Hugging Face's logo

iisking
/
wav2vec2-vocal-disable-classification

Audio Classification
Transformers
Safetensors
wav2vec2
Inference Endpoints
Model card Files Community
iisking commited on
Commit
14075c2
•
1 Parent(s): b822237

Upload folder using huggingface_hub

Browse files
Files changed (44) hide show
  1. checkpoint-1026/config.json +132 -0
  2. checkpoint-1026/model.safetensors +3 -0
  3. checkpoint-1026/optimizer.pt +3 -0
  4. checkpoint-1026/preprocessor_config.json +9 -0
  5. checkpoint-1026/rng_state.pth +3 -0
  6. checkpoint-1026/scheduler.pt +3 -0
  7. checkpoint-1026/trainer_state.json +0 -0
  8. checkpoint-1026/training_args.bin +3 -0
  9. checkpoint-1083/config.json +132 -0
  10. checkpoint-1083/model.safetensors +3 -0
  11. checkpoint-1083/optimizer.pt +3 -0
  12. checkpoint-1083/preprocessor_config.json +9 -0
  13. checkpoint-1083/rng_state.pth +3 -0
  14. checkpoint-1083/scheduler.pt +3 -0
  15. checkpoint-1083/trainer_state.json +0 -0
  16. checkpoint-1083/training_args.bin +3 -0
  17. checkpoint-1140/config.json +132 -0
  18. checkpoint-1140/model.safetensors +3 -0
  19. checkpoint-1140/optimizer.pt +3 -0
  20. checkpoint-1140/preprocessor_config.json +9 -0
  21. checkpoint-1140/rng_state.pth +3 -0
  22. checkpoint-1140/scheduler.pt +3 -0
  23. checkpoint-1140/trainer_state.json +0 -0
  24. checkpoint-1140/training_args.bin +3 -0
  25. checkpoint-513/config.json +132 -0
  26. checkpoint-513/model.safetensors +3 -0
  27. checkpoint-513/optimizer.pt +3 -0
  28. checkpoint-513/preprocessor_config.json +9 -0
  29. checkpoint-513/rng_state.pth +3 -0
  30. checkpoint-513/scheduler.pt +3 -0
  31. checkpoint-513/trainer_state.json +3714 -0
  32. checkpoint-513/training_args.bin +3 -0
  33. checkpoint-969/config.json +132 -0
  34. checkpoint-969/model.safetensors +3 -0
  35. checkpoint-969/optimizer.pt +3 -0
  36. checkpoint-969/preprocessor_config.json +9 -0
  37. checkpoint-969/rng_state.pth +3 -0
  38. checkpoint-969/scheduler.pt +3 -0
  39. checkpoint-969/trainer_state.json +0 -0
  40. checkpoint-969/training_args.bin +3 -0
  41. config.json +132 -0
  42. model.safetensors +3 -0
  43. preprocessor_config.json +9 -0
  44. training_args.bin +3 -0
checkpoint-1026/config.json ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-base",
3
+ "activation_dropout": 0.0,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForSequenceClassification"
11
+ ],
12
+ "attention_dropout": 0.1,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 256,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": false,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "sum",
46
+ "ctc_zero_infinity": false,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": false,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_norm": "group",
52
+ "feat_proj_dropout": 0.1,
53
+ "feat_quantizer_dropout": 0.0,
54
+ "final_dropout": 0.0,
55
+ "freeze_feat_extract_train": true,
56
+ "hidden_act": "gelu",
57
+ "hidden_dropout": 0.1,
58
+ "hidden_size": 768,
59
+ "id2label": {
60
+ "0": "Lesion",
61
+ "1": "Normal",
62
+ "2": "Paralysis",
63
+ "3": "Atrophy",
64
+ "4": "MTD"
65
+ },
66
+ "initializer_range": 0.02,
67
+ "intermediate_size": 3072,
68
+ "label2id": {
69
+ "LABEL_0": 0,
70
+ "LABEL_1": 1,
71
+ "LABEL_2": 2,
72
+ "LABEL_3": 3,
73
+ "LABEL_4": 4
74
+ },
75
+ "layer_norm_eps": 1e-05,
76
+ "layerdrop": 0.0,
77
+ "mask_channel_length": 10,
78
+ "mask_channel_min_space": 1,
79
+ "mask_channel_other": 0.0,
80
+ "mask_channel_prob": 0.0,
81
+ "mask_channel_selection": "static",
82
+ "mask_feature_length": 10,
83
+ "mask_feature_min_masks": 0,
84
+ "mask_feature_prob": 0.0,
85
+ "mask_time_length": 10,
86
+ "mask_time_min_masks": 2,
87
+ "mask_time_min_space": 1,
88
+ "mask_time_other": 0.0,
89
+ "mask_time_prob": 0.05,
90
+ "mask_time_selection": "static",
91
+ "model_type": "wav2vec2",
92
+ "no_mask_channel_overlap": false,
93
+ "no_mask_time_overlap": false,
94
+ "num_adapter_layers": 3,
95
+ "num_attention_heads": 12,
96
+ "num_codevector_groups": 2,
97
+ "num_codevectors_per_group": 320,
98
+ "num_conv_pos_embedding_groups": 16,
99
+ "num_conv_pos_embeddings": 128,
100
+ "num_feat_extract_layers": 7,
101
+ "num_hidden_layers": 12,
102
+ "num_negatives": 100,
103
+ "output_hidden_size": 768,
104
+ "pad_token_id": 0,
105
+ "proj_codevector_dim": 256,
106
+ "tdnn_dilation": [
107
+ 1,
108
+ 2,
109
+ 3,
110
+ 1,
111
+ 1
112
+ ],
113
+ "tdnn_dim": [
114
+ 512,
115
+ 512,
116
+ 512,
117
+ 512,
118
+ 1500
119
+ ],
120
+ "tdnn_kernel": [
121
+ 5,
122
+ 3,
123
+ 3,
124
+ 1,
125
+ 1
126
+ ],
127
+ "torch_dtype": "float32",
128
+ "transformers_version": "4.42.0.dev0",
129
+ "use_weighted_layer_sum": false,
130
+ "vocab_size": 32,
131
+ "xvector_output_dim": 512
132
+ }
checkpoint-1026/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a13ef8d3c6d27544ea4e25f284378552117d51c2a5e61e6f60e5fbdd0bf98520
3
+ size 378305452
checkpoint-1026/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3d284f2e15d41e4d5a5ee56a573eb26d41d8c62a0fbc2a12ea1ad71956029c99
3
+ size 756738746
checkpoint-1026/preprocessor_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "return_attention_mask": false,
8
+ "sampling_rate": 16000
9
+ }
checkpoint-1026/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef723e20ba32c495a504c596a9f26fc96be40c2198744ba26329d568b565be57
3
+ size 14244
checkpoint-1026/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a7dce331dfadec131f73dba48a235d867c3015fbaad1b5bb5236f8009dd468c
3
+ size 1064
checkpoint-1026/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1026/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e48f80f5f7846c1de45b12dd03848923118b7f8dcfa969a22a9b8fb06dc3c376
3
+ size 5112
checkpoint-1083/config.json ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-base",
3
+ "activation_dropout": 0.0,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForSequenceClassification"
11
+ ],
12
+ "attention_dropout": 0.1,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 256,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": false,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "sum",
46
+ "ctc_zero_infinity": false,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": false,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_norm": "group",
52
+ "feat_proj_dropout": 0.1,
53
+ "feat_quantizer_dropout": 0.0,
54
+ "final_dropout": 0.0,
55
+ "freeze_feat_extract_train": true,
56
+ "hidden_act": "gelu",
57
+ "hidden_dropout": 0.1,
58
+ "hidden_size": 768,
59
+ "id2label": {
60
+ "0": "Lesion",
61
+ "1": "Normal",
62
+ "2": "Paralysis",
63
+ "3": "Atrophy",
64
+ "4": "MTD"
65
+ },
66
+ "initializer_range": 0.02,
67
+ "intermediate_size": 3072,
68
+ "label2id": {
69
+ "LABEL_0": 0,
70
+ "LABEL_1": 1,
71
+ "LABEL_2": 2,
72
+ "LABEL_3": 3,
73
+ "LABEL_4": 4
74
+ },
75
+ "layer_norm_eps": 1e-05,
76
+ "layerdrop": 0.0,
77
+ "mask_channel_length": 10,
78
+ "mask_channel_min_space": 1,
79
+ "mask_channel_other": 0.0,
80
+ "mask_channel_prob": 0.0,
81
+ "mask_channel_selection": "static",
82
+ "mask_feature_length": 10,
83
+ "mask_feature_min_masks": 0,
84
+ "mask_feature_prob": 0.0,
85
+ "mask_time_length": 10,
86
+ "mask_time_min_masks": 2,
87
+ "mask_time_min_space": 1,
88
+ "mask_time_other": 0.0,
89
+ "mask_time_prob": 0.05,
90
+ "mask_time_selection": "static",
91
+ "model_type": "wav2vec2",
92
+ "no_mask_channel_overlap": false,
93
+ "no_mask_time_overlap": false,
94
+ "num_adapter_layers": 3,
95
+ "num_attention_heads": 12,
96
+ "num_codevector_groups": 2,
97
+ "num_codevectors_per_group": 320,
98
+ "num_conv_pos_embedding_groups": 16,
99
+ "num_conv_pos_embeddings": 128,
100
+ "num_feat_extract_layers": 7,
101
+ "num_hidden_layers": 12,
102
+ "num_negatives": 100,
103
+ "output_hidden_size": 768,
104
+ "pad_token_id": 0,
105
+ "proj_codevector_dim": 256,
106
+ "tdnn_dilation": [
107
+ 1,
108
+ 2,
109
+ 3,
110
+ 1,
111
+ 1
112
+ ],
113
+ "tdnn_dim": [
114
+ 512,
115
+ 512,
116
+ 512,
117
+ 512,
118
+ 1500
119
+ ],
120
+ "tdnn_kernel": [
121
+ 5,
122
+ 3,
123
+ 3,
124
+ 1,
125
+ 1
126
+ ],
127
+ "torch_dtype": "float32",
128
+ "transformers_version": "4.42.0.dev0",
129
+ "use_weighted_layer_sum": false,
130
+ "vocab_size": 32,
131
+ "xvector_output_dim": 512
132
+ }
checkpoint-1083/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:427a8eee35a2526ab41f269d6fbfbbfb097a6e6e61a84250ec8918b4b44622f8
3
+ size 378305452
checkpoint-1083/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d110dba2d387ff19148f5d61ea322713b94ee1e89ab4bd4fea7891ee91ceac4e
3
+ size 756738746
checkpoint-1083/preprocessor_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "return_attention_mask": false,
8
+ "sampling_rate": 16000
9
+ }
checkpoint-1083/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:470a7e3f81e5ba253d56b1507997d23ebb24f851d7e65798cabd02072ea1ceb3
3
+ size 14308
checkpoint-1083/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff8e391d77e58cfb8f5e0b17cad4a41669b1c8c955c0cade84a2a62e78d84f74
3
+ size 1064
checkpoint-1083/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1083/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e48f80f5f7846c1de45b12dd03848923118b7f8dcfa969a22a9b8fb06dc3c376
3
+ size 5112
checkpoint-1140/config.json ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-base",
3
+ "activation_dropout": 0.0,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForSequenceClassification"
11
+ ],
12
+ "attention_dropout": 0.1,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 256,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": false,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "sum",
46
+ "ctc_zero_infinity": false,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": false,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_norm": "group",
52
+ "feat_proj_dropout": 0.1,
53
+ "feat_quantizer_dropout": 0.0,
54
+ "final_dropout": 0.0,
55
+ "freeze_feat_extract_train": true,
56
+ "hidden_act": "gelu",
57
+ "hidden_dropout": 0.1,
58
+ "hidden_size": 768,
59
+ "id2label": {
60
+ "0": "Lesion",
61
+ "1": "Normal",
62
+ "2": "Paralysis",
63
+ "3": "Atrophy",
64
+ "4": "MTD"
65
+ },
66
+ "initializer_range": 0.02,
67
+ "intermediate_size": 3072,
68
+ "label2id": {
69
+ "LABEL_0": 0,
70
+ "LABEL_1": 1,
71
+ "LABEL_2": 2,
72
+ "LABEL_3": 3,
73
+ "LABEL_4": 4
74
+ },
75
+ "layer_norm_eps": 1e-05,
76
+ "layerdrop": 0.0,
77
+ "mask_channel_length": 10,
78
+ "mask_channel_min_space": 1,
79
+ "mask_channel_other": 0.0,
80
+ "mask_channel_prob": 0.0,
81
+ "mask_channel_selection": "static",
82
+ "mask_feature_length": 10,
83
+ "mask_feature_min_masks": 0,
84
+ "mask_feature_prob": 0.0,
85
+ "mask_time_length": 10,
86
+ "mask_time_min_masks": 2,
87
+ "mask_time_min_space": 1,
88
+ "mask_time_other": 0.0,
89
+ "mask_time_prob": 0.05,
90
+ "mask_time_selection": "static",
91
+ "model_type": "wav2vec2",
92
+ "no_mask_channel_overlap": false,
93
+ "no_mask_time_overlap": false,
94
+ "num_adapter_layers": 3,
95
+ "num_attention_heads": 12,
96
+ "num_codevector_groups": 2,
97
+ "num_codevectors_per_group": 320,
98
+ "num_conv_pos_embedding_groups": 16,
99
+ "num_conv_pos_embeddings": 128,
100
+ "num_feat_extract_layers": 7,
101
+ "num_hidden_layers": 12,
102
+ "num_negatives": 100,
103
+ "output_hidden_size": 768,
104
+ "pad_token_id": 0,
105
+ "proj_codevector_dim": 256,
106
+ "tdnn_dilation": [
107
+ 1,
108
+ 2,
109
+ 3,
110
+ 1,
111
+ 1
112
+ ],
113
+ "tdnn_dim": [
114
+ 512,
115
+ 512,
116
+ 512,
117
+ 512,
118
+ 1500
119
+ ],
120
+ "tdnn_kernel": [
121
+ 5,
122
+ 3,
123
+ 3,
124
+ 1,
125
+ 1
126
+ ],
127
+ "torch_dtype": "float32",
128
+ "transformers_version": "4.42.0.dev0",
129
+ "use_weighted_layer_sum": false,
130
+ "vocab_size": 32,
131
+ "xvector_output_dim": 512
132
+ }
checkpoint-1140/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e738ac66ad882349667ad6d03d6c1789d3b0207b7d97600cc1d1dc33f7355467
3
+ size 378305452
checkpoint-1140/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ed0cad442623e73054478ad1442ff0481162f01043cbebbd25419a7a8f2a8d4
3
+ size 756738746
checkpoint-1140/preprocessor_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "return_attention_mask": false,
8
+ "sampling_rate": 16000
9
+ }
checkpoint-1140/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2531f1129fc963268357262aa0a7dfe489311a979b30084699ee64bdce1e9e49
3
+ size 14308
checkpoint-1140/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:80f1bd4aed8454ffc56bb88312fb47e8f83fd5018a990bdf57eab640964cc26a
3
+ size 1064
checkpoint-1140/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1140/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e48f80f5f7846c1de45b12dd03848923118b7f8dcfa969a22a9b8fb06dc3c376
3
+ size 5112
checkpoint-513/config.json ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-base",
3
+ "activation_dropout": 0.0,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForSequenceClassification"
11
+ ],
12
+ "attention_dropout": 0.1,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 256,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": false,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "sum",
46
+ "ctc_zero_infinity": false,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": false,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_norm": "group",
52
+ "feat_proj_dropout": 0.1,
53
+ "feat_quantizer_dropout": 0.0,
54
+ "final_dropout": 0.0,
55
+ "freeze_feat_extract_train": true,
56
+ "hidden_act": "gelu",
57
+ "hidden_dropout": 0.1,
58
+ "hidden_size": 768,
59
+ "id2label": {
60
+ "0": "Lesion",
61
+ "1": "Normal",
62
+ "2": "Paralysis",
63
+ "3": "Atrophy",
64
+ "4": "MTD"
65
+ },
66
+ "initializer_range": 0.02,
67
+ "intermediate_size": 3072,
68
+ "label2id": {
69
+ "LABEL_0": 0,
70
+ "LABEL_1": 1,
71
+ "LABEL_2": 2,
72
+ "LABEL_3": 3,
73
+ "LABEL_4": 4
74
+ },
75
+ "layer_norm_eps": 1e-05,
76
+ "layerdrop": 0.0,
77
+ "mask_channel_length": 10,
78
+ "mask_channel_min_space": 1,
79
+ "mask_channel_other": 0.0,
80
+ "mask_channel_prob": 0.0,
81
+ "mask_channel_selection": "static",
82
+ "mask_feature_length": 10,
83
+ "mask_feature_min_masks": 0,
84
+ "mask_feature_prob": 0.0,
85
+ "mask_time_length": 10,
86
+ "mask_time_min_masks": 2,
87
+ "mask_time_min_space": 1,
88
+ "mask_time_other": 0.0,
89
+ "mask_time_prob": 0.05,
90
+ "mask_time_selection": "static",
91
+ "model_type": "wav2vec2",
92
+ "no_mask_channel_overlap": false,
93
+ "no_mask_time_overlap": false,
94
+ "num_adapter_layers": 3,
95
+ "num_attention_heads": 12,
96
+ "num_codevector_groups": 2,
97
+ "num_codevectors_per_group": 320,
98
+ "num_conv_pos_embedding_groups": 16,
99
+ "num_conv_pos_embeddings": 128,
100
+ "num_feat_extract_layers": 7,
101
+ "num_hidden_layers": 12,
102
+ "num_negatives": 100,
103
+ "output_hidden_size": 768,
104
+ "pad_token_id": 0,
105
+ "proj_codevector_dim": 256,
