imvladikon commited on
Commit
e8f3ff1
1 Parent(s): 8c65cf6

End of training

Browse files
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "eval_loss": 0.3532930910587311,
4
+ "eval_runtime": 297.867,
5
+ "eval_samples": 5076,
6
+ "eval_samples_per_second": 17.041,
7
+ "eval_steps_per_second": 2.84,
8
+ "eval_wer": 0.22513066847239563,
9
+ "train_loss": 1.3125287554224614,
10
+ "train_runtime": 67690.7896,
11
+ "train_samples": 20306,
12
+ "train_samples_per_second": 6.0,
13
+ "train_steps_per_second": 0.25
14
+ }
checkpoint-4000/config.json DELETED
@@ -1,79 +0,0 @@
1
- {
2
- "_name_or_path": "facebook/wav2vec2-xls-r-300m",
3
- "activation_dropout": 0.0,
4
- "apply_spec_augment": true,
5
- "architectures": [
6
- "Wav2Vec2ForCTC"
7
- ],
8
- "attention_dropout": 0.0,
9
- "bos_token_id": 1,
10
- "classifier_proj_size": 256,
11
- "codevector_dim": 768,
12
- "contrastive_logits_temperature": 0.1,
13
- "conv_bias": true,
14
- "conv_dim": [
15
- 512,
16
- 512,
17
- 512,
18
- 512,
19
- 512,
20
- 512,
21
- 512
22
- ],
23
- "conv_kernel": [
24
- 10,
25
- 3,
26
- 3,
27
- 3,
28
- 3,
29
- 2,
30
- 2
31
- ],
32
- "conv_stride": [
33
- 5,
34
- 2,
35
- 2,
36
- 2,
37
- 2,
38
- 2,
39
- 2
40
- ],
41
- "ctc_loss_reduction": "mean",
42
- "ctc_zero_infinity": false,
43
- "diversity_loss_weight": 0.1,
44
- "do_stable_layer_norm": true,
45
- "eos_token_id": 2,
46
- "feat_extract_activation": "gelu",
47
- "feat_extract_dropout": 0.0,
48
- "feat_extract_norm": "layer",
49
- "feat_proj_dropout": 0.0,
50
- "feat_quantizer_dropout": 0.0,
51
- "final_dropout": 0.0,
52
- "gradient_checkpointing": false,
53
- "hidden_act": "gelu",
54
- "hidden_dropout": 0.0,
55
- "hidden_size": 1024,
56
- "initializer_range": 0.02,
57
- "intermediate_size": 4096,
58
- "layer_norm_eps": 1e-05,
59
- "layerdrop": 0.0,
60
- "mask_feature_length": 10,
61
- "mask_feature_prob": 0.0,
62
- "mask_time_length": 10,
63
- "mask_time_prob": 0.05,
64
- "model_type": "wav2vec2",
65
- "num_attention_heads": 16,
66
- "num_codevector_groups": 2,
67
- "num_codevectors_per_group": 320,
68
- "num_conv_pos_embedding_groups": 16,
69
- "num_conv_pos_embeddings": 128,
70
- "num_feat_extract_layers": 7,
71
- "num_hidden_layers": 24,
72
- "num_negatives": 100,
73
- "pad_token_id": 29,
74
- "proj_codevector_dim": 768,
75
- "torch_dtype": "float32",
76
- "transformers_version": "4.11.3",
77
- "use_weighted_layer_sum": false,
78
- "vocab_size": 32
79
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-4000/optimizer.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:052d9d1f443d1b9f67c2f7a6c2808fe922f1eb7ec26dba77c36c332395ec93d6
3
- size 2490321361
 
 
 
 
checkpoint-4000/preprocessor_config.json DELETED
@@ -1,9 +0,0 @@
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": true,
8
- "sampling_rate": 16000
9
- }
 
 
 
 
 
 
 
 
 
 
checkpoint-4000/pytorch_model.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:092b750aa2825af7fb89c41d0967aa326857a67ed0c101103b080ef3668b8ca5
3
- size 1262054897
 
 
 
 
checkpoint-4000/rng_state.pth DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:1bf4dfa62752c868edada56217703db94c074fedf843e434f438dfbdfd0009a0
3
- size 14631
 
 
 
 
checkpoint-4000/scaler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:93d4c53e2e4d886d0eac00fb8b6e3a12d9b4941d302b762706e463005cd1e4ad
3
- size 559
 
 
 
 
checkpoint-4000/scheduler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:8a20530e63990facdbe1c575c37ff4cc0d4637f57452c1d8f3b74093a7afbf98
3
- size 623
 
 
 
