "learning_rate": 0.00015852534562211984,
287
+ "loss": 0.7343,
288
+ "step": 360
289
+ },
290
+ {
291
+ "epoch": 0.8525345622119815,
292
+ "grad_norm": 1.7923251390457153,
293
+ "learning_rate": 0.00015737327188940093,
294
+ "loss": 0.7986,
295
+ "step": 370
296
+ },
297
+ {
298
+ "epoch": 0.8755760368663594,
299
+ "grad_norm": 2.7093405723571777,
300
+ "learning_rate": 0.00015622119815668204,
301
+ "loss": 0.6377,
302
+ "step": 380
303
+ },
304
+ {
305
+ "epoch": 0.8986175115207373,
306
+ "grad_norm": 4.7555251121521,
307
+ "learning_rate": 0.00015506912442396313,
308
+ "loss": 0.8223,
309
+ "step": 390
310
+ },
311
+ {
312
+ "epoch": 0.9216589861751152,
313
+ "grad_norm": 2.78916072845459,
314
+ "learning_rate": 0.00015391705069124425,
315
+ "loss": 0.7181,
316
+ "step": 400
317
+ },
318
+ {
319
+ "epoch": 0.9216589861751152,
320
+ "eval_accuracy": 0.6962616822429907,
321
+ "eval_loss": 0.8479276299476624,
322
+ "eval_runtime": 11.6609,
323
+ "eval_samples_per_second": 128.464,
324
+ "eval_steps_per_second": 16.122,
325
+ "step": 400
326
+ },
327
+ {
328
+ "epoch": 0.9447004608294931,
329
+ "grad_norm": 2.4783871173858643,
330
+ "learning_rate": 0.00015276497695852537,
331
+ "loss": 0.7422,
332
+ "step": 410
333
+ },
334
+ {
335
+ "epoch": 0.967741935483871,
336
+ "grad_norm": 2.8775382041931152,
337
+ "learning_rate": 0.00015161290322580646,
338
+ "loss": 0.6255,
339
+ "step": 420
340
+ },
341
+ {
342
+ "epoch": 0.9907834101382489,
343
+ "grad_norm": 2.3851194381713867,
344
+ "learning_rate": 0.00015046082949308757,
345
+ "loss": 0.7266,
346
+ "step": 430
347
+ },
348
+ {
349
+ "epoch": 1.0138248847926268,
350
+ "grad_norm": 5.285385608673096,
351
+ "learning_rate": 0.00014930875576036866,
352
+ "loss": 0.6283,
353
+ "step": 440
354
+ },
355
+ {
356
+ "epoch": 1.0368663594470047,
357
+ "grad_norm": 1.691789984703064,
358
+ "learning_rate": 0.00014815668202764978,
359
+ "loss": 0.4918,
360
+ "step": 450
361
+ },
362
+ {
363
+ "epoch": 1.0599078341013826,
364
+ "grad_norm": 2.8921382427215576,
365
+ "learning_rate": 0.00014700460829493087,
366
+ "loss": 0.5787,
367
+ "step": 460
368
+ },
369
+ {
370
+ "epoch": 1.0829493087557605,
371
+ "grad_norm": 3.1509757041931152,
372
+ "learning_rate": 0.00014585253456221199,
373
+ "loss": 0.4906,
374
+ "step": 470
375
+ },
376
+ {
377
+ "epoch": 1.1059907834101383,
378
+ "grad_norm": 3.2979822158813477,
379
+ "learning_rate": 0.0001447004608294931,
380
+ "loss": 0.5715,
381
+ "step": 480
382
+ },
383
+ {
384
+ "epoch": 1.129032258064516,
385
+ "grad_norm": 3.3389899730682373,
386
+ "learning_rate": 0.00014354838709677422,
387
+ "loss": 0.5411,
388
+ "step": 490
389
+ },
390
+ {
391
+ "epoch": 1.1520737327188941,
392
+ "grad_norm": 0.9589664936065674,
393
+ "learning_rate": 0.0001423963133640553,
394
+ "loss": 0.3995,
395
+ "step": 500
396
+ },
397
+ {
398
+ "epoch": 1.1520737327188941,
399
+ "eval_accuracy": 0.7169559412550067,
400
+ "eval_loss": 0.7820530533790588,
401
+ "eval_runtime": 11.5056,
402
+ "eval_samples_per_second": 130.197,
403
+ "eval_steps_per_second": 16.34,
404
+ "step": 500
405
+ },
406
+ {
407
+ "epoch": 1.1751152073732718,
408
+ "grad_norm": 2.248042106628418,
409
+ "learning_rate": 0.00014124423963133643,
410
+ "loss": 0.5057,
411
+ "step": 510
412
+ },
413
+ {
414
+ "epoch": 1.1981566820276497,
415
+ "grad_norm": 3.944963216781616,
416
+ "learning_rate": 0.00014009216589861752,
417
+ "loss": 0.5005,
418
+ "step": 520
419
+ },
420
+ {
421
+ "epoch": 1.2211981566820276,
422
+ "grad_norm": 2.7981412410736084,
423
+ "learning_rate": 0.00013894009216589863,
424
+ "loss": 0.6703,
425
+ "step": 530
426
+ },
427
+ {
428
+ "epoch": 1.2442396313364055,
429