106
+ "tdnn_dilation": [
107
+ 1,
108
+ 2,
109
+ 3,
110
+ 1,
111
+ 1
112
+ ],
113
+ "tdnn_dim": [
114
+ 512,
115
+ 512,
116
+ 512,
117
+ 512,
118
+ 1500
119
+ ],
120
+ "tdnn_kernel": [
121
+ 5,
122
+ 3,
123
+ 3,
124
+ 1,
125
+ 1
126
+ ],
127
+ "torch_dtype": "float32",
128
+ "transformers_version": "4.42.0.dev0",
129
+ "use_weighted_layer_sum": false,
130
+ "vocab_size": 32,
131
+ "xvector_output_dim": 512
132
+ }
checkpoint-513/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a53b1994aba1a9bb0d194f45b0b18aeb8ac6277dede830adf4babb6c4235103
3
+ size 378305452
checkpoint-513/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3308fe2d390fd84ca90ad7f41e21e1324b45546b47b16162347e8c8af457eae0
3
+ size 756738746
checkpoint-513/preprocessor_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "return_attention_mask": false,
8
+ "sampling_rate": 16000
9
+ }
checkpoint-513/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c49d1e323157d81893d74419d5d0847f2ffe7d6e4befd0fa20f240fb36ddb56b
3
+ size 14308
checkpoint-513/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fe098435b37d51f68b03c2e3062fe1ea8e158a3fb57b8e5a2568b8aeb86f5683
3
+ size 1064
checkpoint-513/trainer_state.json ADDED
@@ -0,0 +1,3714 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8051581382751465,
3
+ "best_model_checkpoint": "wav2vec2-vocal-disable-classification/checkpoint-513",
4
+ "epoch": 9.0,
5
+ "eval_steps": 1,
6
+ "global_step": 513,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.017543859649122806,
13
+ "grad_norm": 1.56613290309906,
14
+ "learning_rate": 6.000000000000001e-07,
15
+ "loss": 1.6056,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.03508771929824561,
20
+ "grad_norm": 1.8094905614852905,
21
+ "learning_rate": 1.2000000000000002e-06,
22
+ "loss": 1.6101,
23
+ "step": 2
24
+ },
25
+ {
26
+ "epoch": 0.05263157894736842,
27
+ "grad_norm": 1.138840913772583,
28
+ "learning_rate": 1.8e-06,
29
+ "loss": 1.609,
30
+ "step": 3
31
+ },
32
+ {
33
+ "epoch": 0.07017543859649122,
34
+ "grad_norm": 3.695634603500366,
35
+ "learning_rate": 2.4000000000000003e-06,
36
+ "loss": 1.6097,
37
+ "step": 4
38
+ },
39
+ {
40
+ "epoch": 0.08771929824561403,
41
+ "grad_norm": 1.9893696308135986,
42
+ "learning_rate": 3e-06,
43
+ "loss": 1.6028,
44
+ "step": 5
45
+ },
46
+ {
47
+ "epoch": 0.10526315789473684,
48
+ "grad_norm": 2.095263957977295,
49
+ "learning_rate": 3.6e-06,
50
+ "loss": 1.6016,
51
+ "step": 6
52
+ },
53
+ {
54
+ "epoch": 0.12280701754385964,
55
+ "grad_norm": 1.7659621238708496,
56
+ "learning_rate": 4.2000000000000004e-06,
57
+ "loss": 1.5994,
58
+ "step": 7
59
+ },
60
+ {
61
+ "epoch": 0.14035087719298245,
62
+ "grad_norm": 2.041477680206299,
63
+ "learning_rate": 4.800000000000001e-06,
64
+ "loss": 1.5951,
65
+ "step": 8
66
+ },
67
+ {
68
+ "epoch": 0.15789473684210525,
69
+ "grad_norm": 1.6106765270233154,
70
+ "learning_rate": 5.4e-06,
71
+ "loss": 1.5957,
72
+ "step": 9
73
+ },
74
+ {
75
+ "epoch": 0.17543859649122806,
76
+ "grad_norm": 1.5744407176971436,
77
+ "learning_rate": 6e-06,
78
+ "loss": 1.5951,
79
+ "step": 10
80
+ },
81
+ {
82
+ "epoch": 0.19298245614035087,
83
+ "grad_norm": 1.0527958869934082,
84
+ "learning_rate": 6.6e-06,
85
+ "loss": 1.5994,
86
+ "step": 11
87
+ },
88
+ {
89
+ "epoch": 0.21052631578947367,
90
+ "grad_norm": 1.5314522981643677,
91
+ "learning_rate": 7.2e-06,
92
+ "loss": 1.5766,
93
+ "step": 12
94
+ },
95
+ {
96
+ "epoch": 0.22807017543859648,
97
+ "grad_norm": 1.0843933820724487,
98
+ "learning_rate": 7.8e-06,
99
+ "loss": 1.5976,
100
+ "step": 13
101
+ },
102
+ {
103
+ "epoch": 0.24561403508771928,
104
+ "grad_norm": 1.8420531749725342,
105
+ "learning_rate": 8.400000000000001e-06,
106
+ "loss": 1.5803,
107
+ "step": 14
108
+ },
109
+ {
110
+ "epoch": 0.2631578947368421,
111
+ "grad_norm": 2.5668458938598633,
112
+ "learning_rate": 9e-06,
113
+ "loss": 1.5839,
114
+ "step": 15
115
+ },
116
+ {
117
+ "epoch": 0.2807017543859649,
118
+ "grad_norm": 1.5706558227539062,
119
+ "learning_rate": 9.600000000000001e-06,
120
+ "loss": 1.5522,
121
+ "step": 16
122
+ },
123
+ {
124
+ "epoch": 0.2982456140350877,
125
+ "grad_norm": 1.722334384918213,
126
+ "learning_rate": 1.02e-05,
127
+ "loss": 1.5391,
128
+ "step": 17
129
+ },
130
+ {
131
+ "epoch": 0.3157894736842105,
132
+ "grad_norm": 1.317153811454773,
133
+ "learning_rate": 1.08e-05,
134
+ "loss": 1.6246,
135
+ "step": 18
136
+ },
137
+ {
138
+ "epoch": 0.3333333333333333,
139
+ "grad_norm": 1.3377695083618164,
140
+ "learning_rate": 1.1400000000000001e-05,
141
+ "loss": 1.6166,
142
+ "step": 19
143
+ },
144
+ {
145
+ "epoch": 0.3508771929824561,
146
+ "grad_norm": 3.216266632080078,
147
+ "learning_rate": 1.2e-05,
148
+ "loss": 1.5151,
149
+ "step": 20
150
+ },
151
+ {
152
+ "epoch": 0.3684210526315789,
153
+ "grad_norm": 1.3238260746002197,
154
+ "learning_rate": 1.26e-05,
155
+ "loss": 1.5195,
156
+ "step": 21
157
+ },
158
+ {
159
+ "epoch": 0.38596491228070173,
160
+ "grad_norm": 2.946779251098633,
161
+ "learning_rate": 1.32e-05,
162
+ "loss": 1.503,
163
+ "step": 22
164
+ },
165
+ {
166
+ "epoch": 0.40350877192982454,
167
+ "grad_norm": 1.427848219871521,
168
+ "learning_rate": 1.3800000000000002e-05,
169
+ "loss": 1.5538,
170
+ "step": 23
171
+ },
172
+ {
173
+ "epoch": 0.42105263157894735,
174
+ "grad_norm": 3.5313634872436523,
175
+ "learning_rate": 1.44e-05,
176
+ "loss": 1.545,
177
+ "step": 24
178
+ },
179
+ {
180
+ "epoch": 0.43859649122807015,
181
+ "grad_norm": 1.3384391069412231,
182
+ "learning_rate": 1.5e-05,
183
+ "loss": 1.5406,
184
+ "step": 25
185
+ },
186
+ {
187
+ "epoch": 0.45614035087719296,
188
+ "grad_norm": 1.635528802871704,
189
+ "learning_rate": 1.56e-05,
190
+ "loss": 1.4864,
191
+ "step": 26
192
+ },
193
+ {
194
+ "epoch": 0.47368421052631576,
195
+ "grad_norm": 1.3669475317001343,
196
+ "learning_rate": 1.62e-05,
197
+ "loss": 1.5609,
198
+ "step": 27
199
+ },
200
+ {
201
+ "epoch": 0.49122807017543857,
202
+ "grad_norm": 1.532422661781311,
203
+ "learning_rate": 1.6800000000000002e-05,
204
+ "loss": 1.4829,
205
+ "step": 28
206
+ },
207
+ {
208
+ "epoch": 0.5087719298245614,
209
+ "grad_norm": 14.834467887878418,
210
+ "learning_rate": 1.74e-05,
211
+ "loss": 1.4049,
212
+ "step": 29
213
+ },
214
+ {
215
+ "epoch": 0.5263157894736842,
216
+ "grad_norm": 1.5839215517044067,
217
+ "learning_rate": 1.8e-05,
218
+ "loss": 1.4429,
219
+ "step": 30
220
+ },
221
+ {
222
+ "epoch": 0.543859649122807,
223
+ "grad_norm": 8.556641578674316,
224
+ "learning_rate": 1.86e-05,
225
+ "loss": 1.5076,
226
+ "step": 31
227
+ },
228
+ {
229
+ "epoch": 0.5614035087719298,
230
+ "grad_norm": 2.6420958042144775,
231
+ "learning_rate": 1.9200000000000003e-05,
232
+ "loss": 1.6325,
233
+ "step": 32
234
+ },
235
+ {
236
+ "epoch": 0.5789473684210527,
237
+ "grad_norm": 1.5453814268112183,
238
+ "learning_rate": 1.98e-05,
239
+ "loss": 1.5331,
240
+ "step": 33
241
+ },
242
+ {
243
+ "epoch": 0.5964912280701754,
244
+ "grad_norm": 2.1342880725860596,
245
+ "learning_rate": 2.04e-05,
246
+ "loss": 1.5563,
247
+ "step": 34
248
+ },
249
+ {
250
+ "epoch": 0.6140350877192983,
251
+ "grad_norm": 7.298281669616699,
252
+ "learning_rate": 2.1e-05,
253
+ "loss": 1.3024,
254
+ "step": 35
255
+ },
256
+ {
257
+ "epoch": 0.631578947368421,
258
+ "grad_norm": 2.423535108566284,
259
+ "learning_rate": 2.16e-05,
260
+ "loss": 1.5851,
261
+ "step": 36
262
+ },
263
+ {
264
+ "epoch": 0.6491228070175439,
265
+ "grad_norm": 2.8099899291992188,
266
+ "learning_rate": 2.22e-05,
267
+ "loss": 1.664,
268
+ "step": 37
269
+ },
270
+ {
271
+ "epoch": 0.6666666666666666,
272
+ "grad_norm": 3.804816961288452,
273
+ "learning_rate": 2.2800000000000002e-05,
274
+ "loss": 1.4642,
275
+ "step": 38
276
+ },
277
+ {
278
+ "epoch": 0.6842105263157895,
279
+ "grad_norm": 5.346778392791748,
280
+ "learning_rate": 2.3400000000000003e-05,
281
+ "loss": 1.5016,
282
+ "step": 39
283
+ },
284
+ {
285
+ "epoch": 0.7017543859649122,
286
+ "grad_norm": 5.4332733154296875,
287
+ "learning_rate": 2.4e-05,
288
+ "loss": 1.4937,
289
+ "step": 40
290
+ },
291
+ {
292
+ "epoch": 0.7192982456140351,
293
+ "grad_norm": 3.2267627716064453,
294
+ "learning_rate": 2.4599999999999998e-05,
295
+ "loss": 1.4857,
296
+ "step": 41
297
+ },
298
+ {
299
+ "epoch": 0.7368421052631579,
300
+ "grad_norm": 9.248668670654297,
301
+ "learning_rate": 2.52e-05,
302
+ "loss": 1.4574,
303
+ "step": 42
304
+ },
305
+ {
306
+ "epoch": 0.7543859649122807,
307
+ "grad_norm": 2.983009099960327,
308
+ "learning_rate": 2.58e-05,
309
+ "loss": 1.4802,
310
+ "step": 43
311
+ },
312
+ {
313
+ "epoch": 0.7719298245614035,
314
+ "grad_norm": 9.322035789489746,
315
+ "learning_rate": 2.64e-05,
316
+ "loss": 1.3646,
317
+ "step": 44
318
+ },
319
+ {
320
+ "epoch": 0.7894736842105263,
321
+ "grad_norm": 3.1577627658843994,
322
+ "learning_rate": 2.7000000000000002e-05,
323
+ "loss": 1.3842,
324
+ "step": 45
325
+ },
326
+ {
327
+ "epoch": 0.8070175438596491,
328
+ "grad_norm": 3.0866541862487793,
329
+ "learning_rate": 2.7600000000000003e-05,
330
+ "loss": 1.478,
331
+ "step": 46
332
+ },
333
+ {
334
+ "epoch": 0.8245614035087719,
335
+ "grad_norm": 2.434828281402588,
336
+ "learning_rate": 2.8199999999999998e-05,
337
+ "loss": 1.2678,
338
+ "step": 47
339
+ },
340
+ {
341
+ "epoch": 0.8421052631578947,
342
+ "grad_norm": 19.617595672607422,
343
+ "learning_rate": 2.88e-05,
344
+ "loss": 1.5149,
345
+ "step": 48
346
+ },
347
+ {
348
+ "epoch": 0.8596491228070176,
349
+ "grad_norm": 3.927741527557373,
350
+ "learning_rate": 2.94e-05,
351
+ "loss": 1.3689,
352
+ "step": 49
353
+ },
354
+ {
355
+ "epoch": 0.8771929824561403,
356
+ "grad_norm": 4.532490253448486,
357
+ "learning_rate": 3e-05,
358
+ "loss": 1.5975,
359
+ "step": 50
360
+ },
361
+ {
362
+ "epoch": 0.8947368421052632,
363
+ "grad_norm": 4.9621500968933105,
364
+ "learning_rate": 2.9972477064220187e-05,
365
+ "loss": 1.4361,
366
+ "step": 51
367
+ },
368
+ {
369
+ "epoch": 0.9122807017543859,
370
+ "grad_norm": 6.061293125152588,
371
+ "learning_rate": 2.994495412844037e-05,
372
+ "loss": 1.4083,
373
+ "step": 52
374
+ },
375
+ {
376
+ "epoch": 0.9298245614035088,
377
+ "grad_norm": 4.741385459899902,
378
+ "learning_rate": 2.991743119266055e-05,
379
+ "loss": 1.4497,
380
+ "step": 53
381
+ },
382
+ {
383
+ "epoch": 0.9473684210526315,
384
+ "grad_norm": 3.2169480323791504,
385
+ "learning_rate": 2.9889908256880734e-05,
386
+ "loss": 1.4465,
387
+ "step": 54
388
+ },
389
+ {
390
+ "epoch": 0.9649122807017544,
391
+ "grad_norm": 7.039127349853516,
392
+ "learning_rate": 2.986238532110092e-05,
393
+ "loss": 1.2431,
394
+ "step": 55
395
+ },
396
+ {
397
+ "epoch": 0.9824561403508771,
398
+ "grad_norm": 6.163113594055176,
399
+ "learning_rate": 2.9834862385321102e-05,
400
+ "loss": 1.2167,
401
+ "step": 56
402
+ },
403
+ {
404
+ "epoch": 1.0,
405
+ "grad_norm": 3.4295806884765625,
406
+ "learning_rate": 2.9807339449541285e-05,
407
+ "loss": 1.3198,
408
+ "step": 57
409
+ },
410
+ {
411
+ "epoch": 1.0,
412
+ "eval_accuracy": 0.4482758620689655,
413
+ "eval_loss": 1.3703213930130005,
414
+ "eval_roc_auc": 0.7240004822764974,
415
+ "eval_runtime": 9.1707,
416
+ "eval_samples_per_second": 28.46,
417
+ "eval_steps_per_second": 1.854,
418
+ "step": 57
419
+ },
420
+ {
421
+ "epoch": 1.0175438596491229,
422
+ "grad_norm": 2.1605288982391357,
423
+ "learning_rate": 2.9779816513761467e-05,
424
+ "loss": 1.3161,
425
+ "step": 58
426
+ },
427
+ {
428
+ "epoch": 1.0350877192982457,
429
+ "grad_norm": 2.469620943069458,
430
+ "learning_rate": 2.9752293577981653e-05,
431
+ "loss": 1.4468,
432
+ "step": 59
433
+ },
434
+ {
435
+ "epoch": 1.0526315789473684,
436
+ "grad_norm": 5.3216376304626465,
437
+ "learning_rate": 2.9724770642201836e-05,
438
+ "loss": 1.3517,
439
+ "step": 60
440
+ },
441
+ {
442
+ "epoch": 1.0701754385964912,
443
+ "grad_norm": 5.191178798675537,
444
+ "learning_rate": 2.9697247706422018e-05,
445
+ "loss": 1.3841,
446
+ "step": 61
447
+ },
448
+ {
449
+ "epoch": 1.087719298245614,
450
+ "grad_norm": 2.3336193561553955,
451
+ "learning_rate": 2.9669724770642204e-05,
452
+ "loss": 1.3057,
453
+ "step": 62
454
+ },
455
+ {
456
+ "epoch": 1.1052631578947367,
457
+ "grad_norm": 42.809783935546875,
458
+ "learning_rate": 2.9642201834862386e-05,
459
+ "loss": 1.2346,
460
+ "step": 63
461
+ },
462
+ {
463
+ "epoch": 1.1228070175438596,
464
+ "grad_norm": 6.859128952026367,
465
+ "learning_rate": 2.9614678899082572e-05,
466
+ "loss": 1.4394,
467
+ "step": 64
468
+ },
469
+ {
470
+ "epoch": 1.1403508771929824,
471
+ "grad_norm": 9.006386756896973,
472
+ "learning_rate": 2.958715596330275e-05,
473
+ "loss": 1.2954,
474
+ "step": 65
475
+ },
476
+ {
477
+ "epoch": 1.1578947368421053,
478
+ "grad_norm": 2.543628454208374,
479
+ "learning_rate": 2.9559633027522937e-05,
480
+ "loss": 1.0557,
481
+ "step": 66
482
+ },
483
+ {
484
+ "epoch": 1.1754385964912282,
485
+ "grad_norm": 2.736093044281006,
486
+ "learning_rate": 2.953211009174312e-05,
487
+ "loss": 1.2225,
488
+ "step": 67
489
+ },
490
+ {
491
+ "epoch": 1.1929824561403508,
492
+ "grad_norm": 4.434782028198242,
493
+ "learning_rate": 2.9504587155963306e-05,
494
+ "loss": 1.1875,
495
+ "step": 68
496
+ },
497
+ {
498
+ "epoch": 1.2105263157894737,
499
+ "grad_norm": 8.687559127807617,
500
+ "learning_rate": 2.9477064220183485e-05,
501
+ "loss": 0.937,
502
+ "step": 69
503
+ },
504
+ {
505
+ "epoch": 1.2280701754385965,
506
+ "grad_norm": 3.4958977699279785,
507
+ "learning_rate": 2.944954128440367e-05,
508
+ "loss": 1.2851,
509
+ "step": 70
510
+ },
511
+ {
512
+ "epoch": 1.2456140350877192,
513
+ "grad_norm": 12.490559577941895,
514
+ "learning_rate": 2.9422018348623853e-05,
515
+ "loss": 1.5578,
516
+ "step": 71
517
+ },
518
+ {
519
+ "epoch": 1.263157894736842,
520
+ "grad_norm": 3.466176748275757,
521
+ "learning_rate": 2.939449541284404e-05,
522
+ "loss": 0.9698,
523
+ "step": 72
524
+ },
525
+ {
526
+ "epoch": 1.280701754385965,
527
+ "grad_norm": 9.730321884155273,
528
+ "learning_rate": 2.936697247706422e-05,
529
+ "loss": 1.2044,
530
+ "step": 73
531
+ },
532
+ {
533
+ "epoch": 1.2982456140350878,
534
+ "grad_norm": 12.908202171325684,
535
+ "learning_rate": 2.9339449541284404e-05,
536
+ "loss": 1.3778,
537
+ "step": 74
538
+ },
539
+ {
540
+ "epoch": 1.3157894736842106,
541
+ "grad_norm": 7.409069538116455,
542
+ "learning_rate": 2.931192660550459e-05,
543
+ "loss": 1.6065,
544
+ "step": 75
545
+ },
546
+ {
547
+ "epoch": 1.3333333333333333,
548
+ "grad_norm": 12.185145378112793,
549
+ "learning_rate": 2.9284403669724772e-05,
550
+ "loss": 1.2463,
551
+ "step": 76
552
+ },
553
+ {
554
+ "epoch": 1.3508771929824561,
555
+ "grad_norm": 16.505674362182617,
556
+ "learning_rate": 2.9256880733944955e-05,
557
+ "loss": 1.0027,
558
+ "step": 77
559
+ },
560
+ {
561
+ "epoch": 1.368421052631579,