 
checkpoint-4000/trainer_state.json DELETED
@@ -1,166 +0,0 @@
1
- {
2
- "best_metric": null,
3
- "best_model_checkpoint": null,
4
- "epoch": 1.5759306677171558,
5
- "global_step": 4000,
6
- "is_hyper_param_search": false,
7
- "is_local_process_zero": true,
8
- "is_world_process_zero": true,
9
- "log_history": [
10
- {
11
- "epoch": 0.16,
12
- "learning_rate": 0.0002388,
13
- "loss": 4.5334,
14
- "step": 400
15
- },
16
- {
17
- "epoch": 0.16,
18
- "eval_loss": 2.026416063308716,
19
- "eval_runtime": 471.8471,
20
- "eval_samples_per_second": 10.758,
21
- "eval_steps_per_second": 1.346,
22
- "eval_wer": 0.9800369528744665,
23
- "step": 400
24
- },
25
- {
26
- "epoch": 0.32,
27
- "learning_rate": 0.00029881808566895816,
28
- "loss": 1.0808,
29
- "step": 800
30
- },
31
- {
32
- "epoch": 0.32,
33
- "eval_loss": 0.7069017887115479,
34
- "eval_runtime": 472.368,
35
- "eval_samples_per_second": 10.746,
36
- "eval_steps_per_second": 1.344,
37
- "eval_wer": 0.593412194448574,
38
- "step": 800
39
- },
40
- {
41
- "epoch": 0.47,
42
- "learning_rate": 0.0002972316234796404,
43
- "loss": 0.7499,
44
- "step": 1200
45
- },
46
- {
47
- "epoch": 0.47,
48
- "eval_loss": 0.5792935490608215,
49
- "eval_runtime": 471.2762,
50
- "eval_samples_per_second": 10.771,
51
- "eval_steps_per_second": 1.347,
52
- "eval_wer": 0.49153694225582434,
53
- "step": 1200
54
- },
55
- {
56
- "epoch": 0.63,
57
- "learning_rate": 0.00029564516129032255,
58
- "loss": 0.6397,
59
- "step": 1600
60
- },
61
- {
62
- "epoch": 0.63,
63
- "eval_loss": 0.5181192755699158,
64
- "eval_runtime": 470.1504,
65
- "eval_samples_per_second": 10.797,
66
- "eval_steps_per_second": 1.351,
67
- "eval_wer": 0.41618705799902306,
68
- "step": 1600
69
- },
70
- {
71
- "epoch": 0.79,
72
- "learning_rate": 0.0002940586991010047,
73
- "loss": 0.597,
74
- "step": 2000
75
- },
76
- {
77
- "epoch": 0.79,
78
- "eval_loss": 0.47279468178749084,
79
- "eval_runtime": 470.9283,
80
- "eval_samples_per_second": 10.779,
81
- "eval_steps_per_second": 1.348,
82
- "eval_wer": 0.37248072716461017,
83
- "step": 2000
84
- },
85
- {
86
- "epoch": 0.95,
87
- "learning_rate": 0.00029247223691168694,
88
- "loss": 0.5666,
89
- "step": 2400
90
- },
91
- {
92
- "epoch": 0.95,
93
- "eval_loss": 0.4448830783367157,
94
- "eval_runtime": 471.0174,
95
- "eval_samples_per_second": 10.777,
96
- "eval_steps_per_second": 1.348,
97
- "eval_wer": 0.3520292225030263,
98
- "step": 2400
99
- },
100
- {
101
- "epoch": 1.1,
102
- "learning_rate": 0.0002908857747223691,
103
- "loss": 0.5067,
104
- "step": 2800
105
- },
106
- {
107
- "epoch": 1.1,
108
- "eval_loss": 0.421601802110672,
109
- "eval_runtime": 470.0998,
110
- "eval_samples_per_second": 10.798,
111
- "eval_steps_per_second": 1.351,
112
- "eval_wer": 0.3330855650179455,
113
- "step": 2800
114
- },
115
- {
116
- "epoch": 1.26,
117
- "learning_rate": 0.0002892993125330513,
118
- "loss": 0.477,
119
- "step": 3200
120
- },
121
- {
122
- "epoch": 1.26,
123
- "eval_loss": 0.39891988039016724,
124
- "eval_runtime": 471.7321,
125
- "eval_samples_per_second": 10.76,
126
- "eval_steps_per_second": 1.346,
127
- "eval_wer": 0.30331089260305394,
128
- "step": 3200
129
- },
130
- {
131
- "epoch": 1.42,
132
- "learning_rate": 0.00028771285034373344,
133
- "loss": 0.4633,
134
- "step": 3600
135
- },
136
- {
137
- "epoch": 1.42,
138
- "eval_loss": 0.41667959094047546,
139
- "eval_runtime": 473.0088,
140
- "eval_samples_per_second": 10.731,
141
- "eval_steps_per_second": 1.342,
142
- "eval_wer": 0.3046276042219721,
143
- "step": 3600
144
- },
145
- {
146
- "epoch": 1.58,
147
- "learning_rate": 0.0002861263881544156,
148
- "loss": 0.4606,
149
- "step": 4000
150
- },
151
- {
152
- "epoch": 1.58,
153
- "eval_loss": 0.3892616033554077,
154
- "eval_runtime": 473.5469,
155
- "eval_samples_per_second": 10.719,
156
- "eval_steps_per_second": 1.341,
157
- "eval_wer": 0.29747063945462654,
158
- "step": 4000
159
- }
160
- ],
161
- "max_steps": 76140,
162
- "num_train_epochs": 30,
163
- "total_flos": 5.025867257304392e+18,
164
- "trial_name": null,
165
- "trial_params": null
166
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-4000/training_args.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:c4c1c368c4db40de7b0ac3ee4ff79af2d311395b8ce794d1869af91457b7e050
3
- size 2863
 
 
 