+ "grad_norm": 1.683069109916687,
430
+ "learning_rate": 0.00013778801843317972,
431
+ "loss": 0.5394,
432
+ "step": 540
433
+ },
434
+ {
435
+ "epoch": 1.2672811059907834,
436
+ "grad_norm": 1.2122957706451416,
437
+ "learning_rate": 0.00013663594470046084,
438
+ "loss": 0.4775,
439
+ "step": 550
440
+ },
441
+ {
442
+ "epoch": 1.2903225806451613,
443
+ "grad_norm": 1.4005225896835327,
444
+ "learning_rate": 0.00013548387096774193,
445
+ "loss": 0.4467,
446
+ "step": 560
447
+ },
448
+ {
449
+ "epoch": 1.3133640552995391,
450
+ "grad_norm": 2.5969114303588867,
451
+ "learning_rate": 0.00013433179723502305,
452
+ "loss": 0.4289,
453
+ "step": 570
454
+ },
455
+ {
456
+ "epoch": 1.336405529953917,
457
+ "grad_norm": 3.344553232192993,
458
+ "learning_rate": 0.00013317972350230414,
459
+ "loss": 0.4631,
460
+ "step": 580
461
+ },
462
+ {
463
+ "epoch": 1.359447004608295,
464
+ "grad_norm": 1.6798585653305054,
465
+ "learning_rate": 0.00013202764976958525,
466
+ "loss": 0.4329,
467
+ "step": 590
468
+ },
469
+ {
470
+ "epoch": 1.3824884792626728,
471
+ "grad_norm": 1.3849396705627441,
472
+ "learning_rate": 0.00013087557603686637,
473
+ "loss": 0.5025,
474
+ "step": 600
475
+ },
476
+ {
477
+ "epoch": 1.3824884792626728,
478
+ "eval_accuracy": 0.7837116154873164,
479
+ "eval_loss": 0.6299713253974915,
480
+ "eval_runtime": 11.705,
481
+ "eval_samples_per_second": 127.979,
482
+ "eval_steps_per_second": 16.061,
483
+ "step": 600
484
+ },
485
+ {
486
+ "epoch": 1.4055299539170507,
487
+ "grad_norm": 2.550548791885376,
488
+ "learning_rate": 0.00012972350230414746,
489
+ "loss": 0.4463,
490
+ "step": 610
491
+ },
492
+ {
493
+ "epoch": 1.4285714285714286,
494
+ "grad_norm": 3.063411235809326,
495
+ "learning_rate": 0.00012857142857142858,
496
+ "loss": 0.3624,
497
+ "step": 620
498
+ },
499
+ {
500
+ "epoch": 1.4516129032258065,
501
+ "grad_norm": 6.676961898803711,
502
+ "learning_rate": 0.0001274193548387097,
503
+ "loss": 0.4446,
504
+ "step": 630
505
+ },
506
+ {
507
+ "epoch": 1.4746543778801844,
508
+ "grad_norm": 0.8720624446868896,
509
+ "learning_rate": 0.0001262672811059908,
510
+ "loss": 0.5162,
511
+ "step": 640
512
+ },
513
+ {
514
+ "epoch": 1.4976958525345623,
515
+ "grad_norm": 2.214848041534424,
516
+ "learning_rate": 0.0001251152073732719,
517
+ "loss": 0.2978,
518
+ "step": 650
519
+ },
520
+ {
521
+ "epoch": 1.52073732718894,
522
+ "grad_norm": 5.083272457122803,
523
+ "learning_rate": 0.00012396313364055302,
524
+ "loss": 0.5157,
525
+ "step": 660
526
+ },
527
+ {
528
+ "epoch": 1.543778801843318,
529
+ "grad_norm": 4.042588710784912,
530
+ "learning_rate": 0.0001228110599078341,
531
+ "loss": 0.5338,
532
+ "step": 670
533
+ },
534
+ {
535
+ "epoch": 1.5668202764976957,
536
+ "grad_norm": 3.1029160022735596,
537
+ "learning_rate": 0.00012165898617511522,
538
+ "loss": 0.4767,
539
+ "step": 680
540
+ },
541
+ {
542
+ "epoch": 1.5898617511520738,
543
+ "grad_norm": 1.4430710077285767,
544
+ "learning_rate": 0.00012050691244239631,
545
+ "loss": 0.5531,
546
+ "step": 690
547
+ },
548
+ {
549
+ "epoch": 1.6129032258064515,
550
+ "grad_norm": 11.178030967712402,
551
+ "learning_rate": 0.00011935483870967743,
552
+ "loss": 0.4985,
553
+ "step": 700
554
+ },
555
+ {
556
+ "epoch": 1.6129032258064515,
557
+ "eval_accuracy": 0.7489986648865153,
558
+ "eval_loss": 0.7058817744255066,
559
+ "eval_runtime": 11.9139,
560
+ "eval_samples_per_second": 125.736,
561
+ "eval_steps_per_second": 15.78,
562
+ "step": 700
563
+ },
564
+ {
565
+ "epoch": 1.6359447004608296,
566
+ "grad_norm": 3.918297529220581,
567
+ "learning_rate": 0.00011820276497695852,
568
+ "loss": 0.5471,
569
+ "step": 710
570
+ },
571
+ {