562
+ "grad_norm": 9.078436851501465,
563
+ "learning_rate": 2.9229357798165137e-05,
564
+ "loss": 1.2037,
565
+ "step": 78
566
+ },
567
+ {
568
+ "epoch": 1.3859649122807016,
569
+ "grad_norm": 23.13191032409668,
570
+ "learning_rate": 2.9201834862385323e-05,
571
+ "loss": 1.0528,
572
+ "step": 79
573
+ },
574
+ {
575
+ "epoch": 1.4035087719298245,
576
+ "grad_norm": 7.651644229888916,
577
+ "learning_rate": 2.9174311926605506e-05,
578
+ "loss": 1.0373,
579
+ "step": 80
580
+ },
581
+ {
582
+ "epoch": 1.4210526315789473,
583
+ "grad_norm": 15.649460792541504,
584
+ "learning_rate": 2.9146788990825688e-05,
585
+ "loss": 1.3094,
586
+ "step": 81
587
+ },
588
+ {
589
+ "epoch": 1.4385964912280702,
590
+ "grad_norm": 14.335933685302734,
591
+ "learning_rate": 2.9119266055045874e-05,
592
+ "loss": 1.229,
593
+ "step": 82
594
+ },
595
+ {
596
+ "epoch": 1.456140350877193,
597
+ "grad_norm": 13.704193115234375,
598
+ "learning_rate": 2.9091743119266056e-05,
599
+ "loss": 1.2517,
600
+ "step": 83
601
+ },
602
+ {
603
+ "epoch": 1.4736842105263157,
604
+ "grad_norm": 6.872323036193848,
605
+ "learning_rate": 2.9064220183486242e-05,
606
+ "loss": 1.0614,
607
+ "step": 84
608
+ },
609
+ {
610
+ "epoch": 1.4912280701754386,
611
+ "grad_norm": 58.63323211669922,
612
+ "learning_rate": 2.903669724770642e-05,
613
+ "loss": 1.2651,
614
+ "step": 85
615
+ },
616
+ {
617
+ "epoch": 1.5087719298245614,
618
+ "grad_norm": 16.493391036987305,
619
+ "learning_rate": 2.9009174311926607e-05,
620
+ "loss": 1.0977,
621
+ "step": 86
622
+ },
623
+ {
624
+ "epoch": 1.526315789473684,
625
+ "grad_norm": 76.30412292480469,
626
+ "learning_rate": 2.898165137614679e-05,
627
+ "loss": 1.3534,
628
+ "step": 87
629
+ },
630
+ {
631
+ "epoch": 1.543859649122807,
632
+ "grad_norm": 35.07970428466797,
633
+ "learning_rate": 2.8954128440366975e-05,
634
+ "loss": 1.1015,
635
+ "step": 88
636
+ },
637
+ {
638
+ "epoch": 1.5614035087719298,
639
+ "grad_norm": 41.934513092041016,
640
+ "learning_rate": 2.8926605504587155e-05,
641
+ "loss": 1.252,
642
+ "step": 89
643
+ },
644
+ {
645
+ "epoch": 1.5789473684210527,
646
+ "grad_norm": 6.429330825805664,
647
+ "learning_rate": 2.889908256880734e-05,
648
+ "loss": 0.9942,
649
+ "step": 90
650
+ },
651
+ {
652
+ "epoch": 1.5964912280701755,
653
+ "grad_norm": 4.915737152099609,
654
+ "learning_rate": 2.8871559633027523e-05,
655
+ "loss": 1.0583,
656
+ "step": 91
657
+ },
658
+ {
659
+ "epoch": 1.6140350877192984,
660
+ "grad_norm": 5.171088695526123,
661
+ "learning_rate": 2.884403669724771e-05,
662
+ "loss": 0.9167,
663
+ "step": 92
664
+ },
665
+ {
666
+ "epoch": 1.631578947368421,
667
+ "grad_norm": 4.840869426727295,
668
+ "learning_rate": 2.881651376146789e-05,
669
+ "loss": 1.3165,
670
+ "step": 93
671
+ },
672
+ {
673
+ "epoch": 1.6491228070175439,
674
+ "grad_norm": 8.278257369995117,
675
+ "learning_rate": 2.8788990825688074e-05,
676
+ "loss": 1.0294,
677
+ "step": 94
678
+ },
679
+ {
680
+ "epoch": 1.6666666666666665,
681
+ "grad_norm": 7.380605697631836,
682
+ "learning_rate": 2.876146788990826e-05,
683
+ "loss": 1.16,
684
+ "step": 95
685
+ },
686
+ {
687
+ "epoch": 1.6842105263157894,
688
+ "grad_norm": 27.351318359375,
689
+ "learning_rate": 2.8733944954128442e-05,
690
+ "loss": 1.6427,
691
+ "step": 96
692
+ },
693
+ {
694
+ "epoch": 1.7017543859649122,
695
+ "grad_norm": 5.766673564910889,
696
+ "learning_rate": 2.8706422018348625e-05,
697
+ "loss": 1.1461,
698
+ "step": 97
699
+ },
700
+ {
701
+ "epoch": 1.719298245614035,
702
+ "grad_norm": 6.586304664611816,
703
+ "learning_rate": 2.8678899082568807e-05,
704
+ "loss": 1.1368,
705
+ "step": 98
706
+ },
707
+ {
708
+ "epoch": 1.736842105263158,
709
+ "grad_norm": 13.888910293579102,
710
+ "learning_rate": 2.8651376146788993e-05,
711
+ "loss": 0.8426,
712
+ "step": 99
713
+ },
714
+ {
715
+ "epoch": 1.7543859649122808,
716
+ "grad_norm": 14.992203712463379,
717
+ "learning_rate": 2.8623853211009175e-05,
718
+ "loss": 1.1494,
719
+ "step": 100
720
+ },
721
+ {
722
+ "epoch": 1.7719298245614035,
723
+ "grad_norm": 8.557571411132812,
724
+ "learning_rate": 2.8596330275229358e-05,
725
+ "loss": 1.1022,
726
+ "step": 101
727
+ },
728
+ {
729
+ "epoch": 1.7894736842105263,
730
+ "grad_norm": 19.743364334106445,
731
+ "learning_rate": 2.856880733944954e-05,
732
+ "loss": 1.0737,
733
+ "step": 102
734
+ },
735
+ {
736
+ "epoch": 1.807017543859649,
737
+ "grad_norm": 13.77326488494873,
738
+ "learning_rate": 2.8541284403669726e-05,
739
+ "loss": 0.8547,
740
+ "step": 103
741
+ },
742
+ {
743
+ "epoch": 1.8245614035087718,
744
+ "grad_norm": 4.906980514526367,
745
+ "learning_rate": 2.8513761467889912e-05,
746
+ "loss": 0.8529,
747
+ "step": 104
748
+ },
749
+ {
750
+ "epoch": 1.8421052631578947,
751
+ "grad_norm": 2.8242599964141846,
752
+ "learning_rate": 2.848623853211009e-05,
753
+ "loss": 1.2058,
754
+ "step": 105
755
+ },
756
+ {
757
+ "epoch": 1.8596491228070176,
758
+ "grad_norm": 3.802095890045166,
759
+ "learning_rate": 2.8458715596330277e-05,
760
+ "loss": 0.9236,
761
+ "step": 106
762
+ },
763
+ {
764
+ "epoch": 1.8771929824561404,
765
+ "grad_norm": 4.197514057159424,
766
+ "learning_rate": 2.843119266055046e-05,
767
+ "loss": 0.8856,
768
+ "step": 107
769
+ },
770
+ {
771
+ "epoch": 1.8947368421052633,
772
+ "grad_norm": 13.207415580749512,
773
+ "learning_rate": 2.8403669724770645e-05,
774
+ "loss": 0.769,
775
+ "step": 108
776
+ },
777
+ {
778
+ "epoch": 1.912280701754386,
779
+ "grad_norm": 4.843390941619873,
780
+ "learning_rate": 2.8376146788990824e-05,
781
+ "loss": 1.1706,
782
+ "step": 109
783
+ },
784
+ {
785
+ "epoch": 1.9298245614035088,
786
+ "grad_norm": 13.065381050109863,
787
+ "learning_rate": 2.834862385321101e-05,
788
+ "loss": 1.0655,
789
+ "step": 110
790
+ },
791
+ {
792
+ "epoch": 1.9473684210526314,
793
+ "grad_norm": 15.791166305541992,
794
+ "learning_rate": 2.8321100917431193e-05,
795
+ "loss": 1.2378,
796
+ "step": 111
797
+ },
798
+ {
799
+ "epoch": 1.9649122807017543,
800
+ "grad_norm": 11.551010131835938,
801
+ "learning_rate": 2.829357798165138e-05,
802
+ "loss": 0.9768,
803
+ "step": 112
804
+ },
805
+ {
806
+ "epoch": 1.9824561403508771,
807
+ "grad_norm": 14.838546752929688,
808
+ "learning_rate": 2.826605504587156e-05,
809
+ "loss": 1.0073,
810
+ "step": 113
811
+ },
812
+ {
813
+ "epoch": 2.0,
814
+ "grad_norm": 2.385023355484009,
815
+ "learning_rate": 2.8238532110091744e-05,
816
+ "loss": 0.9486,
817
+ "step": 114
818
+ },
819
+ {
820
+ "epoch": 2.0,
821
+ "eval_accuracy": 0.6091954022988506,
822
+ "eval_loss": 1.0561790466308594,
823
+ "eval_roc_auc": 0.8297140297597192,
824
+ "eval_runtime": 10.2393,
825
+ "eval_samples_per_second": 25.49,
826
+ "eval_steps_per_second": 1.66,
827
+ "step": 114
828
+ },
829
+ {
830
+ "epoch": 2.017543859649123,
831
+ "grad_norm": 13.100618362426758,
832
+ "learning_rate": 2.821100917431193e-05,
833
+ "loss": 0.9394,
834
+ "step": 115
835
+ },
836
+ {
837
+ "epoch": 2.0350877192982457,
838
+ "grad_norm": 22.331451416015625,
839
+ "learning_rate": 2.8183486238532112e-05,
840
+ "loss": 1.0603,
841
+ "step": 116
842
+ },
843
+ {
844
+ "epoch": 2.0526315789473686,
845
+ "grad_norm": 4.956656455993652,
846
+ "learning_rate": 2.8155963302752294e-05,
847
+ "loss": 0.9566,
848
+ "step": 117
849
+ },
850
+ {
851
+ "epoch": 2.0701754385964914,
852
+ "grad_norm": 19.40747833251953,
853
+ "learning_rate": 2.8128440366972477e-05,
854
+ "loss": 1.1504,
855
+ "step": 118
856
+ },
857
+ {
858
+ "epoch": 2.087719298245614,
859
+ "grad_norm": 13.512128829956055,
860
+ "learning_rate": 2.8100917431192663e-05,
861
+ "loss": 1.194,
862
+ "step": 119
863
+ },
864
+ {
865
+ "epoch": 2.1052631578947367,
866
+ "grad_norm": 6.616446495056152,
867
+ "learning_rate": 2.8073394495412845e-05,
868
+ "loss": 1.0321,
869
+ "step": 120
870
+ },
871
+ {
872
+ "epoch": 2.1228070175438596,
873
+ "grad_norm": 2.3998453617095947,
874
+ "learning_rate": 2.8045871559633028e-05,
875
+ "loss": 1.0636,
876
+ "step": 121
877
+ },
878
+ {
879
+ "epoch": 2.1403508771929824,
880
+ "grad_norm": 3.214811086654663,
881
+ "learning_rate": 2.801834862385321e-05,
882
+ "loss": 0.7859,
883
+ "step": 122
884
+ },
885
+ {
886
+ "epoch": 2.1578947368421053,
887
+ "grad_norm": 2.0244710445404053,
888
+ "learning_rate": 2.7990825688073396e-05,
889
+ "loss": 0.8801,
890
+ "step": 123
891
+ },
892
+ {
893
+ "epoch": 2.175438596491228,
894
+ "grad_norm": 3.1682400703430176,
895
+ "learning_rate": 2.7963302752293582e-05,
896
+ "loss": 0.9413,
897
+ "step": 124
898
+ },
899
+ {
900
+ "epoch": 2.192982456140351,
901
+ "grad_norm": 3.509514570236206,
902
+ "learning_rate": 2.793577981651376e-05,
903
+ "loss": 0.8083,
904
+ "step": 125
905
+ },
906
+ {
907
+ "epoch": 2.2105263157894735,
908
+ "grad_norm": 6.175449848175049,
909
+ "learning_rate": 2.7908256880733947e-05,
910
+ "loss": 1.0884,
911
+ "step": 126
912
+ },
913
+ {
914
+ "epoch": 2.2280701754385963,
915
+ "grad_norm": 4.469012260437012,
916
+ "learning_rate": 2.788073394495413e-05,
917
+ "loss": 0.8281,
918
+ "step": 127
919
+ },
920
+ {
921
+ "epoch": 2.245614035087719,
922
+ "grad_norm": 18.924325942993164,
923
+ "learning_rate": 2.7853211009174315e-05,
924
+ "loss": 0.9676,
925
+ "step": 128
926
+ },
927
+ {
928
+ "epoch": 2.263157894736842,
929
+ "grad_norm": 24.372833251953125,
930
+ "learning_rate": 2.7825688073394494e-05,
931
+ "loss": 1.1955,
932
+ "step": 129
933
+ },
934
+ {
935
+ "epoch": 2.280701754385965,
936
+ "grad_norm": 1.3634321689605713,
937
+ "learning_rate": 2.779816513761468e-05,
938
+ "loss": 0.9889,
939
+ "step": 130
940
+ },
941
+ {
942
+ "epoch": 2.2982456140350878,
943
+ "grad_norm": 17.57159996032715,
944
+ "learning_rate": 2.7770642201834863e-05,
945
+ "loss": 1.1324,
946
+ "step": 131
947
+ },
948
+ {
949
+ "epoch": 2.3157894736842106,
950
+ "grad_norm": 16.63871192932129,
951
+ "learning_rate": 2.774311926605505e-05,
952
+ "loss": 1.1951,
953
+ "step": 132
954
+ },
955
+ {
956
+ "epoch": 2.3333333333333335,
957
+ "grad_norm": 16.466339111328125,
958
+ "learning_rate": 2.7715596330275228e-05,
959
+ "loss": 0.9375,
960
+ "step": 133
961
+ },
962
+ {
963
+ "epoch": 2.3508771929824563,
964
+ "grad_norm": 13.808109283447266,
965
+ "learning_rate": 2.7688073394495413e-05,
966
+ "loss": 1.0418,
967
+ "step": 134
968
+ },
969
+ {
970
+ "epoch": 2.3684210526315788,
971
+ "grad_norm": 5.8455729484558105,
972
+ "learning_rate": 2.76605504587156e-05,
973
+ "loss": 1.0906,
974
+ "step": 135
975
+ },
976
+ {
977
+ "epoch": 2.3859649122807016,
978
+ "grad_norm": 18.494245529174805,
979
+ "learning_rate": 2.7633027522935782e-05,
980
+ "loss": 0.9786,
981
+ "step": 136
982
+ },
983
+ {
984
+ "epoch": 2.4035087719298245,
985
+ "grad_norm": 21.47800636291504,
986
+ "learning_rate": 2.7605504587155964e-05,
987
+ "loss": 0.8138,
988
+ "step": 137
989
+ },
990
+ {
991
+ "epoch": 2.4210526315789473,
992
+ "grad_norm": 8.400023460388184,
993
+ "learning_rate": 2.7577981651376147e-05,
994
+ "loss": 1.027,
995
+ "step": 138
996
+ },
997
+ {
998
+ "epoch": 2.43859649122807,
999
+ "grad_norm": 7.198583602905273,
1000
+ "learning_rate": 2.7550458715596333e-05,
1001
+ "loss": 0.7104,
1002
+ "step": 139
1003
+ },
1004
+ {
1005
+ "epoch": 2.456140350877193,
1006
+ "grad_norm": 23.315410614013672,
1007
+ "learning_rate": 2.7522935779816515e-05,
1008
+ "loss": 1.2115,
1009
+ "step": 140
1010
+ },
1011
+ {
1012
+ "epoch": 2.473684210526316,
1013
+ "grad_norm": 18.734041213989258,
1014
+ "learning_rate": 2.7495412844036697e-05,
1015
+ "loss": 0.9739,
1016
+ "step": 141
1017
+ },
1018
+ {
1019
+ "epoch": 2.4912280701754383,
1020
+ "grad_norm": 10.015871047973633,
1021
+ "learning_rate": 2.746788990825688e-05,
1022
+ "loss": 0.9796,
1023
+ "step": 142
1024
+ },
1025
+ {
1026
+ "epoch": 2.5087719298245617,
1027
+ "grad_norm": 11.979524612426758,
1028
+ "learning_rate": 2.7440366972477066e-05,
1029
+ "loss": 0.8025,
1030
+ "step": 143
1031
+ },
1032
+ {
1033
+ "epoch": 2.526315789473684,
1034
+ "grad_norm": 9.438429832458496,
1035
+ "learning_rate": 2.741284403669725e-05,
1036
+ "loss": 0.8838,
1037
+ "step": 144
1038
+ },
1039
+ {
1040
+ "epoch": 2.543859649122807,
1041
+ "grad_norm": 43.801063537597656,
1042
+ "learning_rate": 2.738532110091743e-05,
1043
+ "loss": 0.9144,
1044
+ "step": 145
1045
+ },
1046
+ {
1047
+ "epoch": 2.56140350877193,
1048
+ "grad_norm": 19.495887756347656,
1049
+ "learning_rate": 2.7357798165137617e-05,
1050
+ "loss": 0.8573,
1051
+ "step": 146
1052
+ },
1053
+ {
1054
+ "epoch": 2.5789473684210527,
1055
+ "grad_norm": 21.922138214111328,
1056
+ "learning_rate": 2.73302752293578e-05,
1057
+ "loss": 0.9611,
1058
+ "step": 147
1059
+ },
1060
+ {
1061
+ "epoch": 2.5964912280701755,
1062
+ "grad_norm": 16.30131721496582,
1063
+ "learning_rate": 2.7302752293577985e-05,
1064
+ "loss": 1.3519,
1065
+ "step": 148
1066
+ },
1067
+ {
1068
+ "epoch": 2.6140350877192984,
1069
+ "grad_norm": 43.10526657104492,
1070
+ "learning_rate": 2.7275229357798164e-05,
1071
+ "loss": 1.0194,
1072
+ "step": 149
1073
+ },
1074
+ {
1075
+ "epoch": 2.6315789473684212,
1076
+ "grad_norm": 18.148771286010742,
1077
+ "learning_rate": 2.724770642201835e-05,
1078
+ "loss": 1.0759,
1079
+ "step": 150
1080
+ },
1081
+ {
1082
+ "epoch": 2.6491228070175437,
1083
+ "grad_norm": 28.365219116210938,
1084
+ "learning_rate": 2.7220183486238532e-05,
1085
+ "loss": 0.9998,
1086
+ "step": 151
1087
+ },
1088
+ {
1089
+ "epoch": 2.6666666666666665,
1090
+ "grad_norm": 10.812620162963867,
1091
+ "learning_rate": 2.7192660550458718e-05,
1092
+ "loss": 0.9368,
1093
+ "step": 152
1094
+ },
1095
+ {
1096
+ "epoch": 2.6842105263157894,
1097
+ "grad_norm": 32.16633605957031,
1098
+ "learning_rate": 2.7165137614678897e-05,
1099
+ "loss": 1.0304,
1100
+ "step": 153
1101
+ },
1102
+ {
1103
+ "epoch": 2.7017543859649122,
1104
+ "grad_norm": 63.48967742919922,
1105
+ "learning_rate": 2.7137614678899083e-05,
1106
+ "loss": 1.0795,
1107
+ "step": 154
1108
+ },
1109
+ {
1110
+ "epoch": 2.719298245614035,
1111
+ "grad_norm": 4.874389171600342,
1112
+ "learning_rate": 2.711009174311927e-05,
1113
+ "loss": 1.062,
1114
+ "step": 155
1115
+ },
1116
+ {
1117
+ "epoch": 2.736842105263158,
1118
+ "grad_norm": 18.235118865966797,
1119
+ "learning_rate": 2.708256880733945e-05,
1120
+ "loss": 1.3184,
1121
+ "step": 156
1122
+ },
1123
+ {
1124
+ "epoch": 2.754385964912281,
1125
+ "grad_norm": 22.853351593017578,
1126
+ "learning_rate": 2.7055045871559634e-05,
1127
+ "loss": 1.0108,
1128
+ "step": 157
1129
+ },
1130
+ {
1131
+ "epoch": 2.7719298245614032,
1132
+ "grad_norm": 50.805885314941406,
1133
+ "learning_rate": 2.7027522935779817e-05,
1134
+ "loss": 0.914,
1135
+ "step": 158
1136
+ },
1137
+ {
1138
+ "epoch": 2.7894736842105265,
1139
+ "grad_norm": 5.718901634216309,
1140
+ "learning_rate": 2.7000000000000002e-05,