 
checkpoint-4400/config.json DELETED
@@ -1,79 +0,0 @@
1
- {
2
- "_name_or_path": "facebook/wav2vec2-xls-r-300m",
3
- "activation_dropout": 0.0,
4
- "apply_spec_augment": true,
5
- "architectures": [
6
- "Wav2Vec2ForCTC"
7
- ],
8
- "attention_dropout": 0.0,
9
- "bos_token_id": 1,
10
- "classifier_proj_size": 256,
11
- "codevector_dim": 768,
12
- "contrastive_logits_temperature": 0.1,
13
- "conv_bias": true,
14
- "conv_dim": [
15
- 512,
16
- 512,
17
- 512,
18
- 512,
19
- 512,
20
- 512,
21
- 512
22
- ],
23
- "conv_kernel": [
24
- 10,
25
- 3,
26
- 3,
27
- 3,
28
- 3,
29
- 2,
30
- 2
31
- ],
32
- "conv_stride": [
33
- 5,
34
- 2,
35
- 2,
36
- 2,
37
- 2,
38
- 2,
39
- 2
40
- ],
41
- "ctc_loss_reduction": "mean",
42
- "ctc_zero_infinity": false,
43
- "diversity_loss_weight": 0.1,
44
- "do_stable_layer_norm": true,
45
- "eos_token_id": 2,
46
- "feat_extract_activation": "gelu",
47
- "feat_extract_dropout": 0.0,
48
- "feat_extract_norm": "layer",
49
- "feat_proj_dropout": 0.0,
50
- "feat_quantizer_dropout": 0.0,
51
- "final_dropout": 0.0,
52
- "gradient_checkpointing": false,
53
- "hidden_act": "gelu",
54
- "hidden_dropout": 0.0,
55
- "hidden_size": 1024,
56
- "initializer_range": 0.02,
57
- "intermediate_size": 4096,
58
- "layer_norm_eps": 1e-05,
59
- "layerdrop": 0.0,
60
- "mask_feature_length": 10,
61
- "mask_feature_prob": 0.0,
62
- "mask_time_length": 10,
63
- "mask_time_prob": 0.05,
64
- "model_type": "wav2vec2",
65
- "num_attention_heads": 16,
66
- "num_codevector_groups": 2,
67
- "num_codevectors_per_group": 320,
68
- "num_conv_pos_embedding_groups": 16,
69
- "num_conv_pos_embeddings": 128,
70
- "num_feat_extract_layers": 7,
71
- "num_hidden_layers": 24,
72
- "num_negatives": 100,
73
- "pad_token_id": 29,
74
- "proj_codevector_dim": 768,
75
- "torch_dtype": "float32",
76
- "transformers_version": "4.11.3",
77
- "use_weighted_layer_sum": false,
78
- "vocab_size": 32
79
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-4400/optimizer.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:10430c88cb27d80d4909d60b1f5b4649806024e17f499620af7b33e594972686
3
- size 2490321361
 
 
 
 
checkpoint-4400/preprocessor_config.json DELETED
@@ -1,9 +0,0 @@
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": true,
8
- "sampling_rate": 16000
9
- }
 
 
 
 
 
 
 
 
 
 
checkpoint-4400/pytorch_model.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:9f6b19c2693a520d464d36c22f9e97cab05ee230f04af2c84cb79a18b55c8190
3
- size 1262054897
 
 
 
 
checkpoint-4400/rng_state.pth DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:f850e9eef43b3d8d7daa339e72e185c9b03f2fbdfe48df592252aedbe1875965
3
- size 14567
 
 
 
 
checkpoint-4400/scaler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:4affccbe68b9032483a9a189417afa898f74547fed0e8eff9cfb5b00deb58885
3
- size 559
 
 
 
 
checkpoint-4400/scheduler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:e11bf57746f56037dc02c0f1b9ef426c313cafb919b3099b8b7f85426ae9e2d5
3
- size 623
 
 
 