572
+ "epoch": 1.6589861751152073,
573
+ "grad_norm": 2.7170467376708984,
574
+ "learning_rate": 0.00011705069124423964,
575
+ "loss": 0.4797,
576
+ "step": 720
577
+ },
578
+ {
579
+ "epoch": 1.6820276497695854,
580
+ "grad_norm": 1.0436949729919434,
581
+ "learning_rate": 0.00011589861751152074,
582
+ "loss": 0.427,
583
+ "step": 730
584
+ },
585
+ {
586
+ "epoch": 1.705069124423963,
587
+ "grad_norm": 3.6829638481140137,
588
+ "learning_rate": 0.00011474654377880186,
589
+ "loss": 0.5121,
590
+ "step": 740
591
+ },
592
+ {
593
+ "epoch": 1.728110599078341,
594
+ "grad_norm": 1.8748345375061035,
595
+ "learning_rate": 0.00011359447004608295,
596
+ "loss": 0.4227,
597
+ "step": 750
598
+ },
599
+ {
600
+ "epoch": 1.7511520737327189,
601
+ "grad_norm": 4.548758506774902,
602
+ "learning_rate": 0.00011244239631336406,
603
+ "loss": 0.3164,
604
+ "step": 760
605
+ },
606
+ {
607
+ "epoch": 1.7741935483870968,
608
+ "grad_norm": 3.4847280979156494,
609
+ "learning_rate": 0.00011129032258064515,
610
+ "loss": 0.5092,
611
+ "step": 770
612
+ },
613
+ {
614
+ "epoch": 1.7972350230414746,
615
+ "grad_norm": 1.8869714736938477,
616
+ "learning_rate": 0.00011013824884792627,
617
+ "loss": 0.4472,
618
+ "step": 780
619
+ },
620
+ {
621
+ "epoch": 1.8202764976958525,
622
+ "grad_norm": 3.899409770965576,
623
+ "learning_rate": 0.00010898617511520739,
624
+ "loss": 0.4708,
625
+ "step": 790
626
+ },
627
+ {
628
+ "epoch": 1.8433179723502304,
629
+ "grad_norm": 1.543060541152954,
630
+ "learning_rate": 0.00010783410138248849,
631
+ "loss": 0.4388,
632
+ "step": 800
633
+ },
634
+ {
635
+ "epoch": 1.8433179723502304,
636
+ "eval_accuracy": 0.7857142857142857,
637
+ "eval_loss": 0.5893343091011047,
638
+ "eval_runtime": 11.4174,
639
+ "eval_samples_per_second": 131.203,
640
+ "eval_steps_per_second": 16.466,
641
+ "step": 800
642
+ },
643
+ {
644
+ "epoch": 1.8663594470046083,
645
+ "grad_norm": 5.587724208831787,
646
+ "learning_rate": 0.0001066820276497696,
647
+ "loss": 0.4264,
648
+ "step": 810
649
+ },
650
+ {
651
+ "epoch": 1.8894009216589862,
652
+ "grad_norm": 7.794037342071533,
653
+ "learning_rate": 0.0001055299539170507,
654
+ "loss": 0.4513,
655
+ "step": 820
656
+ },
657
+ {
658
+ "epoch": 1.912442396313364,
659
+ "grad_norm": 3.597796678543091,
660
+ "learning_rate": 0.00010437788018433181,
661
+ "loss": 0.437,
662
+ "step": 830
663
+ },
664
+ {
665
+ "epoch": 1.935483870967742,
666
+ "grad_norm": 2.825336217880249,
667
+ "learning_rate": 0.0001032258064516129,
668
+ "loss": 0.5202,
669
+ "step": 840
670
+ },
671
+ {
672
+ "epoch": 1.9585253456221197,
673
+ "grad_norm": 1.8002281188964844,
674
+ "learning_rate": 0.00010207373271889402,
675
+ "loss": 0.3283,
676
+ "step": 850
677
+ },
678
+ {
679
+ "epoch": 1.9815668202764978,
680
+ "grad_norm": 6.496976375579834,
681
+ "learning_rate": 0.00010092165898617512,
682
+ "loss": 0.2887,
683
+ "step": 860
684
+ },
685
+ {
686
+ "epoch": 2.0046082949308754,
687
+ "grad_norm": 2.1674392223358154,
688
+ "learning_rate": 9.976958525345623e-05,
689
+ "loss": 0.3299,
690
+ "step": 870
691
+ },
692
+ {
693
+ "epoch": 2.0276497695852536,
694
+ "grad_norm": 0.475057989358902,
695
+ "learning_rate": 9.861751152073733e-05,
696
+ "loss": 0.2049,
697
+ "step": 880
698
+ },
699
+ {
700
+ "epoch": 2.0506912442396312,
701
+ "grad_norm": 2.232353687286377,
702
+ "learning_rate": 9.746543778801845e-05,
703
+ "loss": 0.2598,
704
+ "step": 890
705
+ },
706
+ {
707
+ "epoch": 2.0737327188940093,
708
+ "grad_norm": 3.595874309539795,
709
+ "learning_rate": 9.631336405529955e-05,
710
+ "loss": 0.2389,
711
+ "step": 900
712
+ },
713
+ {
714
+ "epoch": 2.0737327188940093,
715