1141
+ "loss": 0.7322,
1142
+ "step": 159
1143
+ },
1144
+ {
1145
+ "epoch": 2.807017543859649,
1146
+ "grad_norm": 60.18952178955078,
1147
+ "learning_rate": 2.6972477064220185e-05,
1148
+ "loss": 1.088,
1149
+ "step": 160
1150
+ },
1151
+ {
1152
+ "epoch": 2.824561403508772,
1153
+ "grad_norm": 5.959671974182129,
1154
+ "learning_rate": 2.6944954128440367e-05,
1155
+ "loss": 0.5275,
1156
+ "step": 161
1157
+ },
1158
+ {
1159
+ "epoch": 2.8421052631578947,
1160
+ "grad_norm": 2.963286876678467,
1161
+ "learning_rate": 2.691743119266055e-05,
1162
+ "loss": 0.8271,
1163
+ "step": 162
1164
+ },
1165
+ {
1166
+ "epoch": 2.8596491228070176,
1167
+ "grad_norm": 18.080677032470703,
1168
+ "learning_rate": 2.6889908256880736e-05,
1169
+ "loss": 0.7593,
1170
+ "step": 163
1171
+ },
1172
+ {
1173
+ "epoch": 2.8771929824561404,
1174
+ "grad_norm": 3.0679006576538086,
1175
+ "learning_rate": 2.6862385321100918e-05,
1176
+ "loss": 0.582,
1177
+ "step": 164
1178
+ },
1179
+ {
1180
+ "epoch": 2.8947368421052633,
1181
+ "grad_norm": 27.97750473022461,
1182
+ "learning_rate": 2.68348623853211e-05,
1183
+ "loss": 1.0477,
1184
+ "step": 165
1185
+ },
1186
+ {
1187
+ "epoch": 2.912280701754386,
1188
+ "grad_norm": 9.784193992614746,
1189
+ "learning_rate": 2.6807339449541286e-05,
1190
+ "loss": 0.7083,
1191
+ "step": 166
1192
+ },
1193
+ {
1194
+ "epoch": 2.9298245614035086,
1195
+ "grad_norm": 7.2379045486450195,
1196
+ "learning_rate": 2.677981651376147e-05,
1197
+ "loss": 0.6767,
1198
+ "step": 167
1199
+ },
1200
+ {
1201
+ "epoch": 2.9473684210526314,
1202
+ "grad_norm": 17.974803924560547,
1203
+ "learning_rate": 2.6752293577981655e-05,
1204
+ "loss": 1.117,
1205
+ "step": 168
1206
+ },
1207
+ {
1208
+ "epoch": 2.9649122807017543,
1209
+ "grad_norm": 8.451967239379883,
1210
+ "learning_rate": 2.6724770642201834e-05,
1211
+ "loss": 1.0163,
1212
+ "step": 169
1213
+ },
1214
+ {
1215
+ "epoch": 2.982456140350877,
1216
+ "grad_norm": 16.834774017333984,
1217
+ "learning_rate": 2.669724770642202e-05,
1218
+ "loss": 0.642,
1219
+ "step": 170
1220
+ },
1221
+ {
1222
+ "epoch": 3.0,
1223
+ "grad_norm": 10.743306159973145,
1224
+ "learning_rate": 2.6669724770642202e-05,
1225
+ "loss": 0.8485,
1226
+ "step": 171
1227
+ },
1228
+ {
1229
+ "epoch": 3.0,
1230
+ "eval_accuracy": 0.5938697318007663,
1231
+ "eval_loss": 1.042966604232788,
1232
+ "eval_roc_auc": 0.8370073235789949,
1233
+ "eval_runtime": 9.8627,
1234
+ "eval_samples_per_second": 26.463,
1235
+ "eval_steps_per_second": 1.724,
1236
+ "step": 171
1237
+ },
1238
+ {
1239
+ "epoch": 3.017543859649123,
1240
+ "grad_norm": 7.450799942016602,
1241
+ "learning_rate": 2.6642201834862388e-05,
1242
+ "loss": 0.7936,
1243
+ "step": 172
1244
+ },
1245
+ {
1246
+ "epoch": 3.0350877192982457,
1247
+ "grad_norm": 13.938980102539062,
1248
+ "learning_rate": 2.6614678899082567e-05,
1249
+ "loss": 0.9981,
1250
+ "step": 173
1251
+ },
1252
+ {
1253
+ "epoch": 3.0526315789473686,
1254
+ "grad_norm": 10.046002388000488,
1255
+ "learning_rate": 2.6587155963302753e-05,
1256
+ "loss": 0.9664,
1257
+ "step": 174
1258
+ },
1259
+ {
1260
+ "epoch": 3.0701754385964914,
1261
+ "grad_norm": 2.47877836227417,
1262
+ "learning_rate": 2.6559633027522936e-05,
1263
+ "loss": 0.4564,
1264
+ "step": 175
1265
+ },
1266
+ {
1267
+ "epoch": 3.087719298245614,
1268
+ "grad_norm": 4.682964324951172,
1269
+ "learning_rate": 2.653211009174312e-05,
1270
+ "loss": 0.7715,
1271
+ "step": 176
1272
+ },
1273
+ {
1274
+ "epoch": 3.1052631578947367,
1275
+ "grad_norm": 31.68276023864746,
1276
+ "learning_rate": 2.6504587155963304e-05,
1277
+ "loss": 1.0283,
1278
+ "step": 177
1279
+ },
1280
+ {
1281
+ "epoch": 3.1228070175438596,
1282
+ "grad_norm": 6.329732418060303,
1283
+ "learning_rate": 2.6477064220183486e-05,
1284
+ "loss": 0.6996,
1285
+ "step": 178
1286
+ },
1287
+ {
1288
+ "epoch": 3.1403508771929824,
1289
+ "grad_norm": 14.793739318847656,
1290
+ "learning_rate": 2.6449541284403672e-05,
1291
+ "loss": 0.5858,
1292
+ "step": 179
1293
+ },
1294
+ {
1295
+ "epoch": 3.1578947368421053,
1296
+ "grad_norm": 123.91106414794922,
1297
+ "learning_rate": 2.6422018348623855e-05,
1298
+ "loss": 0.8345,
1299
+ "step": 180
1300
+ },
1301
+ {
1302
+ "epoch": 3.175438596491228,
1303
+ "grad_norm": 10.042440414428711,
1304
+ "learning_rate": 2.6394495412844037e-05,
1305
+ "loss": 0.8356,
1306
+ "step": 181
1307
+ },
1308
+ {
1309
+ "epoch": 3.192982456140351,
1310
+ "grad_norm": 15.220785140991211,
1311
+ "learning_rate": 2.636697247706422e-05,
1312
+ "loss": 0.5512,
1313
+ "step": 182
1314
+ },
1315
+ {
1316
+ "epoch": 3.2105263157894735,
1317
+ "grad_norm": 9.333229064941406,
1318
+ "learning_rate": 2.6339449541284406e-05,
1319
+ "loss": 0.6247,
1320
+ "step": 183
1321
+ },
1322
+ {
1323
+ "epoch": 3.2280701754385963,
1324
+ "grad_norm": 13.827474594116211,
1325
+ "learning_rate": 2.6311926605504588e-05,
1326
+ "loss": 1.0353,
1327
+ "step": 184
1328
+ },
1329
+ {
1330
+ "epoch": 3.245614035087719,
1331
+ "grad_norm": 4.803915500640869,
1332
+ "learning_rate": 2.628440366972477e-05,
1333
+ "loss": 0.8724,
1334
+ "step": 185
1335
+ },
1336
+ {
1337
+ "epoch": 3.263157894736842,
1338
+ "grad_norm": 9.242655754089355,
1339
+ "learning_rate": 2.6256880733944953e-05,
1340
+ "loss": 0.7395,
1341
+ "step": 186
1342
+ },
1343
+ {
1344
+ "epoch": 3.280701754385965,
1345
+ "grad_norm": 82.34434509277344,
1346
+ "learning_rate": 2.622935779816514e-05,
1347
+ "loss": 0.7797,
1348
+ "step": 187
1349
+ },
1350
+ {
1351
+ "epoch": 3.2982456140350878,
1352
+ "grad_norm": 29.72227668762207,
1353
+ "learning_rate": 2.6201834862385325e-05,
1354
+ "loss": 0.9626,
1355
+ "step": 188
1356
+ },
1357
+ {
1358
+ "epoch": 3.3157894736842106,
1359
+ "grad_norm": 30.35354232788086,
1360
+ "learning_rate": 2.6174311926605504e-05,
1361
+ "loss": 1.0387,
1362
+ "step": 189
1363
+ },
1364
+ {
1365
+ "epoch": 3.3333333333333335,
1366
+ "grad_norm": 9.3469820022583,
1367
+ "learning_rate": 2.614678899082569e-05,
1368
+ "loss": 0.9555,
1369
+ "step": 190
1370
+ },
1371
+ {
1372
+ "epoch": 3.3508771929824563,
1373
+ "grad_norm": 10.0740327835083,
1374
+ "learning_rate": 2.6119266055045872e-05,
1375
+ "loss": 0.7608,
1376
+ "step": 191
1377
+ },
1378
+ {
1379
+ "epoch": 3.3684210526315788,
1380
+ "grad_norm": 16.590267181396484,
1381
+ "learning_rate": 2.6091743119266058e-05,
1382
+ "loss": 0.8894,
1383
+ "step": 192
1384
+ },
1385
+ {
1386
+ "epoch": 3.3859649122807016,
1387
+ "grad_norm": 6.453067779541016,
1388
+ "learning_rate": 2.6064220183486237e-05,
1389
+ "loss": 0.6409,
1390
+ "step": 193
1391
+ },
1392
+ {
1393
+ "epoch": 3.4035087719298245,
1394
+ "grad_norm": 18.96304702758789,
1395
+ "learning_rate": 2.6036697247706423e-05,
1396
+ "loss": 1.1674,
1397
+ "step": 194
1398
+ },
1399
+ {
1400
+ "epoch": 3.4210526315789473,
1401
+ "grad_norm": 4.519999027252197,
1402
+ "learning_rate": 2.6009174311926605e-05,
1403
+ "loss": 0.9088,
1404
+ "step": 195
1405
+ },
1406
+ {
1407
+ "epoch": 3.43859649122807,
1408
+ "grad_norm": 14.560205459594727,
1409
+ "learning_rate": 2.598165137614679e-05,
1410
+ "loss": 1.0394,
1411
+ "step": 196
1412
+ },
1413
+ {
1414
+ "epoch": 3.456140350877193,
1415
+ "grad_norm": 9.643004417419434,
1416
+ "learning_rate": 2.5954128440366974e-05,
1417
+ "loss": 0.7166,
1418
+ "step": 197
1419
+ },
1420
+ {
1421
+ "epoch": 3.473684210526316,
1422
+ "grad_norm": 6.686179161071777,
1423
+ "learning_rate": 2.5926605504587156e-05,
1424
+ "loss": 0.9138,
1425
+ "step": 198
1426
+ },
1427
+ {
1428
+ "epoch": 3.4912280701754383,
1429
+ "grad_norm": 6.065926551818848,
1430
+ "learning_rate": 2.5899082568807342e-05,
1431
+ "loss": 0.5985,
1432
+ "step": 199
1433
+ },
1434
+ {
1435
+ "epoch": 3.5087719298245617,
1436
+ "grad_norm": 8.297146797180176,
1437
+ "learning_rate": 2.5871559633027525e-05,
1438
+ "loss": 0.8459,
1439
+ "step": 200
1440
+ },
1441
+ {
1442
+ "epoch": 3.526315789473684,
1443
+ "grad_norm": 28.6617431640625,
1444
+ "learning_rate": 2.5844036697247707e-05,
1445
+ "loss": 0.4553,
1446
+ "step": 201
1447
+ },
1448
+ {
1449
+ "epoch": 3.543859649122807,
1450
+ "grad_norm": 8.133190155029297,
1451
+ "learning_rate": 2.581651376146789e-05,
1452
+ "loss": 0.8616,
1453
+ "step": 202
1454
+ },
1455
+ {
1456
+ "epoch": 3.56140350877193,
1457
+ "grad_norm": 7.848640441894531,
1458
+ "learning_rate": 2.5788990825688075e-05,
1459
+ "loss": 1.0041,
1460
+ "step": 203
1461
+ },
1462
+ {
1463
+ "epoch": 3.5789473684210527,
1464
+ "grad_norm": 12.316092491149902,
1465
+ "learning_rate": 2.5761467889908258e-05,
1466
+ "loss": 0.8715,
1467
+ "step": 204
1468
+ },
1469
+ {
1470
+ "epoch": 3.5964912280701755,
1471
+ "grad_norm": 7.93389892578125,
1472
+ "learning_rate": 2.573394495412844e-05,
1473
+ "loss": 0.7744,
1474
+ "step": 205
1475
+ },
1476
+ {
1477
+ "epoch": 3.6140350877192984,
1478
+ "grad_norm": 9.929866790771484,
1479
+ "learning_rate": 2.5706422018348623e-05,
1480
+ "loss": 0.6355,
1481
+ "step": 206
1482
+ },
1483
+ {
1484
+ "epoch": 3.6315789473684212,
1485
+ "grad_norm": 11.27926254272461,
1486
+ "learning_rate": 2.567889908256881e-05,
1487
+ "loss": 1.0616,
1488
+ "step": 207
1489
+ },
1490
+ {
1491
+ "epoch": 3.6491228070175437,
1492
+ "grad_norm": 12.538276672363281,
1493
+ "learning_rate": 2.5651376146788994e-05,
1494
+ "loss": 0.8681,
1495
+ "step": 208
1496
+ },
1497
+ {
1498
+ "epoch": 3.6666666666666665,
1499
+ "grad_norm": 14.14212703704834,
1500
+ "learning_rate": 2.5623853211009174e-05,
1501
+ "loss": 0.8615,
1502
+ "step": 209
1503
+ },
1504
+ {
1505
+ "epoch": 3.6842105263157894,
1506
+ "grad_norm": 12.915414810180664,
1507
+ "learning_rate": 2.559633027522936e-05,
1508
+ "loss": 0.905,
1509
+ "step": 210
1510
+ },
1511
+ {
1512
+ "epoch": 3.7017543859649122,
1513
+ "grad_norm": 6.382973670959473,
1514
+ "learning_rate": 2.5568807339449542e-05,
1515
+ "loss": 0.8229,
1516
+ "step": 211
1517
+ },
1518
+ {
1519
+ "epoch": 3.719298245614035,
1520
+ "grad_norm": 8.958497047424316,
1521
+ "learning_rate": 2.5541284403669728e-05,
1522
+ "loss": 0.8367,
1523
+ "step": 212
1524
+ },
1525
+ {
1526
+ "epoch": 3.736842105263158,
1527
+ "grad_norm": 13.200749397277832,
1528
+ "learning_rate": 2.5513761467889907e-05,
1529
+ "loss": 0.7738,
1530
+ "step": 213
1531
+ },
1532
+ {
1533
+ "epoch": 3.754385964912281,
1534
+ "grad_norm": 25.975767135620117,
1535
+ "learning_rate": 2.5486238532110093e-05,
1536
+ "loss": 1.1707,
1537
+ "step": 214
1538
+ },
1539
+ {
1540
+ "epoch": 3.7719298245614032,
1541
+ "grad_norm": 46.98295211791992,
1542
+ "learning_rate": 2.5458715596330275e-05,
1543
+ "loss": 0.815,
1544
+ "step": 215
1545
+ },
1546
+ {
1547
+ "epoch": 3.7894736842105265,
1548
+ "grad_norm": 6.342257499694824,
1549
+ "learning_rate": 2.543119266055046e-05,
1550
+ "loss": 0.9045,
1551
+ "step": 216
1552
+ },
1553
+ {
1554
+ "epoch": 3.807017543859649,
1555
+ "grad_norm": 11.122331619262695,
1556
+ "learning_rate": 2.540366972477064e-05,
1557
+ "loss": 0.4915,
1558
+ "step": 217
1559
+ },
1560
+ {
1561
+ "epoch": 3.824561403508772,
1562
+ "grad_norm": 11.39393424987793,
1563
+ "learning_rate": 2.5376146788990826e-05,
1564
+ "loss": 0.8584,
1565
+ "step": 218
1566
+ },
1567
+ {
1568
+ "epoch": 3.8421052631578947,
1569
+ "grad_norm": 16.47450828552246,
1570
+ "learning_rate": 2.5348623853211012e-05,
1571
+ "loss": 0.9062,
1572
+ "step": 219
1573
+ },
1574
+ {
1575
+ "epoch": 3.8596491228070176,
1576
+ "grad_norm": 8.9651517868042,
1577
+ "learning_rate": 2.5321100917431194e-05,
1578
+ "loss": 0.6332,
1579
+ "step": 220
1580
+ },
1581
+ {
1582
+ "epoch": 3.8771929824561404,
1583
+ "grad_norm": 14.864767074584961,
1584
+ "learning_rate": 2.5293577981651377e-05,
1585
+ "loss": 0.7327,
1586
+ "step": 221
1587
+ },
1588
+ {
1589
+ "epoch": 3.8947368421052633,
1590
+ "grad_norm": 13.695717811584473,
1591
+ "learning_rate": 2.526605504587156e-05,
1592
+ "loss": 0.8439,
1593
+ "step": 222
1594
+ },
1595
+ {
1596
+ "epoch": 3.912280701754386,
1597
+ "grad_norm": 10.765061378479004,
1598
+ "learning_rate": 2.5238532110091745e-05,
1599
+ "loss": 1.0394,
1600
+ "step": 223
1601
+ },
1602
+ {
1603
+ "epoch": 3.9298245614035086,
1604
+ "grad_norm": 15.805130958557129,
1605
+ "learning_rate": 2.5211009174311928e-05,
1606
+ "loss": 1.3535,
1607
+ "step": 224
1608
+ },
1609
+ {
1610
+ "epoch": 3.9473684210526314,
1611
+ "grad_norm": 19.387144088745117,
1612
+ "learning_rate": 2.518348623853211e-05,
1613
+ "loss": 0.7274,
1614
+ "step": 225
1615
+ },
1616
+ {
1617
+ "epoch": 3.9649122807017543,
1618
+ "grad_norm": 37.748497009277344,
1619
+ "learning_rate": 2.5155963302752293e-05,
1620
+ "loss": 0.7626,
1621
+ "step": 226
1622
+ },
1623
+ {
1624
+ "epoch": 3.982456140350877,
1625
+ "grad_norm": 16.551725387573242,
1626
+ "learning_rate": 2.512844036697248e-05,
1627
+ "loss": 0.9978,
1628
+ "step": 227
1629
+ },
1630
+ {
1631
+ "epoch": 4.0,
1632
+ "grad_norm": 25.66643524169922,
1633
+ "learning_rate": 2.510091743119266e-05,
1634
+ "loss": 0.7405,
1635
+ "step": 228
1636
+ },
1637
+ {
1638
+ "epoch": 4.0,
1639
+ "eval_accuracy": 0.6245210727969349,
1640
+ "eval_loss": 0.9546777606010437,
1641
+ "eval_roc_auc": 0.8586184139970431,
1642
+ "eval_runtime": 9.814,
1643
+ "eval_samples_per_second": 26.595,
1644
+ "eval_steps_per_second": 1.732,
1645
+ "step": 228
1646
+ },
1647
+ {
1648
+ "epoch": 4.017543859649122,
1649
+ "grad_norm": 4.81406307220459,
1650
+ "learning_rate": 2.5073394495412843e-05,
1651
+ "loss": 0.5165,
1652
+ "step": 229
1653
+ },
1654
+ {
1655
+ "epoch": 4.035087719298246,
1656
+ "grad_norm": 9.769265174865723,
1657
+ "learning_rate": 2.504587155963303e-05,
1658
+ "loss": 0.6222,
1659
+ "step": 230
1660
+ },
1661
+ {
1662
+ "epoch": 4.052631578947368,
1663
+ "grad_norm": 8.312880516052246,
1664
+ "learning_rate": 2.5018348623853212e-05,
1665
+ "loss": 0.6313,
1666
+ "step": 231
1667
+ },
1668
+ {
1669
+ "epoch": 4.0701754385964914,
1670
+ "grad_norm": 5.13502311706543,
1671
+ "learning_rate": 2.4990825688073398e-05,
1672
+ "loss": 0.7425,
1673
+ "step": 232
1674
+ },
1675
+ {
1676
+ "epoch": 4.087719298245614,
1677
+ "grad_norm": 19.665512084960938,
1678
+ "learning_rate": 2.4963302752293577e-05,
1679
+ "loss": 1.0384,
1680
+ "step": 233
1681
+ },
1682
+ {
1683
+ "epoch": 4.105263157894737,
1684
+ "grad_norm": 29.155986785888672,
1685
+ "learning_rate": 2.4935779816513763e-05,
1686
+ "loss": 1.0976,
1687
+ "step": 234
1688
+ },
1689
+ {
1690
+ "epoch": 4.12280701754386,
1691
+ "grad_norm": 11.831360816955566,
1692
+ "learning_rate": 2.4908256880733945e-05,
1693
+ "loss": 0.7877,
1694
+ "step": 235
1695
+ },
1696
+ {
1697
+ "epoch": 4.140350877192983,
1698
+ "grad_norm": 19.49347686767578,
1699
+ "learning_rate": 2.488073394495413e-05,
1700
+ "loss": 0.6279,
1701
+ "step": 236
1702
+ },
1703
+ {
1704
+ "epoch": 4.157894736842105,
