Hugging Face's logo

yesj1234
/
enko_xlsr_13p_run1

Automatic Speech Recognition
Transformers
PyTorch
wav2vec2
./sample_speech.py
Generated from Trainer
Inference Endpoints
Model card Files Community
yesj1234 commited on
Commit
99d6d7b
•
1 Parent(s): 34cced9

Upload folder using huggingface_hub

Browse files
Files changed (28) hide show
  1. README.md +77 -0
  2. added_tokens.json +4 -0
  3. all_results.json +14 -0
  4. checkpoint-77020/added_tokens.json +4 -0
  5. checkpoint-77020/config.json +117 -0
  6. checkpoint-77020/optimizer.pt +3 -0
  7. checkpoint-77020/preprocessor_config.json +10 -0
  8. checkpoint-77020/pytorch_model.bin +3 -0
  9. checkpoint-77020/rng_state_0.pth +3 -0
  10. checkpoint-77020/rng_state_1.pth +3 -0
  11. checkpoint-77020/rng_state_2.pth +3 -0
  12. checkpoint-77020/rng_state_3.pth +3 -0
  13. checkpoint-77020/scheduler.pt +3 -0
  14. checkpoint-77020/special_tokens_map.json +30 -0
  15. checkpoint-77020/tokenizer_config.json +51 -0
  16. checkpoint-77020/trainer_state.json +1033 -0
  17. checkpoint-77020/training_args.bin +3 -0
  18. checkpoint-77020/vocab.json +103 -0
  19. config.json +117 -0
  20. eval_results.json +9 -0
  21. preprocessor_config.json +10 -0
  22. pytorch_model.bin +3 -0
  23. special_tokens_map.json +30 -0
  24. tokenizer_config.json +51 -0
  25. train_results.json +8 -0
  26. trainer_state.json +1042 -0
  27. training_args.bin +3 -0
  28. vocab.json +103 -0
README.md ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: facebook/wav2vec2-large-xlsr-53
4
+ tags:
5
+ - automatic-speech-recognition
6
+ - ./sample_speech.py
7
+ - generated_from_trainer
8
+ metrics:
9
+ - wer
10
+ model-index:
11
+ - name: enko_xlsr_13p_run1
12
+ results: []
13
+ ---
14
+
15
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
16
+ should probably proofread and complete it, then remove this comment. -->
17
+
18
+ # enko_xlsr_13p_run1
19
+
20
+ This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the ./SAMPLE_SPEECH.PY - NA dataset.
21
+ It achieves the following results on the evaluation set:
22
+ - Loss: 0.3042
23
+ - Wer: 0.1696
24
+
25
+ ## Model description
26
+
27
+ More information needed
28
+
29
+ ## Intended uses & limitations
30
+
31
+ More information needed
32
+
33
+ ## Training and evaluation data
34
+
35
+ More information needed
36
+
37
+ ## Training procedure
38
+
39
+ ### Training hyperparameters
40
+
41
+ The following hyperparameters were used during training:
42
+ - learning_rate: 0.0003
43
+ - train_batch_size: 2
44
+ - eval_batch_size: 2
45
+ - seed: 42
46
+ - distributed_type: multi-GPU
47
+ - num_devices: 4
48
+ - gradient_accumulation_steps: 2
49
+ - total_train_batch_size: 16
50
+ - total_eval_batch_size: 8
51
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
52
+ - lr_scheduler_type: linear
53
+ - lr_scheduler_warmup_steps: 500
54
+ - num_epochs: 10
55
+
56
+ ### Training results
57
+
58
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
59
+ |:-------------:|:-----:|:-----:|:---------------:|:------:|
60
+ | 0.6595 | 1.0 | 7702 | 0.4495 | 0.2974 |
61
+ | 0.5717 | 2.0 | 15404 | 0.3982 | 0.2562 |
62
+ | 0.5134 | 3.0 | 23106 | 0.3769 | 0.2365 |
63
+ | 0.467 | 4.0 | 30808 | 0.3499 | 0.2203 |
64
+ | 0.4156 | 5.0 | 38510 | 0.3391 | 0.2116 |
65
+ | 0.379 | 6.0 | 46212 | 0.3327 | 0.1999 |
66
+ | 0.3475 | 7.0 | 53914 | 0.3127 | 0.1947 |
67
+ | 0.3105 | 8.0 | 61616 | 0.3081 | 0.1814 |
68
+ | 0.281 | 9.0 | 69318 | 0.3068 | 0.1742 |
69
+ | 0.2584 | 10.0 | 77020 | 0.3040 | 0.1713 |
70
+
71
+
72
+ ### Framework versions
73
+
74
+ - Transformers 4.34.1
75
+ - Pytorch 2.1.0+cu121
76
+ - Datasets 2.14.6
77
+ - Tokenizers 0.14.1
added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "</s>": 102,
3
+ "<s>": 101
4
+ }
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 10.0,
3
+ "eval_loss": 0.3041677474975586,
4
+ "eval_runtime": 10.735,
5
+ "eval_samples": 711,
6
+ "eval_samples_per_second": 66.232,
7
+ "eval_steps_per_second": 8.291,
8
+ "eval_wer": 0.1695852534562212,
9
+ "train_loss": 0.4655882824058503,
10