+ "eval_accuracy": 0.807743658210948,
716
+ "eval_loss": 0.5928804278373718,
717
+ "eval_runtime": 11.7831,
718
+ "eval_samples_per_second": 127.131,
719
+ "eval_steps_per_second": 15.955,
720
+ "step": 900
721
+ },
722
+ {
723
+ "epoch": 2.096774193548387,
724
+ "grad_norm": 2.4027860164642334,
725
+ "learning_rate": 9.516129032258065e-05,
726
+ "loss": 0.2023,
727
+ "step": 910
728
+ },
729
+ {
730
+ "epoch": 2.119815668202765,
731
+ "grad_norm": 4.1582560539245605,
732
+ "learning_rate": 9.400921658986176e-05,
733
+ "loss": 0.2389,
734
+ "step": 920
735
+ },
736
+ {
737
+ "epoch": 2.142857142857143,
738
+ "grad_norm": 3.8105199337005615,
739
+ "learning_rate": 9.285714285714286e-05,
740
+ "loss": 0.2054,
741
+ "step": 930
742
+ },
743
+ {
744
+ "epoch": 2.165898617511521,
745
+ "grad_norm": 4.042884826660156,
746
+ "learning_rate": 9.170506912442398e-05,
747
+ "loss": 0.2445,
748
+ "step": 940
749
+ },
750
+ {
751
+ "epoch": 2.1889400921658986,
752
+ "grad_norm": 3.3385071754455566,
753
+ "learning_rate": 9.055299539170508e-05,
754
+ "loss": 0.2578,
755
+ "step": 950
756
+ },
757
+ {
758
+ "epoch": 2.2119815668202767,
759
+ "grad_norm": 2.232977867126465,
760
+ "learning_rate": 8.940092165898618e-05,
761
+ "loss": 0.2168,
762
+ "step": 960
763
+ },
764
+ {
765
+ "epoch": 2.2350230414746544,
766
+ "grad_norm": 4.8774847984313965,
767
+ "learning_rate": 8.824884792626729e-05,
768
+ "loss": 0.1978,
769
+ "step": 970
770
+ },
771
+ {
772
+ "epoch": 2.258064516129032,
773
+ "grad_norm": 2.6131808757781982,
774
+ "learning_rate": 8.709677419354839e-05,
775
+ "loss": 0.223,
776
+ "step": 980
777
+ },
778
+ {
779
+ "epoch": 2.28110599078341,
780
+ "grad_norm": 1.6126481294631958,
781
+ "learning_rate": 8.594470046082949e-05,
782
+ "loss": 0.3882,
783
+ "step": 990
784
+ },
785
+ {
786
+ "epoch": 2.3041474654377883,
787
+ "grad_norm": 1.6977124214172363,
788
+ "learning_rate": 8.479262672811061e-05,
789
+ "loss": 0.2767,
790
+ "step": 1000
791
+ },
792
+ {
793
+ "epoch": 2.3041474654377883,
794
+ "eval_accuracy": 0.8090787716955942,
795
+ "eval_loss": 0.5795237421989441,
796
+ "eval_runtime": 11.3869,
797
+ "eval_samples_per_second": 131.555,
798
+ "eval_steps_per_second": 16.51,
799
+ "step": 1000
800
+ },
801
+ {
802
+ "epoch": 2.327188940092166,
803
+ "grad_norm": 5.384529113769531,
804
+ "learning_rate": 8.364055299539171e-05,
805
+ "loss": 0.2478,
806
+ "step": 1010
807
+ },
808
+ {
809
+ "epoch": 2.3502304147465436,
810
+ "grad_norm": 7.527071952819824,
811
+ "learning_rate": 8.248847926267282e-05,
812
+ "loss": 0.1614,
813
+ "step": 1020
814
+ },
815
+ {
816
+ "epoch": 2.3732718894009217,
817
+ "grad_norm": 3.253967523574829,
818
+ "learning_rate": 8.133640552995392e-05,
819
+ "loss": 0.1988,
820
+ "step": 1030
821
+ },
822
+ {
823
+ "epoch": 2.3963133640552994,
824
+ "grad_norm": 2.3061683177948,
825
+ "learning_rate": 8.018433179723502e-05,
826
+ "loss": 0.2267,
827
+ "step": 1040
828
+ },
829
+ {
830
+ "epoch": 2.4193548387096775,
831
+ "grad_norm": 5.240030288696289,
832
+ "learning_rate": 7.903225806451613e-05,
833
+ "loss": 0.3522,
834
+ "step": 1050
835
+ },
836
+ {
837
+ "epoch": 2.442396313364055,
838
+ "grad_norm": 5.367170810699463,
839
+ "learning_rate": 7.788018433179723e-05,
840
+ "loss": 0.21,
841
+ "step": 1060
842
+ },
843
+ {
844
+ "epoch": 2.4654377880184333,
845
+ "grad_norm": 2.52602219581604,
846
+ "learning_rate": 7.672811059907835e-05,
847
+ "loss": 0.208,
848
+ "step": 1070
849
+ },
850
+ {
851
+ "epoch": 2.488479262672811,
852
+ "grad_norm": 3.110276937484741,
853
+ "learning_rate": 7.557603686635945e-05,
854
+ "loss": 0.1624,
855
+ "step": 1080
856
+ },
857
+ {
858
+ "epoch": 2.511520737327189,