1705
+ "grad_norm": 6.922645092010498,
1706
+ "learning_rate": 2.485321100917431e-05,
1707
+ "loss": 0.7153,
1708
+ "step": 237
1709
+ },
1710
+ {
1711
+ "epoch": 4.175438596491228,
1712
+ "grad_norm": 3.6549301147460938,
1713
+ "learning_rate": 2.4825688073394496e-05,
1714
+ "loss": 0.873,
1715
+ "step": 238
1716
+ },
1717
+ {
1718
+ "epoch": 4.192982456140351,
1719
+ "grad_norm": 10.724910736083984,
1720
+ "learning_rate": 2.4798165137614682e-05,
1721
+ "loss": 0.7728,
1722
+ "step": 239
1723
+ },
1724
+ {
1725
+ "epoch": 4.2105263157894735,
1726
+ "grad_norm": 3.7120726108551025,
1727
+ "learning_rate": 2.4770642201834864e-05,
1728
+ "loss": 0.7132,
1729
+ "step": 240
1730
+ },
1731
+ {
1732
+ "epoch": 4.228070175438597,
1733
+ "grad_norm": 29.73471450805664,
1734
+ "learning_rate": 2.4743119266055047e-05,
1735
+ "loss": 0.5634,
1736
+ "step": 241
1737
+ },
1738
+ {
1739
+ "epoch": 4.245614035087719,
1740
+ "grad_norm": 5.868150234222412,
1741
+ "learning_rate": 2.471559633027523e-05,
1742
+ "loss": 0.6967,
1743
+ "step": 242
1744
+ },
1745
+ {
1746
+ "epoch": 4.2631578947368425,
1747
+ "grad_norm": 9.828577995300293,
1748
+ "learning_rate": 2.4688073394495415e-05,
1749
+ "loss": 0.8814,
1750
+ "step": 243
1751
+ },
1752
+ {
1753
+ "epoch": 4.280701754385965,
1754
+ "grad_norm": 19.289398193359375,
1755
+ "learning_rate": 2.4660550458715597e-05,
1756
+ "loss": 0.6907,
1757
+ "step": 244
1758
+ },
1759
+ {
1760
+ "epoch": 4.298245614035087,
1761
+ "grad_norm": 18.95272445678711,
1762
+ "learning_rate": 2.463302752293578e-05,
1763
+ "loss": 0.7413,
1764
+ "step": 245
1765
+ },
1766
+ {
1767
+ "epoch": 4.315789473684211,
1768
+ "grad_norm": 9.175381660461426,
1769
+ "learning_rate": 2.4605504587155962e-05,
1770
+ "loss": 0.5549,
1771
+ "step": 246
1772
+ },
1773
+ {
1774
+ "epoch": 4.333333333333333,
1775
+ "grad_norm": 11.017579078674316,
1776
+ "learning_rate": 2.4577981651376148e-05,
1777
+ "loss": 0.5834,
1778
+ "step": 247
1779
+ },
1780
+ {
1781
+ "epoch": 4.350877192982456,
1782
+ "grad_norm": 4.641144752502441,
1783
+ "learning_rate": 2.455045871559633e-05,
1784
+ "loss": 0.7171,
1785
+ "step": 248
1786
+ },
1787
+ {
1788
+ "epoch": 4.368421052631579,
1789
+ "grad_norm": 30.48017120361328,
1790
+ "learning_rate": 2.4522935779816513e-05,
1791
+ "loss": 0.5314,
1792
+ "step": 249
1793
+ },
1794
+ {
1795
+ "epoch": 4.385964912280702,
1796
+ "grad_norm": 5.151516914367676,
1797
+ "learning_rate": 2.44954128440367e-05,
1798
+ "loss": 0.822,
1799
+ "step": 250
1800
+ },
1801
+ {
1802
+ "epoch": 4.4035087719298245,
1803
+ "grad_norm": 4.152652263641357,
1804
+ "learning_rate": 2.446788990825688e-05,
1805
+ "loss": 0.8574,
1806
+ "step": 251
1807
+ },
1808
+ {
1809
+ "epoch": 4.421052631578947,
1810
+ "grad_norm": 26.325111389160156,
1811
+ "learning_rate": 2.4440366972477067e-05,
1812
+ "loss": 0.5536,
1813
+ "step": 252
1814
+ },
1815
+ {
1816
+ "epoch": 4.43859649122807,
1817
+ "grad_norm": 3.4374260902404785,
1818
+ "learning_rate": 2.4412844036697247e-05,
1819
+ "loss": 0.7328,
1820
+ "step": 253
1821
+ },
1822
+ {
1823
+ "epoch": 4.456140350877193,
1824
+ "grad_norm": 7.868347644805908,
1825
+ "learning_rate": 2.4385321100917432e-05,
1826
+ "loss": 0.8698,
1827
+ "step": 254
1828
+ },
1829
+ {
1830
+ "epoch": 4.473684210526316,
1831
+ "grad_norm": 14.990694046020508,
1832
+ "learning_rate": 2.4357798165137615e-05,
1833
+ "loss": 0.9928,
1834
+ "step": 255
1835
+ },
1836
+ {
1837
+ "epoch": 4.491228070175438,
1838
+ "grad_norm": 55.8467903137207,
1839
+ "learning_rate": 2.43302752293578e-05,
1840
+ "loss": 0.8542,
1841
+ "step": 256
1842
+ },
1843
+ {
1844
+ "epoch": 4.508771929824562,
1845
+ "grad_norm": 15.516489028930664,
1846
+ "learning_rate": 2.430275229357798e-05,
1847
+ "loss": 0.935,
1848
+ "step": 257
1849
+ },
1850
+ {
1851
+ "epoch": 4.526315789473684,
1852
+ "grad_norm": 31.769515991210938,
1853
+ "learning_rate": 2.4275229357798166e-05,
1854
+ "loss": 0.5695,
1855
+ "step": 258
1856
+ },
1857
+ {
1858
+ "epoch": 4.543859649122807,
1859
+ "grad_norm": 6.501729488372803,
1860
+ "learning_rate": 2.4247706422018348e-05,
1861
+ "loss": 0.7384,
1862
+ "step": 259
1863
+ },
1864
+ {
1865
+ "epoch": 4.56140350877193,
1866
+ "grad_norm": 13.194774627685547,
1867
+ "learning_rate": 2.4220183486238534e-05,
1868
+ "loss": 0.7311,
1869
+ "step": 260
1870
+ },
1871
+ {
1872
+ "epoch": 4.578947368421053,
1873
+ "grad_norm": 9.523067474365234,
1874
+ "learning_rate": 2.4192660550458717e-05,
1875
+ "loss": 0.8254,
1876
+ "step": 261
1877
+ },
1878
+ {
1879
+ "epoch": 4.5964912280701755,
1880
+ "grad_norm": 40.60873794555664,
1881
+ "learning_rate": 2.41651376146789e-05,
1882
+ "loss": 0.7756,
1883
+ "step": 262
1884
+ },
1885
+ {
1886
+ "epoch": 4.614035087719298,
1887
+ "grad_norm": 20.81313133239746,
1888
+ "learning_rate": 2.4137614678899085e-05,
1889
+ "loss": 0.7474,
1890
+ "step": 263
1891
+ },
1892
+ {
1893
+ "epoch": 4.631578947368421,
1894
+ "grad_norm": 30.596458435058594,
1895
+ "learning_rate": 2.4110091743119267e-05,
1896
+ "loss": 0.8261,
1897
+ "step": 264
1898
+ },
1899
+ {
1900
+ "epoch": 4.649122807017544,
1901
+ "grad_norm": 36.82753372192383,
1902
+ "learning_rate": 2.408256880733945e-05,
1903
+ "loss": 0.7422,
1904
+ "step": 265
1905
+ },
1906
+ {
1907
+ "epoch": 4.666666666666667,
1908
+ "grad_norm": 7.798766136169434,
1909
+ "learning_rate": 2.4055045871559632e-05,
1910
+ "loss": 0.7433,
1911
+ "step": 266
1912
+ },
1913
+ {
1914
+ "epoch": 4.684210526315789,
1915
+ "grad_norm": 10.442023277282715,
1916
+ "learning_rate": 2.4027522935779818e-05,
1917
+ "loss": 0.9382,
1918
+ "step": 267
1919
+ },
1920
+ {
1921
+ "epoch": 4.701754385964913,
1922
+ "grad_norm": 7.129739761352539,
1923
+ "learning_rate": 2.4e-05,
1924
+ "loss": 0.4544,
1925
+ "step": 268
1926
+ },
1927
+ {
1928
+ "epoch": 4.719298245614035,
1929
+ "grad_norm": 6.29140567779541,
1930
+ "learning_rate": 2.3972477064220183e-05,
1931
+ "loss": 0.5828,
1932
+ "step": 269
1933
+ },
1934
+ {
1935
+ "epoch": 4.7368421052631575,
1936
+ "grad_norm": 14.20146369934082,
1937
+ "learning_rate": 2.3944954128440366e-05,
1938
+ "loss": 0.7544,
1939
+ "step": 270
1940
+ },
1941
+ {
1942
+ "epoch": 4.754385964912281,
1943
+ "grad_norm": 25.83724594116211,
1944
+ "learning_rate": 2.391743119266055e-05,
1945
+ "loss": 0.8725,
1946
+ "step": 271
1947
+ },
1948
+ {
1949
+ "epoch": 4.771929824561403,
1950
+ "grad_norm": 19.661861419677734,
1951
+ "learning_rate": 2.3889908256880737e-05,
1952
+ "loss": 0.9125,
1953
+ "step": 272
1954
+ },
1955
+ {
1956
+ "epoch": 4.7894736842105265,
1957
+ "grad_norm": 7.706132888793945,
1958
+ "learning_rate": 2.3862385321100916e-05,
1959
+ "loss": 0.4359,
1960
+ "step": 273
1961
+ },
1962
+ {
1963
+ "epoch": 4.807017543859649,
1964
+ "grad_norm": 13.709455490112305,
1965
+ "learning_rate": 2.3834862385321102e-05,
1966
+ "loss": 0.6747,
1967
+ "step": 274
1968
+ },
1969
+ {
1970
+ "epoch": 4.824561403508772,
1971
+ "grad_norm": 27.501550674438477,
1972
+ "learning_rate": 2.3807339449541285e-05,
1973
+ "loss": 0.5231,
1974
+ "step": 275
1975
+ },
1976
+ {
1977
+ "epoch": 4.842105263157895,
1978
+ "grad_norm": 3.0259737968444824,
1979
+ "learning_rate": 2.377981651376147e-05,
1980
+ "loss": 0.5482,
1981
+ "step": 276
1982
+ },
1983
+ {
1984
+ "epoch": 4.859649122807017,
1985
+ "grad_norm": 22.053667068481445,
1986
+ "learning_rate": 2.375229357798165e-05,
1987
+ "loss": 0.6844,
1988
+ "step": 277
1989
+ },
1990
+ {
1991
+ "epoch": 4.87719298245614,
1992
+ "grad_norm": 21.116025924682617,
1993
+ "learning_rate": 2.3724770642201836e-05,
1994
+ "loss": 0.5924,
1995
+ "step": 278
1996
+ },
1997
+ {
1998
+ "epoch": 4.894736842105263,
1999
+ "grad_norm": 45.17483901977539,
2000
+ "learning_rate": 2.3697247706422018e-05,
2001
+ "loss": 0.9959,
2002
+ "step": 279
2003
+ },
2004
+ {
2005
+ "epoch": 4.912280701754386,
2006
+ "grad_norm": 2.0751166343688965,
2007
+ "learning_rate": 2.3669724770642204e-05,
2008
+ "loss": 0.4215,
2009
+ "step": 280
2010
+ },
2011
+ {
2012
+ "epoch": 4.9298245614035086,
2013
+ "grad_norm": 17.698915481567383,
2014
+ "learning_rate": 2.3642201834862386e-05,
2015
+ "loss": 0.8212,
2016
+ "step": 281
2017
+ },
2018
+ {
2019
+ "epoch": 4.947368421052632,
2020
+ "grad_norm": 11.810864448547363,
2021
+ "learning_rate": 2.361467889908257e-05,
2022
+ "loss": 0.7421,
2023
+ "step": 282
2024
+ },
2025
+ {
2026
+ "epoch": 4.964912280701754,
2027
+ "grad_norm": 25.751745223999023,
2028
+ "learning_rate": 2.3587155963302755e-05,
2029
+ "loss": 0.8161,
2030
+ "step": 283
2031
+ },
2032
+ {
2033
+ "epoch": 4.982456140350877,
2034
+ "grad_norm": 9.799457550048828,
2035
+ "learning_rate": 2.3559633027522937e-05,
2036
+ "loss": 0.6857,
2037
+ "step": 284
2038
+ },
2039
+ {
2040
+ "epoch": 5.0,
2041
+ "grad_norm": 10.592196464538574,
2042
+ "learning_rate": 2.353211009174312e-05,
2043
+ "loss": 0.8668,
2044
+ "step": 285
2045
+ },
2046
+ {
2047
+ "epoch": 5.0,
2048
+ "eval_accuracy": 0.7011494252873564,
2049
+ "eval_loss": 0.8508906364440918,
2050
+ "eval_roc_auc": 0.8698378210083979,
2051
+ "eval_runtime": 9.9199,
2052
+ "eval_samples_per_second": 26.311,
2053
+ "eval_steps_per_second": 1.714,
2054
+ "step": 285
2055
+ },
2056
+ {
2057
+ "epoch": 5.017543859649122,
2058
+ "grad_norm": 29.6278133392334,
2059
+ "learning_rate": 2.3504587155963302e-05,
2060
+ "loss": 0.8195,
2061
+ "step": 286
2062
+ },
2063
+ {
2064
+ "epoch": 5.035087719298246,
2065
+ "grad_norm": 8.088125228881836,
2066
+ "learning_rate": 2.3477064220183488e-05,
2067
+ "loss": 0.4163,
2068
+ "step": 287
2069
+ },
2070
+ {
2071
+ "epoch": 5.052631578947368,
2072
+ "grad_norm": 2.1910946369171143,
2073
+ "learning_rate": 2.344954128440367e-05,
2074
+ "loss": 0.4273,
2075
+ "step": 288
2076
+ },
2077
+ {
2078
+ "epoch": 5.0701754385964914,
2079
+ "grad_norm": 27.934329986572266,
2080
+ "learning_rate": 2.3422018348623853e-05,
2081
+ "loss": 0.7804,
2082
+ "step": 289
2083
+ },
2084
+ {
2085
+ "epoch": 5.087719298245614,
2086
+ "grad_norm": 28.2197265625,
2087
+ "learning_rate": 2.3394495412844035e-05,
2088
+ "loss": 0.5328,
2089
+ "step": 290
2090
+ },
2091
+ {
2092
+ "epoch": 5.105263157894737,
2093
+ "grad_norm": 20.080032348632812,
2094
+ "learning_rate": 2.336697247706422e-05,
2095
+ "loss": 0.9274,
2096
+ "step": 291
2097
+ },
2098
+ {
2099
+ "epoch": 5.12280701754386,
2100
+ "grad_norm": 8.307439804077148,
2101
+ "learning_rate": 2.3339449541284407e-05,
2102
+ "loss": 0.7575,
2103
+ "step": 292
2104
+ },
2105
+ {
2106
+ "epoch": 5.140350877192983,
2107
+ "grad_norm": 15.012224197387695,
2108
+ "learning_rate": 2.3311926605504586e-05,
2109
+ "loss": 0.6881,
2110
+ "step": 293
2111
+ },
2112
+ {
2113
+ "epoch": 5.157894736842105,
2114
+ "grad_norm": 19.6754093170166,
2115
+ "learning_rate": 2.3284403669724772e-05,
2116
+ "loss": 0.5243,
2117
+ "step": 294
2118
+ },
2119
+ {
2120
+ "epoch": 5.175438596491228,
2121
+ "grad_norm": 2.4565722942352295,
2122
+ "learning_rate": 2.3256880733944955e-05,
2123
+ "loss": 0.5858,
2124
+ "step": 295
2125
+ },
2126
+ {
2127
+ "epoch": 5.192982456140351,
2128
+ "grad_norm": 12.843107223510742,
2129
+ "learning_rate": 2.322935779816514e-05,
2130
+ "loss": 0.7095,
2131
+ "step": 296
2132
+ },
2133
+ {
2134
+ "epoch": 5.2105263157894735,
2135
+ "grad_norm": 3.9343743324279785,
2136
+ "learning_rate": 2.320183486238532e-05,
2137
+ "loss": 0.6592,
2138
+ "step": 297
2139
+ },
2140
+ {
2141
+ "epoch": 5.228070175438597,
2142
+ "grad_norm": 21.421518325805664,
2143
+ "learning_rate": 2.3174311926605505e-05,
2144
+ "loss": 0.6792,
2145
+ "step": 298
2146
+ },
2147
+ {
2148
+ "epoch": 5.245614035087719,
2149
+ "grad_norm": 29.099624633789062,
2150
+ "learning_rate": 2.3146788990825688e-05,
2151
+ "loss": 0.8901,
2152
+ "step": 299
2153
+ },
2154
+ {
2155
+ "epoch": 5.2631578947368425,
2156
+ "grad_norm": 28.176443099975586,
2157
+ "learning_rate": 2.3119266055045874e-05,
2158
+ "loss": 0.677,
2159
+ "step": 300
2160
+ },
2161
+ {
2162
+ "epoch": 5.280701754385965,
2163
+ "grad_norm": 49.84172439575195,
2164
+ "learning_rate": 2.3091743119266053e-05,
2165
+ "loss": 0.6627,
2166
+ "step": 301
2167
+ },
2168
+ {
2169
+ "epoch": 5.298245614035087,
2170
+ "grad_norm": 11.457085609436035,
2171
+ "learning_rate": 2.306422018348624e-05,
2172
+ "loss": 0.3993,
2173
+ "step": 302
2174
+ },
2175
+ {
2176
+ "epoch": 5.315789473684211,
2177
+ "grad_norm": 187.01239013671875,
2178
+ "learning_rate": 2.3036697247706425e-05,
2179
+ "loss": 0.6346,
2180
+ "step": 303
2181
+ },
2182
+ {
2183
+ "epoch": 5.333333333333333,
2184
+ "grad_norm": 3.9368183612823486,
2185
+ "learning_rate": 2.3009174311926607e-05,
2186
+ "loss": 0.8296,
2187
+ "step": 304
2188
+ },
2189
+ {
2190
+ "epoch": 5.350877192982456,
2191
+ "grad_norm": 100.2872085571289,
2192
+ "learning_rate": 2.298165137614679e-05,
2193
+ "loss": 0.8781,
2194
+ "step": 305
2195
+ },
2196
+ {
2197
+ "epoch": 5.368421052631579,
2198
+ "grad_norm": 24.209882736206055,
2199
+ "learning_rate": 2.2954128440366972e-05,
2200
+ "loss": 0.5137,
2201
+ "step": 306
2202
+ },
2203
+ {
2204
+ "epoch": 5.385964912280702,
2205
+ "grad_norm": 3.6952366828918457,
2206
+ "learning_rate": 2.2926605504587158e-05,
2207
+ "loss": 0.6469,
2208
+ "step": 307
2209
+ },
2210
+ {
2211
+ "epoch": 5.4035087719298245,
2212
+ "grad_norm": 29.038558959960938,
2213
+ "learning_rate": 2.289908256880734e-05,
2214
+ "loss": 0.8628,
2215
+ "step": 308
2216
+ },
2217
+ {
2218
+ "epoch": 5.421052631578947,
2219
+ "grad_norm": 40.4047737121582,
2220
+ "learning_rate": 2.2871559633027523e-05,
2221
+ "loss": 1.065,
2222
+ "step": 309
2223
+ },
2224
+ {
2225
+ "epoch": 5.43859649122807,
2226
+ "grad_norm": 43.313846588134766,
2227
+ "learning_rate": 2.2844036697247705e-05,
2228
+ "loss": 0.5773,
2229
+ "step": 310
2230
+ },
2231
+ {
2232
+ "epoch": 5.456140350877193,
2233
+ "grad_norm": 7.541952133178711,
2234
+ "learning_rate": 2.281651376146789e-05,
2235
+ "loss": 0.7124,
2236
+ "step": 311
2237
+ },
2238
+ {
2239
+ "epoch": 5.473684210526316,
2240
+ "grad_norm": 9.793838500976562,
2241
+ "learning_rate": 2.2788990825688074e-05,
2242
+ "loss": 0.5441,
2243
+ "step": 312
2244
+ },
2245
+ {
2246
+ "epoch": 5.491228070175438,
2247
+ "grad_norm": 17.21137809753418,
2248
+ "learning_rate": 2.2761467889908256e-05,
2249
+ "loss": 0.5927,
2250
+ "step": 313
2251
+ },
2252
+ {
2253
+ "epoch": 5.508771929824562,
2254
+ "grad_norm": 23.36025047302246,
2255
+ "learning_rate": 2.2733944954128442e-05,
2256
+ "loss": 0.8349,
2257
+ "step": 314
2258
+ },
2259
+ {
2260
+ "epoch": 5.526315789473684,
2261
+ "grad_norm": 10.170068740844727,
2262
+ "learning_rate": 2.2706422018348624e-05,
2263
+ "loss": 0.7946,
2264
+ "step": 315
2265
+ },
2266
+ {
2267
+ "epoch": 5.543859649122807,
2268
+ "grad_norm": 7.996306419372559,
2269