+ "train_runtime": 41869.3879,
11
+ "train_samples": 123225,
12
+ "train_samples_per_second": 29.431,
13
+ "train_steps_per_second": 1.84
14
+ }
checkpoint-77020/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "</s>": 102,
3
+ "<s>": 101
4
+ }
checkpoint-77020/config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
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
+ "Wav2Vec2ForCTC"
11
+ ],
12
+ "attention_dropout": 0.1,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 768,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": true,
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": "mean",
46
+ "ctc_zero_infinity": false,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": true,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_dropout": 0.0,
52
+ "feat_extract_norm": "layer",
53
+ "feat_proj_dropout": 0.05,
54
+ "feat_quantizer_dropout": 0.0,
55
+ "final_dropout": 0.0,
56
+ "gradient_checkpointing": false,
57
+ "hidden_act": "gelu",
58
+ "hidden_dropout": 0.05,
59
+ "hidden_size": 1024,
60
+ "initializer_range": 0.02,
61
+ "intermediate_size": 4096,
62
+ "layer_norm_eps": 1e-05,
63
+ "layerdrop": 0.05,
64
+ "mask_channel_length": 10,
65
+ "mask_channel_min_space": 1,
66
+ "mask_channel_other": 0.0,
67
+ "mask_channel_prob": 0.0,
68
+ "mask_channel_selection": "static",
69
+ "mask_feature_length": 10,
70
+ "mask_feature_min_masks": 0,
71
+ "mask_feature_prob": 0.0,
72
+ "mask_time_length": 10,
73
+ "mask_time_min_masks": 2,
74
+ "mask_time_min_space": 1,
75
+ "mask_time_other": 0.0,
76
+ "mask_time_prob": 0.05,
77
+ "mask_time_selection": "static",
78
+ "model_type": "wav2vec2",
79
+ "num_adapter_layers": 3,
80
+ "num_attention_heads": 16,
81
+ "num_codevector_groups": 2,
82
+ "num_codevectors_per_group": 320,
83
+ "num_conv_pos_embedding_groups": 16,
84
+ "num_conv_pos_embeddings": 128,
85
+ "num_feat_extract_layers": 7,
86
+ "num_hidden_layers": 24,
87
+ "num_negatives": 100,
88
+ "output_hidden_size": 1024,
89
+ "pad_token_id": 100,
90
+ "proj_codevector_dim": 768,
91
+ "tdnn_dilation": [
92
+ 1,
93
+ 2,
94
+ 3,
95
+ 1,
96
+ 1
97
+ ],
98
+ "tdnn_dim": [
99
+ 512,
100
+ 512,
101
+ 512,
102
+ 512,
103
+ 1500
104
+ ],
105
+ "tdnn_kernel": [
106
+ 5,
107
+ 3,
108
+ 3,
109
+ 1,
110
+ 1
111
+ ],
112
+ "torch_dtype": "float32",
113
+ "transformers_version": "4.34.1",
114
+ "use_weighted_layer_sum": false,
115
+ "vocab_size": 103,
116
+ "xvector_output_dim": 512
117
+ }
checkpoint-77020/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a43e4f833bac02c6432ff56ca75cd03a8a45a069c0f73721aa095d7bf972a08
3
+ size 2491004342
checkpoint-77020/preprocessor_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0,
7
+ "processor_class": "Wav2Vec2Processor",
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ }
checkpoint-77020/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c9bdf0fd4f9034a342a7bdb6cfbf49f649ce26a10a80c8194598687bd004a2f2
3
+ size 1262324650
checkpoint-77020/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:82267a863b733fa877bbbbcef5ccf30aa4dfec7d449e076de289df51c051c399
3
+ size 15088
checkpoint-77020/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c3d04192ec36ec0d0d205a747cb4c96a57958816f0750a887d973846bf9da8e
3
+ size 15024
checkpoint-77020/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:72be2f34d25492c00ecbf4c26bc66c57c68f53993421523680001a0153ab3acd
3
+ size 15088
checkpoint-77020/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5c45cdb1267d6c8a96008d5eb0a97defd3db0b2d1cd3d335c34f0ca0c220cc1
3
+ size 15024
checkpoint-77020/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f18c026bc0e81eb3c7415165490fe9eb4a8bd63cf970bf8ba23e476fda1205a4
3
+ size 1064
checkpoint-77020/special_tokens_map.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "[PAD]",
18
+ "lstrip": true,
19
+ "normalized": false,
20
+ "rstrip": true,
21
+ "single_word": false
22
+ },
23
+ "unk_token": {
24
+ "content": "[UNK]",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": true,
28
+ "single_word": false
29
+ }
30
+ }