859
+ "grad_norm": 3.7577178478240967,
860
+ "learning_rate": 7.442396313364057e-05,
861
+ "loss": 0.2187,
862
+ "step": 1090
863
+ },
864
+ {
865
+ "epoch": 2.5345622119815667,
866
+ "grad_norm": 0.886064887046814,
867
+ "learning_rate": 7.327188940092167e-05,
868
+ "loss": 0.2387,
869
+ "step": 1100
870
+ },
871
+ {
872
+ "epoch": 2.5345622119815667,
873
+ "eval_accuracy": 0.8090787716955942,
874
+ "eval_loss": 0.6099982857704163,
875
+ "eval_runtime": 11.7513,
876
+ "eval_samples_per_second": 127.476,
877
+ "eval_steps_per_second": 15.998,
878
+ "step": 1100
879
+ },
880
+ {
881
+ "epoch": 2.557603686635945,
882
+ "grad_norm": 0.9772585034370422,
883
+ "learning_rate": 7.211981566820277e-05,
884
+ "loss": 0.2289,
885
+ "step": 1110
886
+ },
887
+ {
888
+ "epoch": 2.5806451612903225,
889
+ "grad_norm": 5.879600524902344,
890
+ "learning_rate": 7.096774193548388e-05,
891
+ "loss": 0.2592,
892
+ "step": 1120
893
+ },
894
+ {
895
+ "epoch": 2.6036866359447006,
896
+ "grad_norm": 5.125580310821533,
897
+ "learning_rate": 6.981566820276498e-05,
898
+ "loss": 0.1801,
899
+ "step": 1130
900
+ },
901
+ {
902
+ "epoch": 2.6267281105990783,
903
+ "grad_norm": 4.4502692222595215,
904
+ "learning_rate": 6.86635944700461e-05,
905
+ "loss": 0.3577,
906
+ "step": 1140
907
+ },
908
+ {
909
+ "epoch": 2.6497695852534564,
910
+ "grad_norm": 0.543267548084259,
911
+ "learning_rate": 6.75115207373272e-05,
912
+ "loss": 0.1313,
913
+ "step": 1150
914
+ },
915
+ {
916
+ "epoch": 2.672811059907834,
917
+ "grad_norm": 1.4891630411148071,
918
+ "learning_rate": 6.63594470046083e-05,
919
+ "loss": 0.1858,
920
+ "step": 1160
921
+ },
922
+ {
923
+ "epoch": 2.6958525345622117,
924
+ "grad_norm": 2.359645366668701,
925
+ "learning_rate": 6.52073732718894e-05,
926
+ "loss": 0.2059,
927
+ "step": 1170
928
+ },
929
+ {
930
+ "epoch": 2.71889400921659,
931
+ "grad_norm": 2.5760185718536377,
932
+ "learning_rate": 6.405529953917051e-05,
933
+ "loss": 0.2378,
934
+ "step": 1180
935
+ },
936
+ {
937
+ "epoch": 2.741935483870968,
938
+ "grad_norm": 0.24703356623649597,
939
+ "learning_rate": 6.290322580645161e-05,
940
+ "loss": 0.1487,
941
+ "step": 1190
942
+ },
943
+ {
944
+ "epoch": 2.7649769585253456,
945
+ "grad_norm": 0.22307877242565155,
946
+ "learning_rate": 6.175115207373272e-05,
947
+ "loss": 0.1691,
948
+ "step": 1200
949
+ },
950
+ {
951
+ "epoch": 2.7649769585253456,
952
+ "eval_accuracy": 0.8070761014686249,
953
+ "eval_loss": 0.6174820065498352,
954
+ "eval_runtime": 11.265,
955
+ "eval_samples_per_second": 132.978,
956
+ "eval_steps_per_second": 16.689,
957
+ "step": 1200
958
+ },
959
+ {
960
+ "epoch": 2.7880184331797233,
961
+ "grad_norm": 2.50034761428833,
962
+ "learning_rate": 6.0599078341013825e-05,
963
+ "loss": 0.2148,
964
+ "step": 1210
965
+ },
966
+ {
967
+ "epoch": 2.8110599078341014,
968
+ "grad_norm": 0.3251860439777374,
969
+ "learning_rate": 5.944700460829493e-05,
970
+ "loss": 0.1538,
971
+ "step": 1220
972
+ },
973
+ {
974
+ "epoch": 2.8341013824884795,
975
+ "grad_norm": 3.687969446182251,
976
+ "learning_rate": 5.829493087557604e-05,
977
+ "loss": 0.2445,
978
+ "step": 1230
979
+ },
980
+ {
981
+ "epoch": 2.857142857142857,
982
+ "grad_norm": 7.214417457580566,
983
+ "learning_rate": 5.714285714285714e-05,
984
+ "loss": 0.229,
985
+ "step": 1240
986
+ },
987
+ {
988
+ "epoch": 2.880184331797235,
989
+ "grad_norm": 2.587062120437622,
990
+ "learning_rate": 5.5990783410138245e-05,
991
+ "loss": 0.1999,
992
+ "step": 1250
993
+ },
994
+ {
995
+ "epoch": 2.903225806451613,
996
+ "grad_norm": 4.365920066833496,
997
+ "learning_rate": 5.4838709677419355e-05,
998
+ "loss": 0.1061,
999
+ "step": 1260
1000
+ },
1001