 
checkpoint-4400/trainer_state.json DELETED
@@ -1,181 +0,0 @@
1
- {
2
- "best_metric": null,
3
- "best_model_checkpoint": null,
4
- "epoch": 1.733504037817609,
5
- "global_step": 4400,
6
- "is_hyper_param_search": false,
7
- "is_local_process_zero": true,
8
- "is_world_process_zero": true,
9
- "log_history": [
10
- {
11
- "epoch": 0.16,
12
- "learning_rate": 0.0002388,
13
- "loss": 4.5334,
14
- "step": 400
15
- },
16
- {
17
- "epoch": 0.16,
18
- "eval_loss": 2.026416063308716,
19
- "eval_runtime": 471.8471,
20
- "eval_samples_per_second": 10.758,
21
- "eval_steps_per_second": 1.346,
22
- "eval_wer": 0.9800369528744665,
23
- "step": 400
24
- },
25
- {
26
- "epoch": 0.32,
27
- "learning_rate": 0.00029881808566895816,
28
- "loss": 1.0808,
29
- "step": 800
30
- },
31
- {
32
- "epoch": 0.32,
33
- "eval_loss": 0.7069017887115479,
34
- "eval_runtime": 472.368,
35
- "eval_samples_per_second": 10.746,
36
- "eval_steps_per_second": 1.344,
37
- "eval_wer": 0.593412194448574,
38
- "step": 800
39
- },
40
- {
41
- "epoch": 0.47,
42
- "learning_rate": 0.0002972316234796404,
43
- "loss": 0.7499,
44
- "step": 1200
45
- },
46
- {
47
- "epoch": 0.47,
48
- "eval_loss": 0.5792935490608215,
49
- "eval_runtime": 471.2762,
50
- "eval_samples_per_second": 10.771,
51
- "eval_steps_per_second": 1.347,
52
- "eval_wer": 0.49153694225582434,
53
- "step": 1200
54
- },
55
- {
56
- "epoch": 0.63,
57
- "learning_rate": 0.00029564516129032255,
58
- "loss": 0.6397,
59
- "step": 1600
60
- },
61
- {
62
- "epoch": 0.63,
63
- "eval_loss": 0.5181192755699158,
64
- "eval_runtime": 470.1504,
65
- "eval_samples_per_second": 10.797,
66
- "eval_steps_per_second": 1.351,
67
- "eval_wer": 0.41618705799902306,
68
- "step": 1600
69
- },
70
- {
71
- "epoch": 0.79,
72
- "learning_rate": 0.0002940586991010047,
73
- "loss": 0.597,
74
- "step": 2000
75
- },
76
- {
77
- "epoch": 0.79,
78
- "eval_loss": 0.47279468178749084,
79
- "eval_runtime": 470.9283,
80
- "eval_samples_per_second": 10.779,
81
- "eval_steps_per_second": 1.348,
82
- "eval_wer": 0.37248072716461017,
83
- "step": 2000
84
- },
85
- {
86
- "epoch": 0.95,
87
- "learning_rate": 0.00029247223691168694,
88
- "loss": 0.5666,
89
- "step": 2400
90
- },
91
- {
92
- "epoch": 0.95,
93
- "eval_loss": 0.4448830783367157,
94
- "eval_runtime": 471.0174,
95
- "eval_samples_per_second": 10.777,
96
- "eval_steps_per_second": 1.348,
97
- "eval_wer": 0.3520292225030263,
98
- "step": 2400
99
- },
100
- {
101
- "epoch": 1.1,
102
- "learning_rate": 0.0002908857747223691,
103
- "loss": 0.5067,
104
- "step": 2800
105
- },
106
- {
107
- "epoch": 1.1,
108
- "eval_loss": 0.421601802110672,
109
- "eval_runtime": 470.0998,
110
- "eval_samples_per_second": 10.798,
111
- "eval_steps_per_second": 1.351,
112
- "eval_wer": 0.3330855650179455,
113
- "step": 2800
114
- },
115
- {
116
- "epoch": 1.26,
117
- "learning_rate": 0.0002892993125330513,
118
- "loss": 0.477,
119
- "step": 3200
120
- },
121
- {
122
- "epoch": 1.26,
123
- "eval_loss": 0.39891988039016724,
124
- "eval_runtime": 471.7321,
125
- "eval_samples_per_second": 10.76,
126
- "eval_steps_per_second": 1.346,
127
- "eval_wer": 0.30331089260305394,
128
- "step": 3200
129
- },
130
- {
131
- "epoch": 1.42,
132
- "learning_rate": 0.00028771285034373344,
133
- "loss": 0.4633,
134
- "step": 3600
135
- },
136
- {
137
- "epoch": 1.42,
138
- "eval_loss": 0.41667959094047546,
139
- "eval_runtime": 473.0088,
140
- "eval_samples_per_second": 10.731,
141
- "eval_steps_per_second": 1.342,
142
- "eval_wer": 0.3046276042219721,
143
- "step": 3600
144
- },
145
- {
146
- "epoch": 1.58,
147
- "learning_rate": 0.0002861263881544156,
148
- "loss": 0.4606,
149
- "step": 4000
150
- },
151
- {
152
- "epoch": 1.58,
153
- "eval_loss": 0.3892616033554077,
154
- "eval_runtime": 473.5469,
155
- "eval_samples_per_second": 10.719,
156
- "eval_steps_per_second": 1.341,
157
- "eval_wer": 0.29747063945462654,
158
- "step": 4000
159
- },
160
- {
161
- "epoch": 1.73,
162
- "learning_rate": 0.0002845399259650978,
163
- "loss": 0.4385,
164
- "step": 4400
165
- },
166
- {
167
- "epoch": 1.73,
168
- "eval_loss": 0.3867528438568115,
169
- "eval_runtime": 471.6975,
170
- "eval_samples_per_second": 10.761,
171
- "eval_steps_per_second": 1.346,
172
- "eval_wer": 0.3090237220464247,
173
- "step": 4400
174
- }
175
- ],
176
- "max_steps": 76140,
177
- "num_train_epochs": 30,
178
- "total_flos": 5.530258185352643e+18,
179
- "trial_name": null,
180
- "trial_params": null
181
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-4400/training_args.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:c4c1c368c4db40de7b0ac3ee4ff79af2d311395b8ce794d1869af91457b7e050
3
- size 2863
 
 
 