+ "learning_rate": 2.267889908256881e-05,
2270
+ "loss": 0.7086,
2271
+ "step": 316
2272
+ },
2273
+ {
2274
+ "epoch": 5.56140350877193,
2275
+ "grad_norm": 4.8672027587890625,
2276
+ "learning_rate": 2.265137614678899e-05,
2277
+ "loss": 0.4384,
2278
+ "step": 317
2279
+ },
2280
+ {
2281
+ "epoch": 5.578947368421053,
2282
+ "grad_norm": 14.543272972106934,
2283
+ "learning_rate": 2.2623853211009175e-05,
2284
+ "loss": 0.9153,
2285
+ "step": 318
2286
+ },
2287
+ {
2288
+ "epoch": 5.5964912280701755,
2289
+ "grad_norm": 28.248517990112305,
2290
+ "learning_rate": 2.2596330275229358e-05,
2291
+ "loss": 0.4361,
2292
+ "step": 319
2293
+ },
2294
+ {
2295
+ "epoch": 5.614035087719298,
2296
+ "grad_norm": 10.02695083618164,
2297
+ "learning_rate": 2.2568807339449544e-05,
2298
+ "loss": 0.5584,
2299
+ "step": 320
2300
+ },
2301
+ {
2302
+ "epoch": 5.631578947368421,
2303
+ "grad_norm": 255.10897827148438,
2304
+ "learning_rate": 2.2541284403669723e-05,
2305
+ "loss": 0.812,
2306
+ "step": 321
2307
+ },
2308
+ {
2309
+ "epoch": 5.649122807017544,
2310
+ "grad_norm": 29.070514678955078,
2311
+ "learning_rate": 2.251376146788991e-05,
2312
+ "loss": 0.7201,
2313
+ "step": 322
2314
+ },
2315
+ {
2316
+ "epoch": 5.666666666666667,
2317
+ "grad_norm": 70.94801330566406,
2318
+ "learning_rate": 2.2486238532110094e-05,
2319
+ "loss": 0.7983,
2320
+ "step": 323
2321
+ },
2322
+ {
2323
+ "epoch": 5.684210526315789,
2324
+ "grad_norm": 9.982966423034668,
2325
+ "learning_rate": 2.2458715596330277e-05,
2326
+ "loss": 0.3869,
2327
+ "step": 324
2328
+ },
2329
+ {
2330
+ "epoch": 5.701754385964913,
2331
+ "grad_norm": 44.79756546020508,
2332
+ "learning_rate": 2.243119266055046e-05,
2333
+ "loss": 0.5092,
2334
+ "step": 325
2335
+ },
2336
+ {
2337
+ "epoch": 5.719298245614035,
2338
+ "grad_norm": 14.900089263916016,
2339
+ "learning_rate": 2.2403669724770642e-05,
2340
+ "loss": 0.7798,
2341
+ "step": 326
2342
+ },
2343
+ {
2344
+ "epoch": 5.7368421052631575,
2345
+ "grad_norm": 5.473045825958252,
2346
+ "learning_rate": 2.2376146788990828e-05,
2347
+ "loss": 0.527,
2348
+ "step": 327
2349
+ },
2350
+ {
2351
+ "epoch": 5.754385964912281,
2352
+ "grad_norm": 5.292675495147705,
2353
+ "learning_rate": 2.234862385321101e-05,
2354
+ "loss": 0.599,
2355
+ "step": 328
2356
+ },
2357
+ {
2358
+ "epoch": 5.771929824561403,
2359
+ "grad_norm": 7.409575462341309,
2360
+ "learning_rate": 2.2321100917431193e-05,
2361
+ "loss": 0.6566,
2362
+ "step": 329
2363
+ },
2364
+ {
2365
+ "epoch": 5.7894736842105265,
2366
+ "grad_norm": 16.337003707885742,
2367
+ "learning_rate": 2.2293577981651375e-05,
2368
+ "loss": 0.421,
2369
+ "step": 330
2370
+ },
2371
+ {
2372
+ "epoch": 5.807017543859649,
2373
+ "grad_norm": 10.473221778869629,
2374
+ "learning_rate": 2.226605504587156e-05,
2375
+ "loss": 0.6791,
2376
+ "step": 331
2377
+ },
2378
+ {
2379
+ "epoch": 5.824561403508772,
2380
+ "grad_norm": 24.943050384521484,
2381
+ "learning_rate": 2.2238532110091743e-05,
2382
+ "loss": 1.0399,
2383
+ "step": 332
2384
+ },
2385
+ {
2386
+ "epoch": 5.842105263157895,
2387
+ "grad_norm": 3.1360652446746826,
2388
+ "learning_rate": 2.2211009174311926e-05,
2389
+ "loss": 0.4585,
2390
+ "step": 333
2391
+ },
2392
+ {
2393
+ "epoch": 5.859649122807017,
2394
+ "grad_norm": 5.742803573608398,
2395
+ "learning_rate": 2.2183486238532112e-05,
2396
+ "loss": 0.7424,
2397
+ "step": 334
2398
+ },
2399
+ {
2400
+ "epoch": 5.87719298245614,
2401
+ "grad_norm": 10.02749252319336,
2402
+ "learning_rate": 2.2155963302752294e-05,
2403
+ "loss": 0.5505,
2404
+ "step": 335
2405
+ },
2406
+ {
2407
+ "epoch": 5.894736842105263,
2408
+ "grad_norm": 11.549918174743652,
2409
+ "learning_rate": 2.212844036697248e-05,
2410
+ "loss": 0.7165,
2411
+ "step": 336
2412
+ },
2413
+ {
2414
+ "epoch": 5.912280701754386,
2415
+ "grad_norm": 11.704695701599121,
2416
+ "learning_rate": 2.210091743119266e-05,
2417
+ "loss": 0.5897,
2418
+ "step": 337
2419
+ },
2420
+ {
2421
+ "epoch": 5.9298245614035086,
2422
+ "grad_norm": 8.045721054077148,
2423
+ "learning_rate": 2.2073394495412845e-05,
2424
+ "loss": 0.5606,
2425
+ "step": 338
2426
+ },
2427
+ {
2428
+ "epoch": 5.947368421052632,
2429
+ "grad_norm": 6.550736904144287,
2430
+ "learning_rate": 2.2045871559633028e-05,
2431
+ "loss": 0.4498,
2432
+ "step": 339
2433
+ },
2434
+ {
2435
+ "epoch": 5.964912280701754,
2436
+ "grad_norm": 7.8976335525512695,
2437
+ "learning_rate": 2.2018348623853213e-05,
2438
+ "loss": 0.6578,
2439
+ "step": 340
2440
+ },
2441
+ {
2442
+ "epoch": 5.982456140350877,
2443
+ "grad_norm": 19.857566833496094,
2444
+ "learning_rate": 2.1990825688073392e-05,
2445
+ "loss": 0.6241,
2446
+ "step": 341
2447
+ },
2448
+ {
2449
+ "epoch": 6.0,
2450
+ "grad_norm": 389.2672119140625,
2451
+ "learning_rate": 2.196330275229358e-05,
2452
+ "loss": 0.5834,
2453
+ "step": 342
2454
+ },
2455
+ {
2456
+ "epoch": 6.0,
2457
+ "eval_accuracy": 0.6436781609195402,
2458
+ "eval_loss": 0.9111022353172302,
2459
+ "eval_roc_auc": 0.8661937516704385,
2460
+ "eval_runtime": 8.4523,
2461
+ "eval_samples_per_second": 30.879,
2462
+ "eval_steps_per_second": 2.011,
2463
+ "step": 342
2464
+ },
2465
+ {
2466
+ "epoch": 6.017543859649122,
2467
+ "grad_norm": 6.004997253417969,
2468
+ "learning_rate": 2.193577981651376e-05,
2469
+ "loss": 0.4383,
2470
+ "step": 343
2471
+ },
2472
+ {
2473
+ "epoch": 6.035087719298246,
2474
+ "grad_norm": 32.85868453979492,
2475
+ "learning_rate": 2.1908256880733947e-05,
2476
+ "loss": 0.8127,
2477
+ "step": 344
2478
+ },
2479
+ {
2480
+ "epoch": 6.052631578947368,
2481
+ "grad_norm": 22.06620216369629,
2482
+ "learning_rate": 2.188073394495413e-05,
2483
+ "loss": 0.5588,
2484
+ "step": 345
2485
+ },
2486
+ {
2487
+ "epoch": 6.0701754385964914,
2488
+ "grad_norm": 12.496269226074219,
2489
+ "learning_rate": 2.185321100917431e-05,
2490
+ "loss": 0.7202,
2491
+ "step": 346
2492
+ },
2493
+ {
2494
+ "epoch": 6.087719298245614,
2495
+ "grad_norm": 39.94806671142578,
2496
+ "learning_rate": 2.1825688073394497e-05,
2497
+ "loss": 0.4362,
2498
+ "step": 347
2499
+ },
2500
+ {
2501
+ "epoch": 6.105263157894737,
2502
+ "grad_norm": 5.034066677093506,
2503
+ "learning_rate": 2.179816513761468e-05,
2504
+ "loss": 0.6235,
2505
+ "step": 348
2506
+ },
2507
+ {
2508
+ "epoch": 6.12280701754386,
2509
+ "grad_norm": 98.4699935913086,
2510
+ "learning_rate": 2.1770642201834862e-05,
2511
+ "loss": 0.8505,
2512
+ "step": 349
2513
+ },
2514
+ {
2515
+ "epoch": 6.140350877192983,
2516
+ "grad_norm": 22.5294132232666,
2517
+ "learning_rate": 2.1743119266055045e-05,
2518
+ "loss": 0.6741,
2519
+ "step": 350
2520
+ },
2521
+ {
2522
+ "epoch": 6.157894736842105,
2523
+ "grad_norm": 33.31329345703125,
2524
+ "learning_rate": 2.171559633027523e-05,
2525
+ "loss": 0.6783,
2526
+ "step": 351
2527
+ },
2528
+ {
2529
+ "epoch": 6.175438596491228,
2530
+ "grad_norm": 2.9709410667419434,
2531
+ "learning_rate": 2.1688073394495413e-05,
2532
+ "loss": 0.3269,
2533
+ "step": 352
2534
+ },
2535
+ {
2536
+ "epoch": 6.192982456140351,
2537
+ "grad_norm": 14.349902153015137,
2538
+ "learning_rate": 2.1660550458715596e-05,
2539
+ "loss": 0.5928,
2540
+ "step": 353
2541
+ },
2542
+ {
2543
+ "epoch": 6.2105263157894735,
2544
+ "grad_norm": 6.400623321533203,
2545
+ "learning_rate": 2.163302752293578e-05,
2546
+ "loss": 0.5708,
2547
+ "step": 354
2548
+ },
2549
+ {
2550
+ "epoch": 6.228070175438597,
2551
+ "grad_norm": 40.32659912109375,
2552
+ "learning_rate": 2.1605504587155964e-05,
2553
+ "loss": 0.4459,
2554
+ "step": 355
2555
+ },
2556
+ {
2557
+ "epoch": 6.245614035087719,
2558
+ "grad_norm": 16.54581069946289,
2559
+ "learning_rate": 2.157798165137615e-05,
2560
+ "loss": 1.0475,
2561
+ "step": 356
2562
+ },
2563
+ {
2564
+ "epoch": 6.2631578947368425,
2565
+ "grad_norm": 7.048498630523682,
2566
+ "learning_rate": 2.155045871559633e-05,
2567
+ "loss": 0.6128,
2568
+ "step": 357
2569
+ },
2570
+ {
2571
+ "epoch": 6.280701754385965,
2572
+ "grad_norm": 18.782760620117188,
2573
+ "learning_rate": 2.1522935779816515e-05,
2574
+ "loss": 0.6483,
2575
+ "step": 358
2576
+ },
2577
+ {
2578
+ "epoch": 6.298245614035087,
2579
+ "grad_norm": 13.411852836608887,
2580
+ "learning_rate": 2.1495412844036697e-05,
2581
+ "loss": 0.6957,
2582
+ "step": 359
2583
+ },
2584
+ {
2585
+ "epoch": 6.315789473684211,
2586
+ "grad_norm": 3.7218034267425537,
2587
+ "learning_rate": 2.1467889908256883e-05,
2588
+ "loss": 0.268,
2589
+ "step": 360
2590
+ },
2591
+ {
2592
+ "epoch": 6.333333333333333,
2593
+ "grad_norm": 5.883011817932129,
2594
+ "learning_rate": 2.1440366972477062e-05,
2595
+ "loss": 0.5415,
2596
+ "step": 361
2597
+ },
2598
+ {
2599
+ "epoch": 6.350877192982456,
2600
+ "grad_norm": 6.700387477874756,
2601
+ "learning_rate": 2.1412844036697248e-05,
2602
+ "loss": 0.4707,
2603
+ "step": 362
2604
+ },
2605
+ {
2606
+ "epoch": 6.368421052631579,
2607
+ "grad_norm": 12.57919692993164,
2608
+ "learning_rate": 2.138532110091743e-05,
2609
+ "loss": 0.5245,
2610
+ "step": 363
2611
+ },
2612
+ {
2613
+ "epoch": 6.385964912280702,
2614
+ "grad_norm": 23.14158058166504,
2615
+ "learning_rate": 2.1357798165137616e-05,
2616
+ "loss": 0.4275,
2617
+ "step": 364
2618
+ },
2619
+ {
2620
+ "epoch": 6.4035087719298245,
2621
+ "grad_norm": 20.404285430908203,
2622
+ "learning_rate": 2.13302752293578e-05,
2623
+ "loss": 0.7325,
2624
+ "step": 365
2625
+ },
2626
+ {
2627
+ "epoch": 6.421052631578947,
2628
+ "grad_norm": 7.9764580726623535,
2629
+ "learning_rate": 2.130275229357798e-05,
2630
+ "loss": 0.469,
2631
+ "step": 366
2632
+ },
2633
+ {
2634
+ "epoch": 6.43859649122807,
2635
+ "grad_norm": 11.24958324432373,
2636
+ "learning_rate": 2.1275229357798167e-05,
2637
+ "loss": 0.7704,
2638
+ "step": 367
2639
+ },
2640
+ {
2641
+ "epoch": 6.456140350877193,
2642
+ "grad_norm": 3.0611751079559326,
2643
+ "learning_rate": 2.124770642201835e-05,
2644
+ "loss": 0.3323,
2645
+ "step": 368
2646
+ },
2647
+ {
2648
+ "epoch": 6.473684210526316,
2649
+ "grad_norm": 40.933162689208984,
2650
+ "learning_rate": 2.1220183486238532e-05,
2651
+ "loss": 0.5066,
2652
+ "step": 369
2653
+ },
2654
+ {
2655
+ "epoch": 6.491228070175438,
2656
+ "grad_norm": 9.126625061035156,
2657
+ "learning_rate": 2.1192660550458715e-05,
2658
+ "loss": 0.4907,
2659
+ "step": 370
2660
+ },
2661
+ {
2662
+ "epoch": 6.508771929824562,
2663
+ "grad_norm": 9.465023040771484,
2664
+ "learning_rate": 2.11651376146789e-05,
2665
+ "loss": 0.5455,
2666
+ "step": 371
2667
+ },
2668
+ {
2669
+ "epoch": 6.526315789473684,
2670
+ "grad_norm": 24.881996154785156,
2671
+ "learning_rate": 2.1137614678899083e-05,
2672
+ "loss": 0.6284,
2673
+ "step": 372
2674
+ },
2675
+ {
2676
+ "epoch": 6.543859649122807,
2677
+ "grad_norm": 32.151466369628906,
2678
+ "learning_rate": 2.1110091743119266e-05,
2679
+ "loss": 0.4083,
2680
+ "step": 373
2681
+ },
2682
+ {
2683
+ "epoch": 6.56140350877193,
2684
+ "grad_norm": 12.968864440917969,
2685
+ "learning_rate": 2.1082568807339448e-05,
2686
+ "loss": 0.5154,
2687
+ "step": 374
2688
+ },
2689
+ {
2690
+ "epoch": 6.578947368421053,
2691
+ "grad_norm": 64.264892578125,
2692
+ "learning_rate": 2.1055045871559634e-05,
2693
+ "loss": 0.2809,
2694
+ "step": 375
2695
+ },
2696
+ {
2697
+ "epoch": 6.5964912280701755,
2698
+ "grad_norm": 19.623029708862305,
2699
+ "learning_rate": 2.102752293577982e-05,
2700
+ "loss": 0.7864,
2701
+ "step": 376
2702
+ },
2703
+ {
2704
+ "epoch": 6.614035087719298,
2705
+ "grad_norm": 7.40114688873291,
2706
+ "learning_rate": 2.1e-05,
2707
+ "loss": 0.5327,
2708
+ "step": 377
2709
+ },
2710
+ {
2711
+ "epoch": 6.631578947368421,
2712
+ "grad_norm": 15.018710136413574,
2713
+ "learning_rate": 2.0972477064220185e-05,
2714
+ "loss": 0.5948,
2715
+ "step": 378
2716
+ },
2717
+ {
2718
+ "epoch": 6.649122807017544,
2719
+ "grad_norm": 6.5761847496032715,
2720
+ "learning_rate": 2.0944954128440367e-05,
2721
+ "loss": 0.5606,
2722
+ "step": 379
2723
+ },
2724
+ {
2725
+ "epoch": 6.666666666666667,
2726
+ "grad_norm": 21.176044464111328,
2727
+ "learning_rate": 2.0917431192660553e-05,
2728
+ "loss": 0.8495,
2729
+ "step": 380
2730
+ },
2731
+ {
2732
+ "epoch": 6.684210526315789,
2733
+ "grad_norm": 15.794927597045898,
2734
+ "learning_rate": 2.0889908256880732e-05,
2735
+ "loss": 0.2922,
2736
+ "step": 381
2737
+ },
2738
+ {
2739
+ "epoch": 6.701754385964913,
2740
+ "grad_norm": 21.976320266723633,
2741
+ "learning_rate": 2.0862385321100918e-05,
2742
+ "loss": 0.4193,
2743
+ "step": 382
2744
+ },
2745
+ {
2746
+ "epoch": 6.719298245614035,
2747
+ "grad_norm": 127.4170150756836,
2748
+ "learning_rate": 2.08348623853211e-05,
2749
+ "loss": 0.7539,
2750
+ "step": 383
2751
+ },
2752
+ {
2753
+ "epoch": 6.7368421052631575,
2754
+ "grad_norm": 23.519792556762695,
2755
+ "learning_rate": 2.0807339449541286e-05,
2756
+ "loss": 0.8324,
2757
+ "step": 384
2758
+ },
2759
+ {
2760
+ "epoch": 6.754385964912281,
2761
+ "grad_norm": 34.533729553222656,
2762
+ "learning_rate": 2.0779816513761465e-05,
2763
+ "loss": 0.635,
2764
+ "step": 385
2765
+ },
2766
+ {
2767
+ "epoch": 6.771929824561403,
2768
+ "grad_norm": 48.43922805786133,
2769
+ "learning_rate": 2.075229357798165e-05,
2770
+ "loss": 0.7622,
2771
+ "step": 386
2772
+ },
2773
+ {
2774
+ "epoch": 6.7894736842105265,
2775
+ "grad_norm": 9.6075439453125,
2776
+ "learning_rate": 2.0724770642201837e-05,
2777
+ "loss": 0.4885,
2778
+ "step": 387
2779
+ },
2780
+ {
2781
+ "epoch": 6.807017543859649,
2782
+ "grad_norm": 25.955398559570312,
2783
+ "learning_rate": 2.069724770642202e-05,
2784
+ "loss": 0.2685,
2785
+ "step": 388
2786
+ },
2787
+ {
2788
+ "epoch": 6.824561403508772,
2789
+ "grad_norm": 17.847700119018555,
2790
+ "learning_rate": 2.0669724770642202e-05,
2791
+ "loss": 0.3958,
2792
+ "step": 389
2793
+ },
2794
+ {
2795
+ "epoch": 6.842105263157895,
2796
+ "grad_norm": 13.096657752990723,
2797
+ "learning_rate": 2.0642201834862385e-05,
2798
+ "loss": 0.9942,
2799
+ "step": 390
2800
+ },
2801
+ {
2802
+ "epoch": 6.859649122807017,
2803
+ "grad_norm": 11.541616439819336,
2804
+ "learning_rate": 2.061467889908257e-05,
2805
+ "loss": 0.5704,
2806
+ "step": 391
2807
+ },
2808
+ {
2809
+ "epoch": 6.87719298245614,
2810
+ "grad_norm": 3.565450668334961,
2811
+ "learning_rate": 2.0587155963302753e-05,
2812
+ "loss": 0.6809,
2813
+ "step": 392
2814
+ },
2815
+ {
2816
+ "epoch": 6.894736842105263,
2817
+ "grad_norm": 27.333837509155273,
2818
+ "learning_rate": 2.0559633027522935e-05,
2819
+ "loss": 0.6584,
2820
+ "step": 393
2821
+ },
2822
+ {
2823
+ "epoch": 6.912280701754386,
2824
+ "grad_norm": 15.244169235229492,
2825
+ "learning_rate": 2.0532110091743118e-05,
2826
+ "loss": 0.6753,
2827
+ "step": 394
2828
+ },
2829
+ {
2830
+ "epoch": 6.9298245614035086,
2831
+ "grad_norm": 18.803529739379883,
2832
+ "learning_rate": 2.0504587155963304e-05,
2833
+ "loss": 0.5975,