+ {
1002
+ "epoch": 2.9262672811059907,
1003
+ "grad_norm": 3.7295572757720947,
1004
+ "learning_rate": 5.368663594470046e-05,
1005
+ "loss": 0.3093,
1006
+ "step": 1270
1007
+ },
1008
+ {
1009
+ "epoch": 2.9493087557603688,
1010
+ "grad_norm": 2.4992685317993164,
1011
+ "learning_rate": 5.253456221198156e-05,
1012
+ "loss": 0.1644,
1013
+ "step": 1280
1014
+ },
1015
+ {
1016
+ "epoch": 2.9723502304147464,
1017
+ "grad_norm": 5.495995998382568,
1018
+ "learning_rate": 5.138248847926268e-05,
1019
+ "loss": 0.2393,
1020
+ "step": 1290
1021
+ },
1022
+ {
1023
+ "epoch": 2.9953917050691246,
1024
+ "grad_norm": 2.1380579471588135,
1025
+ "learning_rate": 5.023041474654379e-05,
1026
+ "loss": 0.1738,
1027
+ "step": 1300
1028
+ },
1029
+ {
1030
+ "epoch": 2.9953917050691246,
1031
+ "eval_accuracy": 0.8197596795727636,
1032
+ "eval_loss": 0.5877332091331482,
1033
+ "eval_runtime": 11.4089,
1034
+ "eval_samples_per_second": 131.301,
1035
+ "eval_steps_per_second": 16.478,
1036
+ "step": 1300
1037
+ },
1038
+ {
1039
+ "epoch": 3.0184331797235022,
1040
+ "grad_norm": 6.119831085205078,
1041
+ "learning_rate": 4.9078341013824885e-05,
1042
+ "loss": 0.075,
1043
+ "step": 1310
1044
+ },
1045
+ {
1046
+ "epoch": 3.0414746543778803,
1047
+ "grad_norm": 0.25446683168411255,
1048
+ "learning_rate": 4.792626728110599e-05,
1049
+ "loss": 0.0528,
1050
+ "step": 1320
1051
+ },
1052
+ {
1053
+ "epoch": 3.064516129032258,
1054
+ "grad_norm": 0.32773900032043457,
1055
+ "learning_rate": 4.67741935483871e-05,
1056
+ "loss": 0.0551,
1057
+ "step": 1330
1058
+ },
1059
+ {
1060
+ "epoch": 3.087557603686636,
1061
+ "grad_norm": 0.8912816643714905,
1062
+ "learning_rate": 4.562211981566821e-05,
1063
+ "loss": 0.0799,
1064
+ "step": 1340
1065
+ },
1066
+ {
1067
+ "epoch": 3.110599078341014,
1068
+ "grad_norm": 0.6732431054115295,
1069
+ "learning_rate": 4.447004608294931e-05,
1070
+ "loss": 0.0327,
1071
+ "step": 1350
1072
+ },
1073
+ {
1074
+ "epoch": 3.133640552995392,
1075
+ "grad_norm": 5.909882545471191,
1076
+ "learning_rate": 4.3317972350230415e-05,
1077
+ "loss": 0.108,
1078
+ "step": 1360
1079
+ },
1080
+ {
1081
+ "epoch": 3.1566820276497696,
1082
+ "grad_norm": 1.3546661138534546,
1083
+ "learning_rate": 4.2165898617511525e-05,
1084
+ "loss": 0.1057,
1085
+ "step": 1370
1086
+ },
1087
+ {
1088
+ "epoch": 3.1797235023041477,
1089
+ "grad_norm": 0.09205944836139679,
1090
+ "learning_rate": 4.101382488479263e-05,
1091
+ "loss": 0.045,
1092
+ "step": 1380
1093
+ },
1094
+ {
1095
+ "epoch": 3.2027649769585254,
1096
+ "grad_norm": 0.12445586174726486,
1097
+ "learning_rate": 3.986175115207373e-05,
1098
+ "loss": 0.0391,
1099
+ "step": 1390
1100
+ },
1101
+ {
1102
+ "epoch": 3.225806451612903,
1103
+ "grad_norm": 0.49267128109931946,
1104
+ "learning_rate": 3.870967741935484e-05,
1105
+ "loss": 0.0397,
1106
+ "step": 1400
1107
+ },
1108
+ {
1109
+ "epoch": 3.225806451612903,
1110
+ "eval_accuracy": 0.835781041388518,
1111
+ "eval_loss": 0.576629102230072,
1112
+ "eval_runtime": 11.5658,
1113
+ "eval_samples_per_second": 129.52,
1114
+ "eval_steps_per_second": 16.255,
1115
+ "step": 1400
1116
+ },
1117
+ {
1118
+ "epoch": 3.248847926267281,
1119
+ "grad_norm": 0.24710910022258759,
1120
+ "learning_rate": 3.7557603686635945e-05,
1121
+ "loss": 0.0982,
1122
+ "step": 1410
1123
+ },
1124
+ {
1125
+ "epoch": 3.271889400921659,
1126
+ "grad_norm": 1.3541345596313477,
1127
+ "learning_rate": 3.640552995391705e-05,
1128
+ "loss": 0.1062,
1129
+ "step": 1420
1130
+ },
1131
+ {
1132
+ "epoch": 3.294930875576037,
1133
+ "grad_norm": 0.07805185765028,
1134
+ "learning_rate": 3.525345622119816e-05,
1135
+ "loss": 0.0367,
1136
+ "step": 1430
1137