 
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:64b32822466e070769a1da09a87aa88cdff1e2a2734026074c4f1c075fb5c0ce
3
  size 3850876337
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:76746edfd1645c596b48e8d49de9250fa4f78eb6ae4945a10266c6e82226bfbc
3
  size 3850876337
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "train_loss": 1.3125287554224614,
4
+ "train_runtime": 67690.7896,
5
+ "train_samples": 20306,
6
+ "train_samples_per_second": 6.0,
7
+ "train_steps_per_second": 0.25
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,655 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 19.99970457902511,
5
+ "global_step": 16920,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.47,
12
+ "learning_rate": 0.00029775,
13
+ "loss": 3.3587,
14
+ "step": 400
15
+ },
16
+ {
17
+ "epoch": 0.47,
18
+ "eval_loss": 1.1883399486541748,
19
+ "eval_runtime": 298.8414,
20
+ "eval_samples_per_second": 16.986,
21
+ "eval_steps_per_second": 2.831,
22
+ "eval_wer": 0.8392195865162833,
23
+ "step": 400
24
+ },
25
+ {
26
+ "epoch": 0.95,
27
+ "learning_rate": 0.00029279055690072635,
28
+ "loss": 1.8377,
29
+ "step": 800
30
+ },
31
+ {
32
+ "epoch": 0.95,
33
+ "eval_loss": 0.8830727338790894,
34
+ "eval_runtime": 295.1554,
35
+ "eval_samples_per_second": 17.198,
36
+ "eval_steps_per_second": 2.866,
37
+ "eval_wer": 0.6852318175085172,
38
+ "step": 800
39
+ },
40
+ {
41
+ "epoch": 1.42,
42
+ "learning_rate": 0.00028552663438256653,
43
+ "loss": 1.7118,
44
+ "step": 1200
45
+ },
46
+ {
47
+ "epoch": 1.42,
48
+ "eval_loss": 0.8031275868415833,
49
+ "eval_runtime": 296.1538,
50
+ "eval_samples_per_second": 17.14,
51
+ "eval_steps_per_second": 2.857,
52
+ "eval_wer": 0.6565588166832427,
53
+ "step": 1200
54
+ },
55
+ {
56
+ "epoch": 1.89,
57
+ "learning_rate": 0.00027826271186440676,
58
+ "loss": 1.6741,
59
+ "step": 1600
60
+ },
61
+ {
62
+ "epoch": 1.89,
63
+ "eval_loss": 0.7517648339271545,
64
+ "eval_runtime": 296.1537,
65
+ "eval_samples_per_second": 17.14,
66
+ "eval_steps_per_second": 2.857,
67
+ "eval_wer": 0.6104492456144063,
68
+ "step": 1600
69
+ },
70
+ {
71
+ "epoch": 2.36,
72
+ "learning_rate": 0.00027099878934624694,
73
+ "loss": 1.6163,
74
+ "step": 2000
75
+ },
76
+ {
77
+ "epoch": 2.36,
78
+ "eval_loss": 0.6887586712837219,
79
+ "eval_runtime": 298.4301,
80
+ "eval_samples_per_second": 17.009,
81
+ "eval_steps_per_second": 2.835,
82
+ "eval_wer": 0.5590706138773092,
83
+ "step": 2000
84
+ },
85
+ {
86
+ "epoch": 2.84,
87
+ "learning_rate": 0.0002637348668280871,
88
+ "loss": 1.5782,
89
+ "step": 2400
90
+ },
91
+ {
92
+ "epoch": 2.84,
93
+ "eval_loss": 0.6580386161804199,
94
+ "eval_runtime": 296.628,
95
+ "eval_samples_per_second": 17.112,
96
+ "eval_steps_per_second": 2.852,
97
+ "eval_wer": 0.5164525890344288,
98
+ "step": 2400
99
+ },
100
+ {
101
+ "epoch": 3.31,
102
+ "learning_rate": 0.0002564709443099273,
103
+ "loss": 1.5548,
104
+ "step": 2800
105
+ },
106
+ {
107
+ "epoch": 3.31,
108
+ "eval_loss": 0.6505530476570129,
109
+ "eval_runtime": 300.5251,
110
+ "eval_samples_per_second": 16.89,
111
+ "eval_steps_per_second": 2.815,
112
+ "eval_wer": 0.5184417123389128,
113
+ "step": 2800
114
+ },
115
+ {
116
+ "epoch": 3.78,
117
+ "learning_rate": 0.00024920702179176754,
118
+ "loss": 1.5249,
119
+ "step": 3200
120
+ },
121
+ {
122
+ "epoch": 3.78,
123
+ "eval_loss": 0.6197642683982849,
124
+ "eval_runtime": 298.2224,
125
+ "eval_samples_per_second": 17.021,
126
+ "eval_steps_per_second": 2.837,
127
+ "eval_wer": 0.5028461391963095,
128
+ "step": 3200
129
+ },
130
+ {
131
+ "epoch": 4.26,
132
+ "learning_rate": 0.00024194309927360772,
133
+ "loss": 1.5078,
134
+ "step": 3600
135
+ },
136
+ {
137
+ "epoch": 4.26,
138
+ "eval_loss": 0.5992246866226196,
139
+ "eval_runtime": 298.0744,
140
+ "eval_samples_per_second": 17.029,
141
+ "eval_steps_per_second": 2.838,
142
+ "eval_wer": 0.4932179359671583,
143
+ "step": 3600
144
+ },
145
+ {
146
+ "epoch": 4.73,
147
+ "learning_rate": 0.0002346791767554479,
148
+ "loss": 1.4836,
149
+ "step": 4000
150
+ },
151
+ {
152
+ "epoch": 4.73,
153
+ "eval_loss": 0.5705241560935974,
154
+ "eval_runtime": 292.5306,
155
+ "eval_samples_per_second": 17.352,
156
+ "eval_steps_per_second": 2.892,
157