2834
+ "step": 395
2835
+ },
2836
+ {
2837
+ "epoch": 6.947368421052632,
2838
+ "grad_norm": 7.317320346832275,
2839
+ "learning_rate": 2.047706422018349e-05,
2840
+ "loss": 0.9889,
2841
+ "step": 396
2842
+ },
2843
+ {
2844
+ "epoch": 6.964912280701754,
2845
+ "grad_norm": 19.616382598876953,
2846
+ "learning_rate": 2.044954128440367e-05,
2847
+ "loss": 0.6709,
2848
+ "step": 397
2849
+ },
2850
+ {
2851
+ "epoch": 6.982456140350877,
2852
+ "grad_norm": 46.652469635009766,
2853
+ "learning_rate": 2.0422018348623855e-05,
2854
+ "loss": 0.5268,
2855
+ "step": 398
2856
+ },
2857
+ {
2858
+ "epoch": 7.0,
2859
+ "grad_norm": 11.762057304382324,
2860
+ "learning_rate": 2.0394495412844037e-05,
2861
+ "loss": 0.9838,
2862
+ "step": 399
2863
+ },
2864
+ {
2865
+ "epoch": 7.0,
2866
+ "eval_accuracy": 0.6819923371647509,
2867
+ "eval_loss": 0.8196430802345276,
2868
+ "eval_roc_auc": 0.8813437942252584,
2869
+ "eval_runtime": 9.6385,
2870
+ "eval_samples_per_second": 27.079,
2871
+ "eval_steps_per_second": 1.764,
2872
+ "step": 399
2873
+ },
2874
+ {
2875
+ "epoch": 7.017543859649122,
2876
+ "grad_norm": 14.474690437316895,
2877
+ "learning_rate": 2.0366972477064223e-05,
2878
+ "loss": 0.3408,
2879
+ "step": 400
2880
+ },
2881
+ {
2882
+ "epoch": 7.035087719298246,
2883
+ "grad_norm": 19.78782081604004,
2884
+ "learning_rate": 2.0339449541284402e-05,
2885
+ "loss": 0.4244,
2886
+ "step": 401
2887
+ },
2888
+ {
2889
+ "epoch": 7.052631578947368,
2890
+ "grad_norm": 10.25345230102539,
2891
+ "learning_rate": 2.0311926605504588e-05,
2892
+ "loss": 0.6707,
2893
+ "step": 402
2894
+ },
2895
+ {
2896
+ "epoch": 7.0701754385964914,
2897
+ "grad_norm": 2.85701060295105,
2898
+ "learning_rate": 2.028440366972477e-05,
2899
+ "loss": 0.2397,
2900
+ "step": 403
2901
+ },
2902
+ {
2903
+ "epoch": 7.087719298245614,
2904
+ "grad_norm": 6.676547527313232,
2905
+ "learning_rate": 2.0256880733944956e-05,
2906
+ "loss": 0.3686,
2907
+ "step": 404
2908
+ },
2909
+ {
2910
+ "epoch": 7.105263157894737,
2911
+ "grad_norm": 165.5542755126953,
2912
+ "learning_rate": 2.0229357798165135e-05,
2913
+ "loss": 0.5689,
2914
+ "step": 405
2915
+ },
2916
+ {
2917
+ "epoch": 7.12280701754386,
2918
+ "grad_norm": 14.226078033447266,
2919
+ "learning_rate": 2.020183486238532e-05,
2920
+ "loss": 0.3423,
2921
+ "step": 406
2922
+ },
2923
+ {
2924
+ "epoch": 7.140350877192983,
2925
+ "grad_norm": 5.127176761627197,
2926
+ "learning_rate": 2.0174311926605507e-05,
2927
+ "loss": 0.6358,
2928
+ "step": 407
2929
+ },
2930
+ {
2931
+ "epoch": 7.157894736842105,
2932
+ "grad_norm": 21.368282318115234,
2933
+ "learning_rate": 2.014678899082569e-05,
2934
+ "loss": 0.5103,
2935
+ "step": 408
2936
+ },
2937
+ {
2938
+ "epoch": 7.175438596491228,
2939
+ "grad_norm": 3.635981321334839,
2940
+ "learning_rate": 2.0119266055045872e-05,
2941
+ "loss": 0.3426,
2942
+ "step": 409
2943
+ },
2944
+ {
2945
+ "epoch": 7.192982456140351,
2946
+ "grad_norm": 8.354199409484863,
2947
+ "learning_rate": 2.0091743119266054e-05,
2948
+ "loss": 0.4433,
2949
+ "step": 410
2950
+ },
2951
+ {
2952
+ "epoch": 7.2105263157894735,
2953
+ "grad_norm": 125.07445526123047,
2954
+ "learning_rate": 2.006422018348624e-05,
2955
+ "loss": 0.5754,
2956
+ "step": 411
2957
+ },
2958
+ {
2959
+ "epoch": 7.228070175438597,
2960
+ "grad_norm": 18.86847496032715,
2961
+ "learning_rate": 2.0036697247706423e-05,
2962
+ "loss": 0.6004,
2963
+ "step": 412
2964
+ },
2965
+ {
2966
+ "epoch": 7.245614035087719,
2967
+ "grad_norm": 13.290031433105469,
2968
+ "learning_rate": 2.0009174311926605e-05,
2969
+ "loss": 0.6924,
2970
+ "step": 413
2971
+ },
2972
+ {
2973
+ "epoch": 7.2631578947368425,
2974
+ "grad_norm": 7.521588325500488,
2975
+ "learning_rate": 1.9981651376146788e-05,
2976
+ "loss": 0.4925,
2977
+ "step": 414
2978
+ },
2979
+ {
2980
+ "epoch": 7.280701754385965,
2981
+ "grad_norm": 14.16098690032959,
2982
+ "learning_rate": 1.9954128440366974e-05,
2983
+ "loss": 0.6182,
2984
+ "step": 415
2985
+ },
2986
+ {
2987
+ "epoch": 7.298245614035087,
2988
+ "grad_norm": 12.972576141357422,
2989
+ "learning_rate": 1.9926605504587156e-05,
2990
+ "loss": 0.4994,
2991
+ "step": 416
2992
+ },
2993
+ {
2994
+ "epoch": 7.315789473684211,
2995
+ "grad_norm": 7.475390911102295,
2996
+ "learning_rate": 1.989908256880734e-05,
2997
+ "loss": 0.5975,
2998
+ "step": 417
2999
+ },
3000
+ {
3001
+ "epoch": 7.333333333333333,
3002
+ "grad_norm": 19.82378387451172,
3003
+ "learning_rate": 1.9871559633027524e-05,
3004
+ "loss": 0.4584,
3005
+ "step": 418
3006
+ },
3007
+ {
3008
+ "epoch": 7.350877192982456,
3009
+ "grad_norm": 7.333564281463623,
3010
+ "learning_rate": 1.9844036697247707e-05,
3011
+ "loss": 0.5651,
3012
+ "step": 419
3013
+ },
3014
+ {
3015
+ "epoch": 7.368421052631579,
3016
+ "grad_norm": 12.829773902893066,
3017
+ "learning_rate": 1.9816513761467893e-05,
3018
+ "loss": 0.4198,
3019
+ "step": 420
3020
+ },
3021
+ {
3022
+ "epoch": 7.385964912280702,
3023
+ "grad_norm": 3.876220464706421,
3024
+ "learning_rate": 1.9788990825688072e-05,
3025
+ "loss": 0.2843,
3026
+ "step": 421
3027
+ },
3028
+ {
3029
+ "epoch": 7.4035087719298245,
3030
+ "grad_norm": 10.361215591430664,
3031
+ "learning_rate": 1.9761467889908258e-05,
3032
+ "loss": 0.4935,
3033
+ "step": 422
3034
+ },
3035
+ {
3036
+ "epoch": 7.421052631578947,
3037
+ "grad_norm": 59.4860954284668,
3038
+ "learning_rate": 1.973394495412844e-05,
3039
+ "loss": 0.5763,
3040
+ "step": 423
3041
+ },
3042
+ {
3043
+ "epoch": 7.43859649122807,
3044
+ "grad_norm": 10.748738288879395,
3045
+ "learning_rate": 1.9706422018348626e-05,
3046
+ "loss": 0.7249,
3047
+ "step": 424
3048
+ },
3049
+ {
3050
+ "epoch": 7.456140350877193,
3051
+ "grad_norm": 19.900938034057617,
3052
+ "learning_rate": 1.9678899082568805e-05,
3053
+ "loss": 0.3448,
3054
+ "step": 425
3055
+ },
3056
+ {
3057
+ "epoch": 7.473684210526316,
3058
+ "grad_norm": 9.298110961914062,
3059
+ "learning_rate": 1.965137614678899e-05,
3060
+ "loss": 0.424,
3061
+ "step": 426
3062
+ },
3063
+ {
3064
+ "epoch": 7.491228070175438,
3065
+ "grad_norm": 6.225700855255127,
3066
+ "learning_rate": 1.9623853211009173e-05,
3067
+ "loss": 0.5622,
3068
+ "step": 427
3069
+ },
3070
+ {
3071
+ "epoch": 7.508771929824562,
3072
+ "grad_norm": 7.592733860015869,
3073
+ "learning_rate": 1.959633027522936e-05,
3074
+ "loss": 0.327,
3075
+ "step": 428
3076
+ },
3077
+ {
3078
+ "epoch": 7.526315789473684,
3079
+ "grad_norm": 15.926929473876953,
3080
+ "learning_rate": 1.9568807339449542e-05,
3081
+ "loss": 0.5397,
3082
+ "step": 429
3083
+ },
3084
+ {
3085
+ "epoch": 7.543859649122807,
3086
+ "grad_norm": 5.76393461227417,
3087
+ "learning_rate": 1.9541284403669724e-05,
3088
+ "loss": 0.6615,
3089
+ "step": 430
3090
+ },
3091
+ {
3092
+ "epoch": 7.56140350877193,
3093
+ "grad_norm": 9.405485153198242,
3094
+ "learning_rate": 1.951376146788991e-05,
3095
+ "loss": 0.5491,
3096
+ "step": 431
3097
+ },
3098
+ {
3099
+ "epoch": 7.578947368421053,
3100
+ "grad_norm": 6.3222479820251465,
3101
+ "learning_rate": 1.9486238532110093e-05,
3102
+ "loss": 0.5539,
3103
+ "step": 432
3104
+ },
3105
+ {
3106
+ "epoch": 7.5964912280701755,
3107
+ "grad_norm": 15.603589057922363,
3108
+ "learning_rate": 1.9458715596330275e-05,
3109
+ "loss": 0.352,
3110
+ "step": 433
3111
+ },
3112
+ {
3113
+ "epoch": 7.614035087719298,
3114
+ "grad_norm": 54.22720718383789,
3115
+ "learning_rate": 1.9431192660550458e-05,
3116
+ "loss": 0.5936,
3117
+ "step": 434
3118
+ },
3119
+ {
3120
+ "epoch": 7.631578947368421,
3121
+ "grad_norm": 7.901265621185303,
3122
+ "learning_rate": 1.9403669724770643e-05,
3123
+ "loss": 0.6319,
3124
+ "step": 435
3125
+ },
3126
+ {
3127
+ "epoch": 7.649122807017544,
3128
+ "grad_norm": 16.199270248413086,
3129
+ "learning_rate": 1.9376146788990826e-05,
3130
+ "loss": 0.5468,
3131
+ "step": 436
3132
+ },
3133
+ {
3134
+ "epoch": 7.666666666666667,
3135
+ "grad_norm": 84.76512908935547,
3136
+ "learning_rate": 1.934862385321101e-05,
3137
+ "loss": 0.5745,
3138
+ "step": 437
3139
+ },
3140
+ {
3141
+ "epoch": 7.684210526315789,
3142
+ "grad_norm": 8.624265670776367,
3143
+ "learning_rate": 1.9321100917431194e-05,
3144
+ "loss": 0.5141,
3145
+ "step": 438
3146
+ },
3147
+ {
3148
+ "epoch": 7.701754385964913,
3149
+ "grad_norm": 4.701018333435059,
3150
+ "learning_rate": 1.9293577981651377e-05,
3151
+ "loss": 0.2706,
3152
+ "step": 439
3153
+ },
3154
+ {
3155
+ "epoch": 7.719298245614035,
3156
+ "grad_norm": 3.979341745376587,
3157
+ "learning_rate": 1.9266055045871563e-05,
3158
+ "loss": 0.4677,
3159
+ "step": 440
3160
+ },
3161
+ {
3162
+ "epoch": 7.7368421052631575,
3163
+ "grad_norm": 10.237481117248535,
3164
+ "learning_rate": 1.923853211009174e-05,
3165
+ "loss": 0.6246,
3166
+ "step": 441
3167
+ },
3168
+ {
3169
+ "epoch": 7.754385964912281,
3170
+ "grad_norm": 16.742145538330078,
3171
+ "learning_rate": 1.9211009174311927e-05,
3172
+ "loss": 0.5573,
3173
+ "step": 442
3174
+ },
3175
+ {
3176
+ "epoch": 7.771929824561403,
3177
+ "grad_norm": 7.850118637084961,
3178
+ "learning_rate": 1.918348623853211e-05,
3179
+ "loss": 0.3705,
3180
+ "step": 443
3181
+ },
3182
+ {
3183
+ "epoch": 7.7894736842105265,
3184
+ "grad_norm": 7.423269271850586,
3185
+ "learning_rate": 1.9155963302752296e-05,
3186
+ "loss": 0.5847,
3187
+ "step": 444
3188
+ },
3189
+ {
3190
+ "epoch": 7.807017543859649,
3191
+ "grad_norm": 3.7411139011383057,
3192
+ "learning_rate": 1.9128440366972475e-05,
3193
+ "loss": 0.3861,
3194
+ "step": 445
3195
+ },
3196
+ {
3197
+ "epoch": 7.824561403508772,
3198
+ "grad_norm": 131.05946350097656,
3199
+ "learning_rate": 1.910091743119266e-05,
3200
+ "loss": 0.5242,
3201
+ "step": 446
3202
+ },
3203
+ {
3204
+ "epoch": 7.842105263157895,
3205
+ "grad_norm": 12.686647415161133,
3206
+ "learning_rate": 1.9073394495412843e-05,
3207
+ "loss": 0.2422,
3208
+ "step": 447
3209
+ },
3210
+ {
3211
+ "epoch": 7.859649122807017,
3212
+ "grad_norm": 7.073160171508789,
3213
+ "learning_rate": 1.904587155963303e-05,
3214
+ "loss": 0.2323,
3215
+ "step": 448
3216
+ },
3217
+ {
3218
+ "epoch": 7.87719298245614,
3219
+ "grad_norm": 4.771604061126709,
3220
+ "learning_rate": 1.901834862385321e-05,
3221
+ "loss": 0.2788,
3222
+ "step": 449
3223
+ },
3224
+ {
3225
+ "epoch": 7.894736842105263,
3226
+ "grad_norm": 11.123143196105957,
3227
+ "learning_rate": 1.8990825688073394e-05,
3228
+ "loss": 0.3865,
3229
+ "step": 450
3230
+ },
3231
+ {
3232
+ "epoch": 7.912280701754386,
3233
+ "grad_norm": 19.361055374145508,
3234
+ "learning_rate": 1.896330275229358e-05,
3235
+ "loss": 0.4511,
3236
+ "step": 451
3237
+ },
3238
+ {
3239
+ "epoch": 7.9298245614035086,
3240
+ "grad_norm": 8.789932250976562,
3241
+ "learning_rate": 1.8935779816513762e-05,
3242
+ "loss": 0.3365,
3243
+ "step": 452
3244
+ },
3245
+ {
3246
+ "epoch": 7.947368421052632,
3247
+ "grad_norm": 12.437773704528809,
3248
+ "learning_rate": 1.8908256880733945e-05,
3249
+ "loss": 0.3662,
3250
+ "step": 453
3251
+ },
3252
+ {
3253
+ "epoch": 7.964912280701754,
3254
+ "grad_norm": 17.096485137939453,
3255
+ "learning_rate": 1.8880733944954127e-05,
3256
+ "loss": 0.5647,
3257
+ "step": 454
3258
+ },
3259
+ {
3260
+ "epoch": 7.982456140350877,
3261
+ "grad_norm": 20.25951385498047,
3262
+ "learning_rate": 1.8853211009174313e-05,
3263
+ "loss": 0.4497,
3264
+ "step": 455
3265
+ },
3266
+ {
3267
+ "epoch": 8.0,
3268
+ "grad_norm": 14.128129005432129,
3269
+ "learning_rate": 1.8825688073394496e-05,
3270
+ "loss": 0.3699,
3271
+ "step": 456
3272
+ },
3273
+ {
3274
+ "epoch": 8.0,
3275
+ "eval_accuracy": 0.6743295019157088,
3276
+ "eval_loss": 0.9114387035369873,
3277
+ "eval_roc_auc": 0.8783069858814981,
3278
+ "eval_runtime": 9.7288,
3279
+ "eval_samples_per_second": 26.827,
3280
+ "eval_steps_per_second": 1.747,
3281
+ "step": 456
3282
+ },
3283
+ {
3284
+ "epoch": 8.017543859649123,
3285
+ "grad_norm": 14.421273231506348,
3286
+ "learning_rate": 1.8798165137614678e-05,
3287
+ "loss": 0.4197,
3288
+ "step": 457
3289
+ },
3290
+ {
3291
+ "epoch": 8.035087719298245,
3292
+ "grad_norm": 7.326334476470947,
3293
+ "learning_rate": 1.877064220183486e-05,
3294
+ "loss": 0.3367,
3295
+ "step": 458
3296
+ },
3297
+ {
3298
+ "epoch": 8.052631578947368,
3299
+ "grad_norm": 9.997421264648438,
3300
+ "learning_rate": 1.8743119266055047e-05,
3301
+ "loss": 0.4693,
3302
+ "step": 459
3303
+ },
3304
+ {
3305
+ "epoch": 8.070175438596491,
3306
+ "grad_norm": 16.052030563354492,
3307
+ "learning_rate": 1.8715596330275232e-05,
3308
+ "loss": 0.3885,
3309
+ "step": 460
3310
+ },
3311
+ {
3312
+ "epoch": 8.087719298245615,
3313
+ "grad_norm": 5.081548690795898,
3314
+ "learning_rate": 1.868807339449541e-05,
3315
+ "loss": 0.3361,
3316
+ "step": 461
3317
+ },
3318
+ {
3319
+ "epoch": 8.105263157894736,
3320
+ "grad_norm": 12.288408279418945,
3321
+ "learning_rate": 1.8660550458715597e-05,
3322
+ "loss": 0.4407,
3323
+ "step": 462
3324
+ },
3325
+ {
3326
+ "epoch": 8.12280701754386,
3327
+ "grad_norm": 20.765512466430664,
3328
+ "learning_rate": 1.863302752293578e-05,
3329
+ "loss": 0.3498,
3330
+ "step": 463
3331
+ },
3332
+ {
3333
+ "epoch": 8.140350877192983,
3334
+ "grad_norm": 16.51128578186035,
3335
+ "learning_rate": 1.8605504587155966e-05,
3336
+ "loss": 0.5519,
3337
+ "step": 464
3338
+ },
3339
+ {
3340
+ "epoch": 8.157894736842104,
3341
+ "grad_norm": 9.807088851928711,
3342
+ "learning_rate": 1.8577981651376145e-05,
3343
+ "loss": 0.4753,
3344
+ "step": 465
3345
+ },
3346
+ {
3347
+ "epoch": 8.175438596491228,
3348
+ "grad_norm": 27.620849609375,
3349
+ "learning_rate": 1.855045871559633e-05,
3350
+ "loss": 0.4292,
3351
+ "step": 466
3352
+ },
3353
+ {
3354
+ "epoch": 8.192982456140351,
3355
+ "grad_norm": 3.228712320327759,
3356
+ "learning_rate": 1.8522935779816513e-05,
3357
+ "loss": 0.2354,
3358
+ "step": 467
3359
+ },
3360
+ {
3361
+ "epoch": 8.210526315789474,
3362
+ "grad_norm": 19.046974182128906,
3363
+ "learning_rate": 1.84954128440367e-05,
3364
+ "loss": 0.7982,
3365
+ "step": 468
3366
+ },
3367
+ {
3368
+ "epoch": 8.228070175438596,
3369
+ "grad_norm": 35.34560775756836,
3370
+ "learning_rate": 1.8467889908256878e-05,
3371
+ "loss": 0.6045,
3372
+ "step": 469
3373
+ },
3374
+ {
3375
+ "epoch": 8.24561403508772,
3376
+ "grad_norm": 19.82710838317871,
3377
+ "learning_rate": 1.8440366972477064e-05,
3378
+ "loss": 0.2748,
3379
+ "step": 470
3380
+ },
3381
+ {
3382
+ "epoch": 8.263157894736842,
3383
+ "grad_norm": 107.26939392089844,
3384
+ "learning_rate": 1.841284403669725e-05,
3385
+ "loss": 0.5046,
3386
+ "step": 471
3387
+ },
3388
+ {
3389
+ "epoch": 8.280701754385966,
3390
+ "grad_norm": 9.757340431213379,
3391
+ "learning_rate": 1.8385321100917432e-05,
3392
+ "loss": 0.2989,
3393
+ "step": 472
3394
+ },
3395
+ {
3396
+ "epoch": 8.298245614035087,
3397
+ "grad_norm": 42.89109802246094,
3398