+ },
1138
+ {
1139
+ "epoch": 3.3179723502304146,
1140
+ "grad_norm": 0.704824686050415,
1141
+ "learning_rate": 3.410138248847927e-05,
1142
+ "loss": 0.0576,
1143
+ "step": 1440
1144
+ },
1145
+ {
1146
+ "epoch": 3.3410138248847927,
1147
+ "grad_norm": 3.216744899749756,
1148
+ "learning_rate": 3.294930875576037e-05,
1149
+ "loss": 0.123,
1150
+ "step": 1450
1151
+ },
1152
+ {
1153
+ "epoch": 3.3640552995391704,
1154
+ "grad_norm": 3.2812583446502686,
1155
+ "learning_rate": 3.1797235023041475e-05,
1156
+ "loss": 0.0535,
1157
+ "step": 1460
1158
+ },
1159
+ {
1160
+ "epoch": 3.3870967741935485,
1161
+ "grad_norm": 0.09345371276140213,
1162
+ "learning_rate": 3.0645161290322585e-05,
1163
+ "loss": 0.0363,
1164
+ "step": 1470
1165
+ },
1166
+ {
1167
+ "epoch": 3.410138248847926,
1168
+ "grad_norm": 0.5610162019729614,
1169
+ "learning_rate": 2.9493087557603688e-05,
1170
+ "loss": 0.0903,
1171
+ "step": 1480
1172
+ },
1173
+ {
1174
+ "epoch": 3.4331797235023043,
1175
+ "grad_norm": 1.413180947303772,
1176
+ "learning_rate": 2.8341013824884795e-05,
1177
+ "loss": 0.0792,
1178
+ "step": 1490
1179
+ },
1180
+ {
1181
+ "epoch": 3.456221198156682,
1182
+ "grad_norm": 6.735473155975342,
1183
+ "learning_rate": 2.7188940092165898e-05,
1184
+ "loss": 0.03,
1185
+ "step": 1500
1186
+ },
1187
+ {
1188
+ "epoch": 3.456221198156682,
1189
+ "eval_accuracy": 0.8371161548731643,
1190
+ "eval_loss": 0.5680701732635498,
1191
+ "eval_runtime": 11.6369,
1192
+ "eval_samples_per_second": 128.728,
1193
+ "eval_steps_per_second": 16.155,
1194
+ "step": 1500
1195
+ },
1196
+ {
1197
+ "epoch": 3.47926267281106,
1198
+ "grad_norm": 1.4329415559768677,
1199
+ "learning_rate": 2.6036866359447005e-05,
1200
+ "loss": 0.0206,
1201
+ "step": 1510
1202
+ },
1203
+ {
1204
+ "epoch": 3.5023041474654377,
1205
+ "grad_norm": 0.0513407364487648,
1206
+ "learning_rate": 2.488479262672811e-05,
1207
+ "loss": 0.0637,
1208
+ "step": 1520
1209
+ },
1210
+ {
1211
+ "epoch": 3.525345622119816,
1212
+ "grad_norm": 0.09985367208719254,
1213
+ "learning_rate": 2.3732718894009218e-05,
1214
+ "loss": 0.0829,
1215
+ "step": 1530
1216
+ },
1217
+ {
1218
+ "epoch": 3.5483870967741935,
1219
+ "grad_norm": 0.0632900595664978,
1220
+ "learning_rate": 2.258064516129032e-05,
1221
+ "loss": 0.0329,
1222
+ "step": 1540
1223
+ },
1224
+ {
1225
+ "epoch": 3.571428571428571,
1226
+ "grad_norm": 0.23229588568210602,
1227
+ "learning_rate": 2.1428571428571428e-05,
1228
+ "loss": 0.0709,
1229
+ "step": 1550
1230
+ },
1231
+ {
1232
+ "epoch": 3.5944700460829493,
1233
+ "grad_norm": 0.15025608241558075,
1234
+ "learning_rate": 2.0276497695852538e-05,
1235
+ "loss": 0.1135,
1236
+ "step": 1560
1237
+ },
1238
+ {
1239
+ "epoch": 3.6175115207373274,
1240
+ "grad_norm": 5.933778285980225,
1241
+ "learning_rate": 1.912442396313364e-05,
1242
+ "loss": 0.1093,
1243
+ "step": 1570
1244
+ },
1245
+ {
1246
+ "epoch": 3.640552995391705,
1247
+ "grad_norm": 0.06949874013662338,
1248
+ "learning_rate": 1.7972350230414748e-05,
1249
+ "loss": 0.0498,
1250
+ "step": 1580
1251
+ },
1252
+ {
1253
+ "epoch": 3.6635944700460827,
1254
+ "grad_norm": 0.09838402271270752,
1255
+ "learning_rate": 1.682027649769585e-05,
1256
+ "loss": 0.0598,
1257
+ "step": 1590
1258
+ },
1259
+ {
1260
+ "epoch": 3.686635944700461,
1261
+ "grad_norm": 0.9366612434387207,
1262
+ "learning_rate": 1.5668202764976958e-05,
1263
+ "loss": 0.092,
1264
+ "step": 1600
1265
+ },
1266
+ {
1267
+ "epoch": 3.686635944700461,
1268
+ "eval_accuracy": 0.8451268357810414,
1269
+ "eval_loss": 0.5304816365242004,
1270
+ "eval_runtime": 11.6024,
1271
+ "eval_samples_per_second": 129.111,
1272
+ "eval_steps_per_second": 16.203,
1273