+ "eval_wer": 0.46513743995598533,
158
+ "step": 4000
159
+ },
160
+ {
161
+ "epoch": 5.2,
162
+ "learning_rate": 0.0002274152542372881,
163
+ "loss": 1.4505,
164
+ "step": 4400
165
+ },
166
+ {
167
+ "epoch": 5.2,
168
+ "eval_loss": 0.5488837957382202,
169
+ "eval_runtime": 300.7764,
170
+ "eval_samples_per_second": 16.876,
171
+ "eval_steps_per_second": 2.813,
172
+ "eval_wer": 0.4507691982140212,
173
+ "step": 4400
174
+ },
175
+ {
176
+ "epoch": 5.67,
177
+ "learning_rate": 0.00022015133171912832,
178
+ "loss": 1.4481,
179
+ "step": 4800
180
+ },
181
+ {
182
+ "epoch": 5.67,
183
+ "eval_loss": 0.5577110052108765,
184
+ "eval_runtime": 299.1357,
185
+ "eval_samples_per_second": 16.969,
186
+ "eval_steps_per_second": 2.828,
187
+ "eval_wer": 0.45624986774446114,
188
+ "step": 4800
189
+ },
190
+ {
191
+ "epoch": 6.15,
192
+ "learning_rate": 0.0002128874092009685,
193
+ "loss": 1.4136,
194
+ "step": 5200
195
+ },
196
+ {
197
+ "epoch": 6.15,
198
+ "eval_loss": 0.5452219247817993,
199
+ "eval_runtime": 298.6905,
200
+ "eval_samples_per_second": 16.994,
201
+ "eval_steps_per_second": 2.832,
202
+ "eval_wer": 0.4370992657172482,
203
+ "step": 5200
204
+ },
205
+ {
206
+ "epoch": 6.62,
207
+ "learning_rate": 0.00020564164648910412,
208
+ "loss": 1.3861,
209
+ "step": 5600
210
+ },
211
+ {
212
+ "epoch": 6.62,
213
+ "eval_loss": 0.510087788105011,
214
+ "eval_runtime": 298.7447,
215
+ "eval_samples_per_second": 16.991,
216
+ "eval_steps_per_second": 2.832,
217
+ "eval_wer": 0.40865903464037073,
218
+ "step": 5600
219
+ },
220
+ {
221
+ "epoch": 7.09,
222
+ "learning_rate": 0.0001983777239709443,
223
+ "loss": 1.3772,
224
+ "step": 6000
225
+ },
226
+ {
227
+ "epoch": 7.09,
228
+ "eval_loss": 0.49333110451698303,
229
+ "eval_runtime": 298.2699,
230
+ "eval_samples_per_second": 17.018,
231
+ "eval_steps_per_second": 2.836,
232
+ "eval_wer": 0.39513722834712317,
233
+ "step": 6000
234
+ },
235
+ {
236
+ "epoch": 7.56,
237
+ "learning_rate": 0.00019111380145278448,
238
+ "loss": 1.3478,
239
+ "step": 6400
240
+ },
241
+ {
242
+ "epoch": 7.56,
243
+ "eval_loss": 0.4849308133125305,
244
+ "eval_runtime": 299.279,
245
+ "eval_samples_per_second": 16.961,
246
+ "eval_steps_per_second": 2.827,
247
+ "eval_wer": 0.3921747042766151,
248
+ "step": 6400
249
+ },
250
+ {
251
+ "epoch": 8.04,
252
+ "learning_rate": 0.0001838498789346247,
253
+ "loss": 1.3394,
254
+ "step": 6800
255
+ },
256
+ {
257
+ "epoch": 8.04,
258
+ "eval_loss": 0.4805210828781128,
259
+ "eval_runtime": 297.1556,
260
+ "eval_samples_per_second": 17.082,
261
+ "eval_steps_per_second": 2.847,
262
+ "eval_wer": 0.3891698584336712,
263
+ "step": 6800
264
+ },
265
+ {
266
+ "epoch": 8.51,
267
+ "learning_rate": 0.00017658595641646487,
268
+ "loss": 1.3095,
269
+ "step": 7200
270
+ },
271
+ {
272
+ "epoch": 8.51,
273
+ "eval_loss": 0.48388615250587463,
274
+ "eval_runtime": 298.6068,
275
+ "eval_samples_per_second": 16.999,
276
+ "eval_steps_per_second": 2.833,
277
+ "eval_wer": 0.38335061472374465,
278
+ "step": 7200
279
+ },
280
+ {
281
+ "epoch": 8.98,
282
+ "learning_rate": 0.00016932203389830508,
283
+ "loss": 1.306,
284
+ "step": 7600
285
+ },
286
+ {
287
+ "epoch": 8.98,
288
+ "eval_loss": 0.461063027381897,
289
+ "eval_runtime": 296.7161,
290
+ "eval_samples_per_second": 17.107,
291
+ "eval_steps_per_second": 2.851,
292
+ "eval_wer": 0.35865586050743803,
293
+ "step": 7600
294
+ },
295
+ {
296
+ "epoch": 9.46,
297
+ "learning_rate": 0.00016205811138014526,
298
+ "loss": 1.2707,
299
+ "step": 8000
300
+ },
301
+ {
302
+ "epoch": 9.46,
303
+ "eval_loss": 0.4544948935508728,
304
+ "eval_runtime": 298.2639,
305
+ "eval_samples_per_second": 17.018,
306
+ "eval_steps_per_second": 2.836,
307
+ "eval_wer": 0.3730241022494022,
308
+ "step": 8000
309
+ },
310
+ {
311
+ "epoch": 9.93,
312
+ "learning_rate": 0.00015479418886198547,
313
+ "loss": 1.2626,
314
+ "step": 8400
315
+ },
316
+ {
317
+ "epoch": 9.93,
318
+ "eval_loss": 0.4515869617462158,
319
+ "eval_runtime": 299.4234,
320
+ "eval_samples_per_second": 16.953,
321
+ "eval_steps_per_second": 2.825,
322
+ "eval_wer": 0.35241339907315317,
323
+ "step": 8400
324
+ },
325
+ {
326
+ "epoch": 10.4,
327
+ "learning_rate": 0.00014753026634382565,
328
+ "loss": 1.2412,