checkpoint-77020/tokenizer_config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "99": {
4
+ "content": "[UNK]",
5
+ "lstrip": true,
6
+ "normalized": false,
7
+ "rstrip": true,
8
+ "single_word": false,
9
+ "special": false
10
+ },
11
+ "100": {
12
+ "content": "[PAD]",
13
+ "lstrip": true,
14
+ "normalized": false,
15
+ "rstrip": true,
16
+ "single_word": false,
17
+ "special": false
18
+ },
19
+ "101": {
20
+ "content": "<s>",
21
+ "lstrip": false,
22
+ "normalized": true,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "102": {
28
+ "content": "</s>",
29
+ "lstrip": false,
30
+ "normalized": true,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ }
35
+ },
36
+ "bos_token": "<s>",
37
+ "clean_up_tokenization_spaces": true,
38
+ "config": null,
39
+ "do_lower_case": false,
40
+ "eos_token": "</s>",
41
+ "model_max_length": 1000000000000000019884624838656,
42
+ "pad_token": "[PAD]",
43
+ "processor_class": "Wav2Vec2Processor",
44
+ "replace_word_delimiter_char": " ",
45
+ "target_lang": null,
46
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
47
+ "tokenizer_type": "wav2vec2",
48
+ "trust_remote_code": false,
49
+ "unk_token": "[UNK]",
50
+ "word_delimiter_token": "|"
51
+ }
checkpoint-77020/trainer_state.json ADDED
@@ -0,0 +1,1033 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.3040466010570526,
3
+ "best_model_checkpoint": "./enko_xlsr_13p_run1/checkpoint-77020",
4
+ "epoch": 10.0,
5
+ "eval_steps": 500,
6
+ "global_step": 77020,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.06,
13
+ "learning_rate": 0.0003,
14
+ "loss": 4.7797,
15
+ "step": 500
16
+ },
17
+ {
18
+ "epoch": 0.13,
19
+ "learning_rate": 0.00029803972817564033,
20
+ "loss": 1.2377,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.19,
25
+ "learning_rate": 0.0002960794563512807,
26
+ "loss": 0.9452,
27
+ "step": 1500
28
+ },
29
+ {
30
+ "epoch": 0.26,
31
+ "learning_rate": 0.000294119184526921,
32
+ "loss": 0.8622,
33
+ "step": 2000
34
+ },
35
+ {
36
+ "epoch": 0.32,
37
+ "learning_rate": 0.0002921589127025614,
38
+ "loss": 0.82,
39
+ "step": 2500
40
+ },
41
+ {
42
+ "epoch": 0.39,
43
+ "learning_rate": 0.0002901986408782018,
44
+ "loss": 0.7742,
45
+ "step": 3000
46
+ },
47
+ {
48
+ "epoch": 0.45,
49
+ "learning_rate": 0.00028823836905384213,
50
+ "loss": 0.75,
51
+ "step": 3500
52
+ },
53
+ {
54
+ "epoch": 0.52,
55
+ "learning_rate": 0.0002862780972294825,
56
+ "loss": 0.7459,
57
+ "step": 4000
58
+ },
59
+ {
60
+ "epoch": 0.58,
61
+ "learning_rate": 0.00028431782540512285,
62
+ "loss": 0.7282,
63
+ "step": 4500
64
+ },
65
+ {
66
+ "epoch": 0.65,
67
+ "learning_rate": 0.00028235755358076316,
68
+ "loss": 0.7104,
69
+ "step": 5000
70
+ },
71
+ {
72
+ "epoch": 0.71,
73
+ "learning_rate": 0.0002803972817564035,
74
+ "loss": 0.692,
75
+ "step": 5500
76
+ },
77
+ {
78
+ "epoch": 0.78,
79
+ "learning_rate": 0.0002784370099320439,
80
+ "loss": 0.6954,
81
+ "step": 6000
82
+ },
83
+ {
84
+ "epoch": 0.84,
85
+ "learning_rate": 0.00027647673810768424,
86
+ "loss": 0.6844,
87
+ "step": 6500
88
+ },
89
+ {
90
+ "epoch": 0.91,
91
+ "learning_rate": 0.0002745164662833246,
92
+ "loss": 0.6779,
93
+ "step": 7000
94
+ },
95
+ {
96
+ "epoch": 0.97,
97
+ "learning_rate": 0.00027255619445896496,
98
+ "loss": 0.6595,
99
+ "step": 7500
100
+ },
101
+ {
102
+ "epoch": 1.0,
103
+ "eval_loss": 0.44951537251472473,
104
+ "eval_runtime": 11.7967,
105
+ "eval_samples_per_second": 60.271,
106
+ "eval_steps_per_second": 7.544,
107
+ "eval_wer": 0.297432521395655,
108
+ "step": 7702
109
+ },
110
+ {
111
+ "epoch": 1.04,
112
+ "learning_rate": 0.0002705959226346053,
113
+ "loss": 0.6426,
114
+ "step": 8000
115
+ },
116
+ {
117
+ "epoch": 1.1,
118
+ "learning_rate": 0.0002686356508102457,
119
+ "loss": 0.6215,
120
+ "step": 8500
121
+ },
122
+ {
123
+ "epoch": 1.17,
124
+ "learning_rate": 0.00026667537898588604,
125
+ "loss": 0.6209,
126
+ "step": 9000
127
+ },
128
+ {
129
+ "epoch": 1.23,
130
+ "learning_rate": 0.0002647151071615264,
131
+ "loss": 0.6134,
132
+ "step": 9500
133
+ },
134
+ {
135
+ "epoch": 1.3,
136