+ "learning_rate": 1.8357798165137615e-05,
3399
+ "loss": 0.4882,
3400
+ "step": 473
3401
+ },
3402
+ {
3403
+ "epoch": 8.31578947368421,
3404
+ "grad_norm": 27.338111877441406,
3405
+ "learning_rate": 1.8330275229357797e-05,
3406
+ "loss": 0.4523,
3407
+ "step": 474
3408
+ },
3409
+ {
3410
+ "epoch": 8.333333333333334,
3411
+ "grad_norm": 3.047417163848877,
3412
+ "learning_rate": 1.8302752293577983e-05,
3413
+ "loss": 0.2369,
3414
+ "step": 475
3415
+ },
3416
+ {
3417
+ "epoch": 8.350877192982455,
3418
+ "grad_norm": 27.468341827392578,
3419
+ "learning_rate": 1.8275229357798166e-05,
3420
+ "loss": 0.4376,
3421
+ "step": 476
3422
+ },
3423
+ {
3424
+ "epoch": 8.368421052631579,
3425
+ "grad_norm": 4.080908298492432,
3426
+ "learning_rate": 1.8247706422018348e-05,
3427
+ "loss": 0.3097,
3428
+ "step": 477
3429
+ },
3430
+ {
3431
+ "epoch": 8.385964912280702,
3432
+ "grad_norm": 15.838717460632324,
3433
+ "learning_rate": 1.822018348623853e-05,
3434
+ "loss": 0.424,
3435
+ "step": 478
3436
+ },
3437
+ {
3438
+ "epoch": 8.403508771929825,
3439
+ "grad_norm": 9.841987609863281,
3440
+ "learning_rate": 1.8192660550458716e-05,
3441
+ "loss": 0.5209,
3442
+ "step": 479
3443
+ },
3444
+ {
3445
+ "epoch": 8.421052631578947,
3446
+ "grad_norm": 11.688602447509766,
3447
+ "learning_rate": 1.8165137614678902e-05,
3448
+ "loss": 0.2871,
3449
+ "step": 480
3450
+ },
3451
+ {
3452
+ "epoch": 8.43859649122807,
3453
+ "grad_norm": 42.60127639770508,
3454
+ "learning_rate": 1.813761467889908e-05,
3455
+ "loss": 0.7017,
3456
+ "step": 481
3457
+ },
3458
+ {
3459
+ "epoch": 8.456140350877194,
3460
+ "grad_norm": 21.921663284301758,
3461
+ "learning_rate": 1.8110091743119267e-05,
3462
+ "loss": 0.6333,
3463
+ "step": 482
3464
+ },
3465
+ {
3466
+ "epoch": 8.473684210526315,
3467
+ "grad_norm": 14.230140686035156,
3468
+ "learning_rate": 1.808256880733945e-05,
3469
+ "loss": 0.4494,
3470
+ "step": 483
3471
+ },
3472
+ {
3473
+ "epoch": 8.491228070175438,
3474
+ "grad_norm": 4.411128997802734,
3475
+ "learning_rate": 1.8055045871559636e-05,
3476
+ "loss": 0.2,
3477
+ "step": 484
3478
+ },
3479
+ {
3480
+ "epoch": 8.508771929824562,
3481
+ "grad_norm": 4.632619857788086,
3482
+ "learning_rate": 1.8027522935779815e-05,
3483
+ "loss": 0.3362,
3484
+ "step": 485
3485
+ },
3486
+ {
3487
+ "epoch": 8.526315789473685,
3488
+ "grad_norm": 11.009140014648438,
3489
+ "learning_rate": 1.8e-05,
3490
+ "loss": 0.2433,
3491
+ "step": 486
3492
+ },
3493
+ {
3494
+ "epoch": 8.543859649122806,
3495
+ "grad_norm": 35.23042297363281,
3496
+ "learning_rate": 1.7972477064220183e-05,
3497
+ "loss": 0.4174,
3498
+ "step": 487
3499
+ },
3500
+ {
3501
+ "epoch": 8.56140350877193,
3502
+ "grad_norm": 6.151736259460449,
3503
+ "learning_rate": 1.794495412844037e-05,
3504
+ "loss": 0.3024,
3505
+ "step": 488
3506
+ },
3507
+ {
3508
+ "epoch": 8.578947368421053,
3509
+ "grad_norm": 25.69590950012207,
3510
+ "learning_rate": 1.7917431192660548e-05,
3511
+ "loss": 0.7038,
3512
+ "step": 489
3513
+ },
3514
+ {
3515
+ "epoch": 8.596491228070175,
3516
+ "grad_norm": 20.76173973083496,
3517
+ "learning_rate": 1.7889908256880734e-05,
3518
+ "loss": 0.6203,
3519
+ "step": 490
3520
+ },
3521
+ {
3522
+ "epoch": 8.614035087719298,
3523
+ "grad_norm": 17.6284122467041,
3524
+ "learning_rate": 1.786238532110092e-05,
3525
+ "loss": 0.3738,
3526
+ "step": 491
3527
+ },
3528
+ {
3529
+ "epoch": 8.631578947368421,
3530
+ "grad_norm": 6.477032661437988,
3531
+ "learning_rate": 1.7834862385321102e-05,
3532
+ "loss": 0.4933,
3533
+ "step": 492
3534
+ },
3535
+ {
3536
+ "epoch": 8.649122807017545,
3537
+ "grad_norm": 11.045804023742676,
3538
+ "learning_rate": 1.7807339449541285e-05,
3539
+ "loss": 0.4582,
3540
+ "step": 493
3541
+ },
3542
+ {
3543
+ "epoch": 8.666666666666666,
3544
+ "grad_norm": 7.45139741897583,
3545
+ "learning_rate": 1.7779816513761467e-05,
3546
+ "loss": 0.2789,
3547
+ "step": 494
3548
+ },
3549
+ {
3550
+ "epoch": 8.68421052631579,
3551
+ "grad_norm": 6.092661380767822,
3552
+ "learning_rate": 1.7752293577981653e-05,
3553
+ "loss": 0.5297,
3554
+ "step": 495
3555
+ },
3556
+ {
3557
+ "epoch": 8.701754385964913,
3558
+ "grad_norm": 9.822498321533203,
3559
+ "learning_rate": 1.7724770642201835e-05,
3560
+ "loss": 0.4994,
3561
+ "step": 496
3562
+ },
3563
+ {
3564
+ "epoch": 8.719298245614034,
3565
+ "grad_norm": 25.944725036621094,
3566
+ "learning_rate": 1.7697247706422018e-05,
3567
+ "loss": 0.5637,
3568
+ "step": 497
3569
+ },
3570
+ {
3571
+ "epoch": 8.736842105263158,
3572
+ "grad_norm": 7.313636302947998,
3573
+ "learning_rate": 1.76697247706422e-05,
3574
+ "loss": 0.3809,
3575
+ "step": 498
3576
+ },
3577
+ {
3578
+ "epoch": 8.75438596491228,
3579
+ "grad_norm": 28.630786895751953,
3580
+ "learning_rate": 1.7642201834862386e-05,
3581
+ "loss": 0.2553,
3582
+ "step": 499
3583
+ },
3584
+ {
3585
+ "epoch": 8.771929824561404,
3586
+ "grad_norm": 13.779499053955078,
3587
+ "learning_rate": 1.761467889908257e-05,
3588
+ "loss": 0.6088,
3589
+ "step": 500
3590
+ },
3591
+ {
3592
+ "epoch": 8.789473684210526,
3593
+ "grad_norm": 13.866922378540039,
3594
+ "learning_rate": 1.758715596330275e-05,
3595
+ "loss": 0.3418,
3596
+ "step": 501
3597
+ },
3598
+ {
3599
+ "epoch": 8.807017543859649,
3600
+ "grad_norm": 9.200101852416992,
3601
+ "learning_rate": 1.7559633027522937e-05,
3602
+ "loss": 0.5261,
3603
+ "step": 502
3604
+ },
3605
+ {
3606
+ "epoch": 8.824561403508772,
3607
+ "grad_norm": 4.6678853034973145,
3608
+ "learning_rate": 1.753211009174312e-05,
3609
+ "loss": 0.2657,
3610
+ "step": 503
3611
+ },
3612
+ {
3613
+ "epoch": 8.842105263157894,
3614
+ "grad_norm": 13.394104957580566,
3615
+ "learning_rate": 1.7504587155963305e-05,
3616
+ "loss": 0.4874,
3617
+ "step": 504
3618
+ },
3619
+ {
3620
+ "epoch": 8.859649122807017,
3621
+ "grad_norm": 8.163538932800293,
3622
+ "learning_rate": 1.7477064220183484e-05,
3623
+ "loss": 0.4425,
3624
+ "step": 505
3625
+ },
3626
+ {
3627
+ "epoch": 8.87719298245614,
3628
+ "grad_norm": 11.24685001373291,
3629
+ "learning_rate": 1.744954128440367e-05,
3630
+ "loss": 0.4143,
3631
+ "step": 506
3632
+ },
3633
+ {
3634
+ "epoch": 8.894736842105264,
3635
+ "grad_norm": 15.51749324798584,
3636
+ "learning_rate": 1.7422018348623853e-05,
3637
+ "loss": 0.2075,
3638
+ "step": 507
3639
+ },
3640
+ {
3641
+ "epoch": 8.912280701754385,
3642
+ "grad_norm": 18.03982162475586,
3643
+ "learning_rate": 1.739449541284404e-05,
3644
+ "loss": 0.3895,
3645
+ "step": 508
3646
+ },
3647
+ {
3648
+ "epoch": 8.929824561403509,
3649
+ "grad_norm": 30.238689422607422,
3650
+ "learning_rate": 1.7366972477064218e-05,
3651
+ "loss": 0.502,
3652
+ "step": 509
3653
+ },
3654
+ {
3655
+ "epoch": 8.947368421052632,
3656
+ "grad_norm": 17.355592727661133,
3657
+ "learning_rate": 1.7339449541284404e-05,
3658
+ "loss": 0.3032,
3659
+ "step": 510
3660
+ },
3661
+ {
3662
+ "epoch": 8.964912280701755,
3663
+ "grad_norm": 8.36916732788086,
3664
+ "learning_rate": 1.7311926605504586e-05,
3665
+ "loss": 0.4566,
3666
+ "step": 511
3667
+ },
3668
+ {
3669
+ "epoch": 8.982456140350877,
3670
+ "grad_norm": 4.937262535095215,
3671
+ "learning_rate": 1.7284403669724772e-05,
3672
+ "loss": 0.2615,
3673
+ "step": 512
3674
+ },
3675
+ {
3676
+ "epoch": 9.0,
3677
+ "grad_norm": 61.639869689941406,
3678
+ "learning_rate": 1.7256880733944954e-05,
3679
+ "loss": 0.6482,
3680
+ "step": 513
3681
+ },
3682
+ {
3683
+ "epoch": 9.0,
3684
+ "eval_accuracy": 0.7126436781609196,
3685
+ "eval_loss": 0.8051581382751465,
3686
+ "eval_roc_auc": 0.8928142130921077,
3687
+ "eval_runtime": 9.5241,
3688
+ "eval_samples_per_second": 27.404,
3689
+ "eval_steps_per_second": 1.785,
3690
+ "step": 513
3691
+ }
3692
+ ],
3693
+ "logging_steps": 1,
3694
+ "max_steps": 1140,
3695
+ "num_input_tokens_seen": 0,
3696
+ "num_train_epochs": 20,
3697
+ "save_steps": 500,
3698
+ "stateful_callbacks": {
3699
+ "TrainerControl": {
3700
+ "args": {
3701
+ "should_epoch_stop": false,
3702
+ "should_evaluate": false,
3703
+ "should_log": false,
3704
+ "should_save": true,
3705
+ "should_training_stop": false
3706
+ },
3707
+ "attributes": {}
3708
+ }
3709
+ },
3710
+ "total_flos": 2.23308947617632e+17,
3711
+ "train_batch_size": 16,
3712
+ "trial_name": null,
3713
+ "trial_params": null
3714
+ }
checkpoint-513/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e48f80f5f7846c1de45b12dd03848923118b7f8dcfa969a22a9b8fb06dc3c376
3
+ size 5112
checkpoint-969/config.json ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-base",
3
+ "activation_dropout": 0.0,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForSequenceClassification"
11
+ ],
12
+ "attention_dropout": 0.1,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 256,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": false,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "sum",
46
+ "ctc_zero_infinity": false,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": false,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_norm": "group",
52
+ "feat_proj_dropout": 0.1,
53
+ "feat_quantizer_dropout": 0.0,
54
+ "final_dropout": 0.0,
55
+ "freeze_feat_extract_train": true,
56
+ "hidden_act": "gelu",
57
+ "hidden_dropout": 0.1,
58
+ "hidden_size": 768,
59
+ "id2label": {
60
+ "0": "Lesion",
61
+ "1": "Normal",
62
+ "2": "Paralysis",
63
+ "3": "Atrophy",
64
+ "4": "MTD"
65
+ },
66
+ "initializer_range": 0.02,
67
+ "intermediate_size": 3072,
68
+ "label2id": {
69
+ "LABEL_0": 0,
70
+ "LABEL_1": 1,
71
+ "LABEL_2": 2,
72
+ "LABEL_3": 3,
73
+ "LABEL_4": 4
74
+ },
75
+ "layer_norm_eps": 1e-05,
76
+ "layerdrop": 0.0,
77
+ "mask_channel_length": 10,
78
+ "mask_channel_min_space": 1,
79
+ "mask_channel_other": 0.0,
80
+ "mask_channel_prob": 0.0,
81
+ "mask_channel_selection": "static",
82
+ "mask_feature_length": 10,
83
+ "mask_feature_min_masks": 0,
84
+ "mask_feature_prob": 0.0,
85
+ "mask_time_length": 10,
86
+ "mask_time_min_masks": 2,
87
+ "mask_time_min_space": 1,
88
+ "mask_time_other": 0.0,
89
+ "mask_time_prob": 0.05,
90
+ "mask_time_selection": "static",
91
+ "model_type": "wav2vec2",
92
+ "no_mask_channel_overlap": false,
93
+ "no_mask_time_overlap": false,
94
+ "num_adapter_layers": 3,
95
+ "num_attention_heads": 12,
96
+ "num_codevector_groups": 2,
97
+ "num_codevectors_per_group": 320,
98
+ "num_conv_pos_embedding_groups": 16,
99
+ "num_conv_pos_embeddings": 128,
100
+ "num_feat_extract_layers": 7,
101
+ "num_hidden_layers": 12,
102
+ "num_negatives": 100,
103
+ "output_hidden_size": 768,
104
+ "pad_token_id": 0,
105
+ "proj_codevector_dim": 256,
106
+ "tdnn_dilation": [
107
+ 1,
108
+ 2,
109
+ 3,
110
+ 1,
111
+ 1
112
+ ],
113
+ "tdnn_dim": [
114
+ 512,
115
+ 512,
116
+ 512,
117
+ 512,
118
+ 1500
119
+ ],
120
+ "tdnn_kernel": [
121
+ 5,
122
+ 3,
123
+ 3,
124
+ 1,
125
+ 1
126
+ ],
127
+ "torch_dtype": "float32",
128
+ "transformers_version": "4.42.0.dev0",
129
+ "use_weighted_layer_sum": false,
130
+ "vocab_size": 32,
131
+ "xvector_output_dim": 512
132
+ }
checkpoint-969/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ede01aa08c9694b47b2723b335c6a0322d96a35cb0d10a85f273d4f756da31ce
3
+ size 378305452
checkpoint-969/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:952db691bdb4f601366f5dabdf701a5b9b6ff4ca884558dd25854799a8b4c3cf
3
+ size 756738746
checkpoint-969/preprocessor_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "return_attention_mask": false,
8
+ "sampling_rate": 16000
9
+ }
checkpoint-969/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c394441342b3ba7451be20e313f197e3db830f6a92f409d6343fb9e24983c60b
3
+ size 14308
checkpoint-969/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fe5dff51f041c7df15c690efba8c78b414c7c56840a3deecdacb4b6517657f6e
3
+ size 1064
checkpoint-969/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-969/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e48f80f5f7846c1de45b12dd03848923118b7f8dcfa969a22a9b8fb06dc3c376
3
+ size 5112
config.json ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-base",
3
+ "activation_dropout": 0.0,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForSequenceClassification"
11
+ ],
12
+ "attention_dropout": 0.1,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 256,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": false,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "sum",
46
+ "ctc_zero_infinity": false,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": false,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_norm": "group",
52
+ "feat_proj_dropout": 0.1,
53
+ "feat_quantizer_dropout": 0.0,
54
+ "final_dropout": 0.0,
55
+ "freeze_feat_extract_train": true,
56
+ "hidden_act": "gelu",
57
+ "hidden_dropout": 0.1,
58
+ "hidden_size": 768,
59
+ "id2label": {
60
+ "0": "Lesion",
61
+ "1": "Normal",
62
+ "2": "Paralysis",
63
+ "3": "Atrophy",
64
+ "4": "MTD"
65
+ },
66
+ "initializer_range": 0.02,
67
+ "intermediate_size": 3072,
68
+ "label2id": {
69
+ "LABEL_0": 0,
70
+ "LABEL_1": 1,
71
+ "LABEL_2": 2,
72
+ "LABEL_3": 3,
73
+ "LABEL_4": 4
74
+ },
75
+ "layer_norm_eps": 1e-05,
76
+ "layerdrop": 0.0,
77
+ "mask_channel_length": 10,
78
+ "mask_channel_min_space": 1,
79
+ "mask_channel_other": 0.0,
80
+ "mask_channel_prob": 0.0,
81
+ "mask_channel_selection": "static",
82
+ "mask_feature_length": 10,
83
+ "mask_feature_min_masks": 0,
84
+ "mask_feature_prob": 0.0,
85
+ "mask_time_length": 10,
86
+ "mask_time_min_masks": 2,
87
+ "mask_time_min_space": 1,
88
+ "mask_time_other": 0.0,
89
+ "mask_time_prob": 0.05,
90
+ "mask_time_selection": "static",
91
+ "model_type": "wav2vec2",
92
+ "no_mask_channel_overlap": false,
93
+ "no_mask_time_overlap": false,
94
+ "num_adapter_layers": 3,
95
+ "num_attention_heads": 12,
96
+ "num_codevector_groups": 2,
97
+ "num_codevectors_per_group": 320,
98
+ "num_conv_pos_embedding_groups": 16,
99
+ "num_conv_pos_embeddings": 128,
100
+ "num_feat_extract_layers": 7,
101
+ "num_hidden_layers": 12,
102
+ "num_negatives": 100,
103
+ "output_hidden_size": 768,
104
+ "pad_token_id": 0,
105
+ "proj_codevector_dim": 256,
106
+ "tdnn_dilation": [
107
+ 1,
108
+ 2,
109
+ 3,
110
+ 1,
111
+ 1
112
+ ],
113
+ "tdnn_dim": [
114
+ 512,
115
+ 512,
116
+ 512,
117
+ 512,
118
+ 1500
119
+ ],
120
+ "tdnn_kernel": [
121
+ 5,
122
+ 3,
123
+ 3,
124
+ 1,
125
+ 1
126
+ ],
127
+ "torch_dtype": "float32",
128
+ "transformers_version": "4.42.0.dev0",
129
+ "use_weighted_layer_sum": false,
130
+ "vocab_size": 32,
131
+ "xvector_output_dim": 512
132
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a53b1994aba1a9bb0d194f45b0b18aeb8ac6277dede830adf4babb6c4235103
3
+ size 378305452
preprocessor_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "return_attention_mask": false,
8
+ "sampling_rate": 16000
9
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e48f80f5f7846c1de45b12dd03848923118b7f8dcfa969a22a9b8fb06dc3c376
3
+ size 5112