329
+ "step": 8800
330
+ },
331
+ {
332
+ "epoch": 10.4,
333
+ "eval_loss": 0.4314204454421997,
334
+ "eval_runtime": 294.8808,
335
+ "eval_samples_per_second": 17.214,
336
+ "eval_steps_per_second": 2.869,
337
+ "eval_wer": 0.3310197431068413,
338
+ "step": 8800
339
+ },
340
+ {
341
+ "epoch": 10.87,
342
+ "learning_rate": 0.00014026634382566583,
343
+ "loss": 1.2456,
344
+ "step": 9200
345
+ },
346
+ {
347
+ "epoch": 10.87,
348
+ "eval_loss": 0.4400792419910431,
349
+ "eval_runtime": 296.6883,
350
+ "eval_samples_per_second": 17.109,
351
+ "eval_steps_per_second": 2.851,
352
+ "eval_wer": 0.3458958461180354,
353
+ "step": 9200
354
+ },
355
+ {
356
+ "epoch": 11.35,
357
+ "learning_rate": 0.00013302058111380144,
358
+ "loss": 1.2081,
359
+ "step": 9600
360
+ },
361
+ {
362
+ "epoch": 11.35,
363
+ "eval_loss": 0.4399240016937256,
364
+ "eval_runtime": 295.1566,
365
+ "eval_samples_per_second": 17.198,
366
+ "eval_steps_per_second": 2.866,
367
+ "eval_wer": 0.33563281630234676,
368
+ "step": 9600
369
+ },
370
+ {
371
+ "epoch": 11.82,
372
+ "learning_rate": 0.00012575665859564165,
373
+ "loss": 1.1998,
374
+ "step": 10000
375
+ },
376
+ {
377
+ "epoch": 11.82,
378
+ "eval_loss": 0.41947221755981445,
379
+ "eval_runtime": 296.8576,
380
+ "eval_samples_per_second": 17.099,
381
+ "eval_steps_per_second": 2.85,
382
+ "eval_wer": 0.3215396660812155,
383
+ "step": 10000
384
+ },
385
+ {
386
+ "epoch": 12.29,
387
+ "learning_rate": 0.00011851089588377723,
388
+ "loss": 1.1826,
389
+ "step": 10400
390
+ },
391
+ {
392
+ "epoch": 12.29,
393
+ "eval_loss": 0.4220864474773407,
394
+ "eval_runtime": 296.3734,
395
+ "eval_samples_per_second": 17.127,
396
+ "eval_steps_per_second": 2.855,
397
+ "eval_wer": 0.3177518674482087,
398
+ "step": 10400
399
+ },
400
+ {
401
+ "epoch": 12.77,
402
+ "learning_rate": 0.00011124697336561742,
403
+ "loss": 1.1573,
404
+ "step": 10800
405
+ },
406
+ {
407
+ "epoch": 12.77,
408
+ "eval_loss": 0.4097737967967987,
409
+ "eval_runtime": 297.6541,
410
+ "eval_samples_per_second": 17.053,
411
+ "eval_steps_per_second": 2.842,
412
+ "eval_wer": 0.3084199166261083,
413
+ "step": 10800
414
+ },
415
+ {
416
+ "epoch": 13.24,
417
+ "learning_rate": 0.00010398305084745762,
418
+ "loss": 1.1416,
419
+ "step": 11200
420
+ },
421
+ {
422
+ "epoch": 13.24,
423
+ "eval_loss": 0.4085627496242523,
424
+ "eval_runtime": 296.7019,
425
+ "eval_samples_per_second": 17.108,
426
+ "eval_steps_per_second": 2.851,
427
+ "eval_wer": 0.3119114628520642,
428
+ "step": 11200
429
+ },
430
+ {
431
+ "epoch": 13.71,
432
+ "learning_rate": 9.671912832929781e-05,
433
+ "loss": 1.1174,
434
+ "step": 11600
435
+ },
436
+ {
437
+ "epoch": 13.71,
438
+ "eval_loss": 0.3854358196258545,
439
+ "eval_runtime": 297.2728,
440
+ "eval_samples_per_second": 17.075,
441
+ "eval_steps_per_second": 2.846,
442
+ "eval_wer": 0.2910468290412003,
443
+ "step": 11600
444
+ },
445
+ {
446
+ "epoch": 14.18,
447
+ "learning_rate": 8.947336561743341e-05,
448
+ "loss": 1.1048,
449
+ "step": 12000
450
+ },
451
+ {
452
+ "epoch": 14.18,
453
+ "eval_loss": 0.38590070605278015,
454
+ "eval_runtime": 297.414,
455
+ "eval_samples_per_second": 17.067,
456
+ "eval_steps_per_second": 2.845,
457
+ "eval_wer": 0.282413187464291,
458
+ "step": 12000
459
+ },
460
+ {
461
+ "epoch": 14.66,
462
+ "learning_rate": 8.22094430992736e-05,
463
+ "loss": 1.0748,
464
+ "step": 12400
465
+ },
466
+ {
467
+ "epoch": 14.66,
468
+ "eval_loss": 0.3854171633720398,
469
+ "eval_runtime": 295.6597,
470
+ "eval_samples_per_second": 17.168,
471
+ "eval_steps_per_second": 2.861,
472
+ "eval_wer": 0.275705186533212,
473
+ "step": 12400
474
+ },
475
+ {
476
+ "epoch": 15.13,
477
+ "learning_rate": 7.494552058111379e-05,
478
+ "loss": 1.0697,
479
+ "step": 12800
480
+ },
481
+ {
482
+ "epoch": 15.13,
483
+ "eval_loss": 0.37401217222213745,
484
+ "eval_runtime": 297.4379,
485
+ "eval_samples_per_second": 17.066,
486
+ "eval_steps_per_second": 2.844,
487
+ "eval_wer": 0.2723829273969994,
488
+ "step": 12800
489
+ },
490
+ {
491
+ "epoch": 15.6,
492
+ "learning_rate": 6.7681598062954e-05,
493
+ "loss": 1.0477,
494
+ "step": 13200
495
+ },
496
+ {
497
+ "epoch": 15.6,
498
+ "eval_loss": 0.3693487048149109,
499