+ "learning_rate": 0.00026275483533716676,
137
+ "loss": 0.6127,
138
+ "step": 10000
139
+ },
140
+ {
141
+ "epoch": 1.36,
142
+ "learning_rate": 0.00026079456351280707,
143
+ "loss": 0.5987,
144
+ "step": 10500
145
+ },
146
+ {
147
+ "epoch": 1.43,
148
+ "learning_rate": 0.0002588342916884474,
149
+ "loss": 0.6028,
150
+ "step": 11000
151
+ },
152
+ {
153
+ "epoch": 1.49,
154
+ "learning_rate": 0.0002568740198640878,
155
+ "loss": 0.5893,
156
+ "step": 11500
157
+ },
158
+ {
159
+ "epoch": 1.56,
160
+ "learning_rate": 0.00025491374803972815,
161
+ "loss": 0.59,
162
+ "step": 12000
163
+ },
164
+ {
165
+ "epoch": 1.62,
166
+ "learning_rate": 0.0002529534762153685,
167
+ "loss": 0.594,
168
+ "step": 12500
169
+ },
170
+ {
171
+ "epoch": 1.69,
172
+ "learning_rate": 0.00025099320439100887,
173
+ "loss": 0.5903,
174
+ "step": 13000
175
+ },
176
+ {
177
+ "epoch": 1.75,
178
+ "learning_rate": 0.0002490329325666492,
179
+ "loss": 0.5779,
180
+ "step": 13500
181
+ },
182
+ {
183
+ "epoch": 1.82,
184
+ "learning_rate": 0.0002470726607422896,
185
+ "loss": 0.582,
186
+ "step": 14000
187
+ },
188
+ {
189
+ "epoch": 1.88,
190
+ "learning_rate": 0.00024511238891792995,
191
+ "loss": 0.5781,
192
+ "step": 14500
193
+ },
194
+ {
195
+ "epoch": 1.95,
196
+ "learning_rate": 0.0002431521170935703,
197
+ "loss": 0.5717,
198
+ "step": 15000
199
+ },
200
+ {
201
+ "epoch": 2.0,
202
+ "eval_loss": 0.39816993474960327,
203
+ "eval_runtime": 11.9909,
204
+ "eval_samples_per_second": 59.295,
205
+ "eval_steps_per_second": 7.422,
206
+ "eval_wer": 0.25622119815668204,
207
+ "step": 15404
208
+ },
209
+ {
210
+ "epoch": 2.01,
211
+ "learning_rate": 0.00024119184526921064,
212
+ "loss": 0.5632,
213
+ "step": 15500
214
+ },
215
+ {
216
+ "epoch": 2.08,
217
+ "learning_rate": 0.000239231573444851,
218
+ "loss": 0.5298,
219
+ "step": 16000
220
+ },
221
+ {
222
+ "epoch": 2.14,
223
+ "learning_rate": 0.00023727130162049136,
224
+ "loss": 0.5391,
225
+ "step": 16500
226
+ },
227
+ {
228
+ "epoch": 2.21,
229
+ "learning_rate": 0.00023531102979613172,
230
+ "loss": 0.5274,
231
+ "step": 17000
232
+ },
233
+ {
234
+ "epoch": 2.27,
235
+ "learning_rate": 0.00023335075797177208,
236
+ "loss": 0.5368,
237
+ "step": 17500
238
+ },
239
+ {
240
+ "epoch": 2.34,
241
+ "learning_rate": 0.00023139048614741244,
242
+ "loss": 0.5191,
243
+ "step": 18000
244
+ },
245
+ {
246
+ "epoch": 2.4,
247
+ "learning_rate": 0.00022943021432305277,
248
+ "loss": 0.523,
249
+ "step": 18500
250
+ },
251
+ {
252
+ "epoch": 2.47,
253
+ "learning_rate": 0.00022746994249869313,
254
+ "loss": 0.529,
255
+ "step": 19000
256
+ },
257
+ {
258
+ "epoch": 2.53,
259
+ "learning_rate": 0.0002255096706743335,
260
+ "loss": 0.5293,
261
+ "step": 19500
262
+ },
263
+ {
264
+ "epoch": 2.6,
265
+ "learning_rate": 0.00022354939884997385,
266
+ "loss": 0.5091,
267
+ "step": 20000
268
+ },
269
+ {
270
+ "epoch": 2.66,
271
+ "learning_rate": 0.0002215891270256142,
272
+ "loss": 0.5258,
273
+ "step": 20500
274
+ },
275
+ {
276
+ "epoch": 2.73,
277
+ "learning_rate": 0.00021962885520125454,
278
+ "loss": 0.5107,
279
+ "step": 21000
280
+ },
281
+ {
282
+ "epoch": 2.79,
283
+ "learning_rate": 0.0002176685833768949,
284
+ "loss": 0.5197,
285
+ "step": 21500
286
+ },
287
+ {
288
+ "epoch": 2.86,
289
+ "learning_rate": 0.00021570831155253526,
290
+ "loss": 0.5162,
291
+ "step": 22000
292
+ },
293
+ {
294
+ "epoch": 2.92,
295
+ "learning_rate": 0.00021374803972817562,
296
+ "loss": 0.5162,
297
+ "step": 22500
298
+ },
299
+ {
300
+ "epoch": 2.99,
301
+ "learning_rate": 0.00021178776790381598,
302
+ "loss": 0.5134,
303
+ "step": 23000
304
+ },
305
+ {
306
+ "epoch": 3.0,
307
+ "eval_loss": 0.3769395649433136,
308
+ "eval_runtime": 12.4234,
309
+ "eval_samples_per_second": 57.231,
310
+ "eval_steps_per_second": 7.164,
311
+ "eval_wer": 0.23647136273864383,
312
+ "step": 23106
313
+ },
314
+ {
315
+ "epoch": 3.05,
316
+ "learning_rate": 0.00020982749607945634,
317
+ "loss": 0.4797,
318
+ "step": 23500
319
+ },
320
+ {
321
+ "epoch": 3.12,
322
+ "learning_rate": 0.00020786722425509668,