+ "step": 1600
1274
+ },
1275
+ {
1276
+ "epoch": 3.709677419354839,
1277
+ "grad_norm": 0.04733530059456825,
1278
+ "learning_rate": 1.4516129032258066e-05,
1279
+ "loss": 0.0276,
1280
+ "step": 1610
1281
+ },
1282
+ {
1283
+ "epoch": 3.7327188940092166,
1284
+ "grad_norm": 0.08603022992610931,
1285
+ "learning_rate": 1.3364055299539171e-05,
1286
+ "loss": 0.0347,
1287
+ "step": 1620
1288
+ },
1289
+ {
1290
+ "epoch": 3.7557603686635943,
1291
+ "grad_norm": 0.041543856263160706,
1292
+ "learning_rate": 1.2211981566820276e-05,
1293
+ "loss": 0.026,
1294
+ "step": 1630
1295
+ },
1296
+ {
1297
+ "epoch": 3.7788018433179724,
1298
+ "grad_norm": 0.24026305973529816,
1299
+ "learning_rate": 1.1059907834101383e-05,
1300
+ "loss": 0.0496,
1301
+ "step": 1640
1302
+ },
1303
+ {
1304
+ "epoch": 3.80184331797235,
1305
+ "grad_norm": 0.03894612938165665,
1306
+ "learning_rate": 9.90783410138249e-06,
1307
+ "loss": 0.0365,
1308
+ "step": 1650
1309
+ },
1310
+ {
1311
+ "epoch": 3.824884792626728,
1312
+ "grad_norm": 4.442405700683594,
1313
+ "learning_rate": 8.755760368663595e-06,
1314
+ "loss": 0.0402,
1315
+ "step": 1660
1316
+ },
1317
+ {
1318
+ "epoch": 3.847926267281106,
1319
+ "grad_norm": 0.032657474279403687,
1320
+ "learning_rate": 7.603686635944701e-06,
1321
+ "loss": 0.0596,
1322
+ "step": 1670
1323
+ },
1324
+ {
1325
+ "epoch": 3.870967741935484,
1326
+ "grad_norm": 2.9635491371154785,
1327
+ "learning_rate": 6.451612903225806e-06,
1328
+ "loss": 0.0835,
1329
+ "step": 1680
1330
+ },
1331
+ {
1332
+ "epoch": 3.8940092165898617,
1333
+ "grad_norm": 0.06601913273334503,
1334
+ "learning_rate": 5.299539170506913e-06,
1335
+ "loss": 0.0277,
1336
+ "step": 1690
1337
+ },
1338
+ {
1339
+ "epoch": 3.9170506912442398,
1340
+ "grad_norm": 0.22990980744361877,
1341
+ "learning_rate": 4.147465437788019e-06,
1342
+ "loss": 0.0416,
1343
+ "step": 1700
1344
+ },
1345
+ {
1346
+ "epoch": 3.9170506912442398,
1347
+ "eval_accuracy": 0.8471295060080107,
1348
+ "eval_loss": 0.5442608594894409,
1349
+ "eval_runtime": 11.486,
1350
+ "eval_samples_per_second": 130.42,
1351
+ "eval_steps_per_second": 16.368,
1352
+ "step": 1700
1353
+ },
1354
+ {
1355
+ "epoch": 3.9400921658986174,
1356
+ "grad_norm": 0.06300857663154602,
1357
+ "learning_rate": 2.9953917050691243e-06,
1358
+ "loss": 0.0331,
1359
+ "step": 1710
1360
+ },
1361
+ {
1362
+ "epoch": 3.9631336405529956,
1363
+ "grad_norm": 1.9846687316894531,
1364
+ "learning_rate": 1.8433179723502305e-06,
1365
+ "loss": 0.04,
1366
+ "step": 1720
1367
+ },
1368
+ {
1369
+ "epoch": 3.986175115207373,
1370
+ "grad_norm": 0.4808693826198578,
1371
+ "learning_rate": 6.912442396313364e-07,
1372
+ "loss": 0.0494,
1373
+ "step": 1730
1374
+ },
1375
+ {
1376
+ "epoch": 4.0,
1377
+ "step": 1736,
1378
+ "total_flos": 2.1525139607212524e+18,
1379
+ "train_loss": 0.4232822818690181,
1380
+ "train_runtime": 559.5799,
1381
+ "train_samples_per_second": 49.637,
1382
+ "train_steps_per_second": 3.102
1383
+ }
1384
+ ],
1385
+ "logging_steps": 10,
1386
+ "max_steps": 1736,
1387
+ "num_input_tokens_seen": 0,
1388
+ "num_train_epochs": 4,
1389
+ "save_steps": 100,
1390
+ "stateful_callbacks": {
1391
+ "TrainerControl": {
1392
+ "args": {
1393
+ "should_epoch_stop": false,
1394
+ "should_evaluate": false,
1395
+ "should_log": false,
1396
+ "should_save": true,
1397
+ "should_training_stop": false
1398
+ },
1399
+ "attributes": {}
1400
+ }
1401
+ },
1402
+ "total_flos": 2.1525139607212524e+18,
1403
+ "train_batch_size": 16,
1404
+ "trial_name": null,
1405
+ "trial_params": null
1406
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d66167632c09a446e5d3275f18d6a203e0198247882b9b893a75ae4a978e75ca
3
+ size 5112