+ "eval_runtime": 298.4761,
500
+ "eval_samples_per_second": 17.006,
501
+ "eval_steps_per_second": 2.834,
502
+ "eval_wer": 0.2642994688617559,
503
+ "step": 13200
504
+ },
505
+ {
506
+ "epoch": 16.08,
507
+ "learning_rate": 6.0417675544794184e-05,
508
+ "loss": 1.0356,
509
+ "step": 13600
510
+ },
511
+ {
512
+ "epoch": 16.08,
513
+ "eval_loss": 0.3726678788661957,
514
+ "eval_runtime": 298.583,
515
+ "eval_samples_per_second": 17.0,
516
+ "eval_steps_per_second": 2.833,
517
+ "eval_wer": 0.2561102058954229,
518
+ "step": 13600
519
+ },
520
+ {
521
+ "epoch": 16.55,
522
+ "learning_rate": 5.315375302663438e-05,
523
+ "loss": 1.0083,
524
+ "step": 14000
525
+ },
526
+ {
527
+ "epoch": 16.55,
528
+ "eval_loss": 0.36522331833839417,
529
+ "eval_runtime": 301.7885,
530
+ "eval_samples_per_second": 16.82,
531
+ "eval_steps_per_second": 2.803,
532
+ "eval_wer": 0.25007935332331716,
533
+ "step": 14000
534
+ },
535
+ {
536
+ "epoch": 17.02,
537
+ "learning_rate": 4.5889830508474574e-05,
538
+ "loss": 1.0,
539
+ "step": 14400
540
+ },
541
+ {
542
+ "epoch": 17.02,
543
+ "eval_loss": 0.36411818861961365,
544
+ "eval_runtime": 297.5091,
545
+ "eval_samples_per_second": 17.062,
546
+ "eval_steps_per_second": 2.844,
547
+ "eval_wer": 0.24572021076242673,
548
+ "step": 14400
549
+ },
550
+ {
551
+ "epoch": 17.49,
552
+ "learning_rate": 3.862590799031477e-05,
553
+ "loss": 0.9779,
554
+ "step": 14800
555
+ },
556
+ {
557
+ "epoch": 17.49,
558
+ "eval_loss": 0.35678642988204956,
559
+ "eval_runtime": 295.5669,
560
+ "eval_samples_per_second": 17.174,
561
+ "eval_steps_per_second": 2.862,
562
+ "eval_wer": 0.24089552870474215,
563
+ "step": 14800
564
+ },
565
+ {
566
+ "epoch": 17.97,
567
+ "learning_rate": 3.1361985472154963e-05,
568
+ "loss": 0.9596,
569
+ "step": 15200
570
+ },
571
+ {
572
+ "epoch": 17.97,
573
+ "eval_loss": 0.3557918071746826,
574
+ "eval_runtime": 296.6965,
575
+ "eval_samples_per_second": 17.108,
576
+ "eval_steps_per_second": 2.851,
577
+ "eval_wer": 0.23759443045474746,
578
+ "step": 15200
579
+ },
580
+ {
581
+ "epoch": 18.44,
582
+ "learning_rate": 2.4098062953995155e-05,
583
+ "loss": 0.946,
584
+ "step": 15600
585
+ },
586
+ {
587
+ "epoch": 18.44,
588
+ "eval_loss": 0.35907429456710815,
589
+ "eval_runtime": 295.7242,
590
+ "eval_samples_per_second": 17.165,
591
+ "eval_steps_per_second": 2.861,
592
+ "eval_wer": 0.23114036015828343,
593
+ "step": 15600
594
+ },
595
+ {
596
+ "epoch": 18.91,
597
+ "learning_rate": 1.685230024213075e-05,
598
+ "loss": 0.9389,
599
+ "step": 16000
600
+ },
601
+ {
602
+ "epoch": 18.91,
603
+ "eval_loss": 0.3540380001068115,
604
+ "eval_runtime": 296.3149,
605
+ "eval_samples_per_second": 17.13,
606
+ "eval_steps_per_second": 2.855,
607
+ "eval_wer": 0.22828364051886493,
608
+ "step": 16000
609
+ },
610
+ {
611
+ "epoch": 19.39,
612
+ "learning_rate": 9.588377723970942e-06,
613
+ "loss": 0.9173,
614
+ "step": 16400
615
+ },
616
+ {
617
+ "epoch": 19.39,
618
+ "eval_loss": 0.35515815019607544,
619
+ "eval_runtime": 296.2267,
620
+ "eval_samples_per_second": 17.136,
621
+ "eval_steps_per_second": 2.856,
622
+ "eval_wer": 0.2265061260765601,
623
+ "step": 16400
624
+ },
625
+ {
626
+ "epoch": 19.86,
627
+ "learning_rate": 2.3244552058111378e-06,
628
+ "loss": 0.9122,
629
+ "step": 16800
630
+ },
631
+ {
632
+ "epoch": 19.86,
633
+ "eval_loss": 0.353456050157547,
634
+ "eval_runtime": 298.3571,
635
+ "eval_samples_per_second": 17.013,
636
+ "eval_steps_per_second": 2.836,
637
+ "eval_wer": 0.22502486404130606,
638
+ "step": 16800
639
+ },
640
+ {
641
+ "epoch": 20.0,
642
+ "step": 16920,
643
+ "total_flos": 1.9284410553255697e+20,
644
+ "train_loss": 1.3125287554224614,
645
+ "train_runtime": 67690.7896,
646
+ "train_samples_per_second": 6.0,
647
+ "train_steps_per_second": 0.25
648
+ }
649
+ ],
650
+ "max_steps": 16920,
651
+ "num_train_epochs": 20,
652
+ "total_flos": 1.9284410553255697e+20,
653
+ "trial_name": null,
654
+ "trial_params": null
655
+ }
validation_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "eval_loss": 0.3532930910587311,
4
+ "eval_runtime": 297.867,
5
+ "eval_samples": 5076,
6
+ "eval_samples_per_second": 17.041,
7
+ "eval_steps_per_second": 2.84,
8
+ "eval_wer": 0.22513066847239563
9
+ }