323
+ "loss": 0.4777,
324
+ "step": 24000
325
+ },
326
+ {
327
+ "epoch": 3.18,
328
+ "learning_rate": 0.00020590695243073704,
329
+ "loss": 0.4729,
330
+ "step": 24500
331
+ },
332
+ {
333
+ "epoch": 3.25,
334
+ "learning_rate": 0.0002039466806063774,
335
+ "loss": 0.4749,
336
+ "step": 25000
337
+ },
338
+ {
339
+ "epoch": 3.31,
340
+ "learning_rate": 0.00020198640878201776,
341
+ "loss": 0.4803,
342
+ "step": 25500
343
+ },
344
+ {
345
+ "epoch": 3.38,
346
+ "learning_rate": 0.00020002613695765812,
347
+ "loss": 0.4738,
348
+ "step": 26000
349
+ },
350
+ {
351
+ "epoch": 3.44,
352
+ "learning_rate": 0.00019806586513329845,
353
+ "loss": 0.4782,
354
+ "step": 26500
355
+ },
356
+ {
357
+ "epoch": 3.51,
358
+ "learning_rate": 0.0001961055933089388,
359
+ "loss": 0.4994,
360
+ "step": 27000
361
+ },
362
+ {
363
+ "epoch": 3.57,
364
+ "learning_rate": 0.00019414532148457917,
365
+ "loss": 0.4676,
366
+ "step": 27500
367
+ },
368
+ {
369
+ "epoch": 3.64,
370
+ "learning_rate": 0.00019218504966021953,
371
+ "loss": 0.457,
372
+ "step": 28000
373
+ },
374
+ {
375
+ "epoch": 3.7,
376
+ "learning_rate": 0.0001902247778358599,
377
+ "loss": 0.469,
378
+ "step": 28500
379
+ },
380
+ {
381
+ "epoch": 3.77,
382
+ "learning_rate": 0.00018826450601150025,
383
+ "loss": 0.4573,
384
+ "step": 29000
385
+ },
386
+ {
387
+ "epoch": 3.83,
388
+ "learning_rate": 0.00018630423418714058,
389
+ "loss": 0.4672,
390
+ "step": 29500
391
+ },
392
+ {
393
+ "epoch": 3.9,
394
+ "learning_rate": 0.00018434396236278094,
395
+ "loss": 0.4736,
396
+ "step": 30000
397
+ },
398
+ {
399
+ "epoch": 3.96,
400
+ "learning_rate": 0.0001823836905384213,
401
+ "loss": 0.467,
402
+ "step": 30500
403
+ },
404
+ {
405
+ "epoch": 4.0,
406
+ "eval_loss": 0.34988921880722046,
407
+ "eval_runtime": 12.3506,
408
+ "eval_samples_per_second": 57.568,
409
+ "eval_steps_per_second": 7.206,
410
+ "eval_wer": 0.22027649769585253,
411
+ "step": 30808
412
+ },
413
+ {
414
+ "epoch": 4.02,
415
+ "learning_rate": 0.00018042341871406166,
416
+ "loss": 0.4541,
417
+ "step": 31000
418
+ },
419
+ {
420
+ "epoch": 4.09,
421
+ "learning_rate": 0.00017846314688970202,
422
+ "loss": 0.4273,
423
+ "step": 31500
424
+ },
425
+ {
426
+ "epoch": 4.15,
427
+ "learning_rate": 0.00017650287506534236,
428
+ "loss": 0.4334,
429
+ "step": 32000
430
+ },
431
+ {
432
+ "epoch": 4.22,
433
+ "learning_rate": 0.00017454260324098272,
434
+ "loss": 0.4203,
435
+ "step": 32500
436
+ },
437
+ {
438
+ "epoch": 4.28,
439
+ "learning_rate": 0.00017258233141662308,
440
+ "loss": 0.4331,
441
+ "step": 33000
442
+ },
443
+ {
444
+ "epoch": 4.35,
445
+ "learning_rate": 0.00017062205959226344,
446
+ "loss": 0.4306,
447
+ "step": 33500
448
+ },
449
+ {
450
+ "epoch": 4.41,
451
+ "learning_rate": 0.00016866178776790382,
452
+ "loss": 0.4251,
453
+ "step": 34000
454
+ },
455
+ {
456
+ "epoch": 4.48,
457
+ "learning_rate": 0.00016670151594354418,
458
+ "loss": 0.4231,
459
+ "step": 34500
460
+ },
461
+ {
462
+ "epoch": 4.54,
463
+ "learning_rate": 0.0001647412441191845,
464
+ "loss": 0.4256,
465
+ "step": 35000
466
+ },
467
+ {
468
+ "epoch": 4.61,
469
+ "learning_rate": 0.00016278097229482488,
470
+ "loss": 0.4332,
471
+ "step": 35500
472
+ },
473
+ {
474
+ "epoch": 4.67,
475
+ "learning_rate": 0.00016082070047046524,
476
+ "loss": 0.4238,
477
+ "step": 36000
478
+ },
479
+ {
480
+ "epoch": 4.74,
481
+ "learning_rate": 0.0001588604286461056,
482
+ "loss": 0.4251,
483
+ "step": 36500
484
+ },
485
+ {
486
+ "epoch": 4.8,
487
+ "learning_rate": 0.00015690015682174596,
488
+ "loss": 0.425,
489
+ "step": 37000
490
+ },
491
+ {
492
+ "epoch": 4.87,
493
+ "learning_rate": 0.0001549398849973863,
494
+ "loss": 0.4208,
495
+ "step": 37500
496
+ },
497
+ {
498
+ "epoch": 4.93,
499
+ "learning_rate": 0.00015297961317302665,
500
+ "loss": 0.4245,
501
+ "step": 38000
502
+ },
503
+ {
504
+ "epoch": 5.0,
505
+ "learning_rate": 0.000151019341348667,
506
+ "loss": 0.4156,
507
+ "step": 38500
508
+ },
509
+ {
510
+ "epoch": 5.0,
511
+ "eval_loss": 0.3391418159008026,
512
+ "eval_runtime": 12.5768,
513
+ "eval_samples_per_second": 56.533,
514
+ "eval_steps_per_second": 7.077,
515
+ "eval_wer": 0.21158657011191573,
516
+ "step": 38510
517
+ },
518
+ {
519
+ "epoch": 5.06,
520
+ "learning_rate": 0.00014905906952430737,
521
+ "loss": 0.3845,
522
+ "step": 39000
523
+ },
524
+ {
525
+ "epoch": 5.13,
526
+ "learning_rate": 0.0001470987976999477,
527
+ "loss": 0.393,
528
+ "step": 39500
529
+ },
530
+ {
531
+ "epoch": 5.19,
532
+ "learning_rate": 0.00014513852587558806,
533
+ "loss": 0.3938,
534
+ "step": 40000
535
+ },
536
+ {
537
+ "epoch": 5.26,
538
+ "learning_rate": 0.00014317825405122842,
539
+ "loss": 0.3902,
540
+ "step": 40500
541
+ },
542
+ {
543
+ "epoch": 5.32,
544
+ "learning_rate": 0.00014121798222686878,
545
+ "loss": 0.3817,
546
+ "step": 41000
547
+ },
548
+ {
549
+ "epoch": 5.39,
550
+ "learning_rate": 0.00013925771040250914,
551
+ "loss": 0.3937,
552
+ "step": 41500
553
+ },
554
+ {
555
+ "epoch": 5.45,
556
+ "learning_rate": 0.0001372974385781495,
557
+ "loss": 0.3821,
558
+ "step": 42000
559
+ },
560
+ {
561
+ "epoch": 5.52,
562
+ "learning_rate": 0.00013533716675378984,
563
+ "loss": 0.3903,
564
+ "step": 42500
565
+ },
566
+ {
567
+ "epoch": 5.58,
568
+ "learning_rate": 0.0001333768949294302,
569
+ "loss": 0.381,
570
+ "step": 43000
571
+ },
572
+ {
573
+ "epoch": 5.65,
574
+ "learning_rate": 0.00013141662310507056,
575
+ "loss": 0.3866,
576
+ "step": 43500
577
+ },
578
+ {
579
+ "epoch": 5.71,
580
+ "learning_rate": 0.00012945635128071092,
581
+ "loss": 0.3816,
582
+ "step": 44000
583
+ },
584
+ {
585
+ "epoch": 5.78,
586
+ "learning_rate": 0.00012749607945635128,
587
+ "loss": 0.3845,
588
+ "step": 44500
589
+ },
590
+ {
591
+ "epoch": 5.84,
592
+ "learning_rate": 0.0001255358076319916,
593
+ "loss": 0.3777,
594
+ "step": 45000
595
+ },
596
+ {
597
+ "epoch": 5.91,
598
+ "learning_rate": 0.00012357553580763197,
599
+ "loss": 0.3777,
600
+ "step": 45500
601
+ },
602
+ {
603
+ "epoch": 5.97,
604
+ "learning_rate": 0.00012161526398327234,
605
+ "loss": 0.379,
606
+ "step": 46000
607
+ },
608
+ {
609
+ "epoch": 6.0,
610
+ "eval_loss": 0.3327275514602661,
611
+ "eval_runtime": 12.31,
612
+ "eval_samples_per_second": 57.758,
613
+ "eval_steps_per_second": 7.23,
614
+ "eval_wer": 0.1998683344305464,
615
+ "step": 46212
616
+ },
617
+ {
618
+ "epoch": 6.04,
619
+ "learning_rate": 0.00011965499215891269,
620
+ "loss": 0.3657,
621
+ "step": 46500
622
+ },
623
+ {
624
+ "epoch": 6.1,
625
+ "learning_rate": 0.00011769472033455305,
626
+ "loss": 0.3527,
627
+ "step": 47000
628
+ },
629
+ {
630
+ "epoch": 6.17,
631
+ "learning_rate": 0.00011573444851019341,
632
+ "loss": 0.3505,
633
+ "step": 47500
634
+ },
635
+ {
636
+ "epoch": 6.23,
637
+ "learning_rate": 0.00011377417668583375,
638
+ "loss": 0.3472,
639
+ "step": 48000
640
+ },
641
+ {
642
+ "epoch": 6.3,
643
+ "learning_rate": 0.00011181390486147411,
644
+ "loss": 0.3497,
645
+ "step": 48500
646
+ },
647
+ {
648
+ "epoch": 6.36,
649
+ "learning_rate": 0.00010985363303711446,
650
+ "loss": 0.349,
651
+ "step": 49000
652
+ },
653
+ {
654
+ "epoch": 6.43,
655
+ "learning_rate": 0.00010789336121275482,
656
+ "loss": 0.3545,
657
+ "step": 49500
658
+ },
659
+ {
660
+ "epoch": 6.49,
661
+ "learning_rate": 0.00010593308938839518,
662
+ "loss": 0.3453,
663
+ "step": 50000
664
+ },
665
+ {
666
+ "epoch": 6.56,
667
+ "learning_rate": 0.00010397281756403553,
668
+ "loss": 0.3501,
669
+ "step": 50500
670
+ },
671
+ {
672
+ "epoch": 6.62,
673
+ "learning_rate": 0.00010201254573967589,
674
+ "loss": 0.3473,
675
+ "step": 51000
676
+ },
677
+ {
678
+ "epoch": 6.69,
679
+ "learning_rate": 0.00010005227391531626,
680
+ "loss": 0.3462,
681
+ "step": 51500
682
+ },
683
+ {
684
+ "epoch": 6.75,
685
+ "learning_rate": 9.80920020909566e-05,
686
+ "loss": 0.3385,
687
+ "step": 52000
688
+ },
689
+ {
690
+ "epoch": 6.82,
691
+ "learning_rate": 9.613173026659697e-05,
692
+ "loss": 0.3459,
693
+ "step": 52500
694
+ },
695
+ {
696
+ "epoch": 6.88,
697
+ "learning_rate": 9.417145844223731e-05,
698
+ "loss": 0.3441,
699
+ "step": 53000
700
+ },
701
+ {
702
+ "epoch": 6.95,
703
+ "learning_rate": 9.221118661787767e-05,
704
+ "loss": 0.3475,
705
+ "step": 53500
706
+ },
707
+ {
708
+ "epoch": 7.0,
709
+ "eval_loss": 0.31270119547843933,
710
+ "eval_runtime": 12.4561,
711
+ "eval_samples_per_second": 57.08,
712
+ "eval_steps_per_second": 7.145,
713
+ "eval_wer": 0.1947333772218565,
714
+ "step": 53914
715
+ },
716
+ {
717
+ "epoch": 7.01,
718
+ "learning_rate": 9.025091479351803e-05,
719
+ "loss": 0.3379,
720
+ "step": 54000
721
+ },
722
+ {
723
+ "epoch": 7.08,
724
+ "learning_rate": 8.829064296915838e-05,
725
+ "loss": 0.3159,
726
+ "step": 54500
727
+ },
728
+ {
729
+ "epoch": 7.14,
730
+ "learning_rate": 8.633037114479874e-05,
731
+ "loss": 0.3254,
732
+ "step": 55000
733
+ },
734
+ {
735
+ "epoch": 7.21,
736
+ "learning_rate": 8.43700993204391e-05,
737
+ "loss": 0.3195,
738
+ "step": 55500
739
+ },
740
+ {
741
+ "epoch": 7.27,
742
+ "learning_rate": 8.240982749607945e-05,
743
+ "loss": 0.3225,
744
+ "step": 56000
745
+ },
746
+ {
747
+ "epoch": 7.34,
748
+ "learning_rate": 8.044955567171981e-05,
749
+ "loss": 0.3128,
750
+ "step": 56500
751
+ },
752
+ {
753
+ "epoch": 7.4,
754
+ "learning_rate": 7.848928384736017e-05,
755
+ "loss": 0.318,
756
+ "step": 57000
757
+ },
758
+ {
759
+ "epoch": 7.47,
760
+ "learning_rate": 7.652901202300051e-05,
761
+ "loss": 0.3165,
762
+ "step": 57500
763
+ },
764
+ {
765
+ "epoch": 7.53,
766
+ "learning_rate": 7.456874019864087e-05,
767
+ "loss": 0.3191,
768
+ "step": 58000
769
+ },
770
+ {
771
+ "epoch": 7.6,
772
+ "learning_rate": 7.260846837428123e-05,
773
+ "loss": 0.3169,
774
+ "step": 58500
775
+ },
776
+ {
777
+ "epoch": 7.66,
778
+ "learning_rate": 7.064819654992158e-05,
779
+ "loss": 0.3118,
780
+ "step": 59000
781
+ },
782
+ {
783
+ "epoch": 7.73,
784
+ "learning_rate": 6.868792472556194e-05,
785
+ "loss": 0.3192,
786
+ "step": 59500
787
+ },
788
+ {
789
+ "epoch": 7.79,
790
+ "learning_rate": 6.67276529012023e-05,
791
+ "loss": 0.3055,
792
+ "step": 60000
793
+ },
794
+ {
795
+ "epoch": 7.86,
796
+ "learning_rate": 6.476738107684265e-05,
797
+ "loss": 0.3134,
798
+ "step": 60500
799
+ },
800
+ {
801
+ "epoch": 7.92,
802
+ "learning_rate": 6.2807109252483e-05,
803
+ "loss": 0.3081,
804
+ "step": 61000
805
+ },
806
+ {
807
+ "epoch": 7.98,
808
+ "learning_rate": 6.084683742812336e-05,
809
+ "loss": 0.3105,
810
+ "step": 61500
811
+ },
812
+ {
813
+ "epoch": 8.0,
814
+ "eval_loss": 0.30813026428222656,
815
+ "eval_runtime": 12.2002,
816
+ "eval_samples_per_second": 58.278,
817
+ "eval_steps_per_second": 7.295,
818
+ "eval_wer": 0.1814351547070441,
819
+ "step": 61616
820
+ },
821
+ {
822
+ "epoch": 8.05,
823
+ "learning_rate": 5.888656560376372e-05,
824
+ "loss": 0.2986,
825
+ "step": 62000
826
+ },
827
+ {
828
+ "epoch": 8.11,
829
+ "learning_rate": 5.692629377940407e-05,
830
+ "loss": 0.2908,
831
+ "step": 62500
832
+ },
833
+ {
834
+ "epoch": 8.18,
835
+ "learning_rate": 5.4966021955044426e-05,
836
+ "loss": 0.2835,
837
+ "step": 63000
838
+ },
839
+ {
840
+ "epoch": 8.24,
841
+ "learning_rate": 5.3005750130684786e-05,
842
+ "loss": 0.2904,
843
+ "step": 63500
844
+ },
845
+ {
846
+ "epoch": 8.31,
847
+ "learning_rate": 5.104547830632514e-05,
848
+ "loss": 0.289,
849
+ "step": 64000
850
+ },
851
+ {
852
+ "epoch": 8.37,
853
+ "learning_rate": 4.908520648196549e-05,
854
+ "loss": 0.2905,
855
+ "step": 64500
856
+ },
857
+ {
858
+ "epoch": 8.44,
859
+ "learning_rate": 4.7124934657605846e-05,
860
+ "loss": 0.288,
861
+ "step": 65000
862
+ },
863
+ {
864
+ "epoch": 8.5,
865
+ "learning_rate": 4.516466283324621e-05,
866
+ "loss": 0.2813,
867
+ "step": 65500
868
+ },
869
+ {
870
+ "epoch": 8.57,
871
+ "learning_rate": 4.3204391008886566e-05,
872
+ "loss": 0.2831,
873
+ "step": 66000
874
+ },
875
+ {
876
+ "epoch": 8.63,
877
+ "learning_rate": 4.124411918452692e-05,
878
+ "loss": 0.2914,
879
+ "step": 66500
880
+ },
881
+ {
882
+ "epoch": 8.7,
883
+ "learning_rate": 3.928384736016727e-05,
884
+ "loss": 0.2807,
885
+ "step": 67000
886
+ },
887
+ {
888
+ "epoch": 8.76,
889
+ "learning_rate": 3.7323575535807626e-05,
890
+ "loss": 0.2833,
891
+ "step": 67500
892
+ },
893
+ {
894
+ "epoch": 8.83,
895
+ "learning_rate": 3.5363303711447986e-05,
896
+ "loss": 0.2825,
897
+ "step": 68000
898
+ },
899
+ {
900
+ "epoch": 8.89,
901
+ "learning_rate": 3.340303188708834e-05,
902
+ "loss": 0.2824,
903
+ "step": 68500
904
+ },
905
+ {
906
+ "epoch": 8.96,
907
+ "learning_rate": 3.14427600627287e-05,
908
+ "loss": 0.281,
909
+ "step": 69000
910
+ },
911
+ {
912
+ "epoch": 9.0,
913
+ "eval_loss": 0.3067900538444519,
914
+ "eval_runtime": 12.6252,
915
+ "eval_samples_per_second": 56.316,
916
+ "eval_steps_per_second": 7.049,
917
+ "eval_wer": 0.17419354838709677,
918
+ "step": 69318
919
+ },
920
+ {
921
+ "epoch": 9.02,
922
+ "learning_rate": 2.948248823836905e-05,
923
+ "loss": 0.2725,
924
+ "step": 69500
925
+ },
926
+ {
927
+ "epoch": 9.09,
928
+ "learning_rate": 2.7522216414009406e-05,
929
+ "loss": 0.2696,
930
+ "step": 70000
931
+ },
932
+ {
933
+ "epoch": 9.15,
934
+ "learning_rate": 2.5561944589649766e-05,
935
+ "loss": 0.2713,
936
+ "step": 70500
937
+ },
938
+ {
939
+ "epoch": 9.22,
940
+ "learning_rate": 2.360167276529012e-05,
941
+ "loss": 0.2634,
942
+ "step": 71000
943
+ },
944
+ {
945
+ "epoch": 9.28,
946
+ "learning_rate": 2.1641400940930475e-05,
947
+ "loss": 0.2635,
948
+ "step": 71500
949
+ },
950
+ {
951
+ "epoch": 9.35,
952
+ "learning_rate": 1.968112911657083e-05,
953
+ "loss": 0.2658,
954
+ "step": 72000
955
+ },
956
+ {
957
+ "epoch": 9.41,
958
+ "learning_rate": 1.7720857292211185e-05,
959
+ "loss": 0.2648,
960
+ "step": 72500
961
+ },
962
+ {
963
+ "epoch": 9.48,
964
+ "learning_rate": 1.576058546785154e-05,
965
+ "loss": 0.273,
966
+ "step": 73000
967
+ },
968
+ {
969
+ "epoch": 9.54,
970
+ "learning_rate": 1.3800313643491895e-05,
971
+ "loss": 0.2601,
972
+ "step": 73500
973
+ },
974
+ {
975
+ "epoch": 9.61,
976
+ "learning_rate": 1.1840041819132252e-05,
977
+ "loss": 0.2598,
978
+ "step": 74000
979
+ },
980
+ {
981
+ "epoch": 9.67,
982
+ "learning_rate": 9.879769994772608e-06,
983
+ "loss": 0.2698,
984
+ "step": 74500
985
+ },
986
+ {
987
+ "epoch": 9.74,
988
+ "learning_rate": 7.919498170412963e-06,
989
+ "loss": 0.2574,
990
+ "step": 75000
991
+ },
992
+ {
993
+ "epoch": 9.8,
994
+ "learning_rate": 5.959226346053318e-06,
995
+ "loss": 0.2646,
996
+ "step": 75500
997
+ },
998
+ {
999
+ "epoch": 9.87,
1000
+ "learning_rate": 3.998954521693675e-06,
1001
+ "loss": 0.2584,
1002
+ "step": 76000
1003
+ },
1004
+ {
1005
+ "epoch": 9.93,
1006
+ "learning_rate": 2.0386826973340303e-06,
1007
+ "loss": 0.2623,
1008
+ "step": 76500
1009
+ },
1010
+ {
1011
+ "epoch": 10.0,
1012
+ "learning_rate": 7.841087297438577e-08,
1013
+ "loss": 0.2584,
1014
+ "step": 77000
1015
+ },
1016
+ {
1017
+ "epoch": 10.0,
1018
+ "eval_loss": 0.3040466010570526,
1019
+ "eval_runtime": 12.4236,
1020
+ "eval_samples_per_second": 57.23,
1021
+ "eval_steps_per_second": 7.164,
1022
+ "eval_wer": 0.17129690585911783,
1023
+ "step": 77020
1024
+ }
1025
+ ],
1026
+ "logging_steps": 500,
1027
+ "max_steps": 77020,
1028
+ "num_train_epochs": 10,
1029
+ "save_steps": 500,
1030
+ "total_flos": 1.409382714117005e+20,
1031
+ "trial_name": null,
1032
+ "trial_params": null
1033
+ }
checkpoint-77020/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:57454b8ef61e73ee41e0231ebd404fc4c44e52d4e9fbcfd019aea1cdd0b25721
3
+ size 4536
checkpoint-77020/vocab.json ADDED
@@ -0,0 +1,103 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "\"": 1,
3
+ "#": 2,
4
+ "$": 3,
5
+ "&": 4,
6
+ "0": 5,
7
+ "1": 6,
8
+ "2": 7,
9
+ "3": 8,
10
+ "4": 9,
11
+ "5": 10,
12
+ "6": 11,
13
+ "7": 12,
14
+ "8": 13,
15
+ "9": 14,
16
+ ";": 15,
17
+ "[PAD]": 100,
18
+ "[UNK]": 99,
19
+ "\\": 16,
20
+ "_": 17,
21
+ "`": 18,
22
+ "a": 19,
23
+ "b": 20,
24
+ "c": 21,
25
+ "d": 22,
26
+ "e": 23,
27
+ "f": 24,
28
+ "g": 25,
29
+ "h": 26,
30
+ "i": 27,
31
+ "j": 28,
32
+ "k": 29,
33
+ "l": 30,
34
+ "m": 31,
35
+ "n": 32,
36
+ "o": 33,
37
+ "p": 34,
38
+ "q": 35,
39
+ "r": 36,
40
+ "s": 37,
41
+ "t": 38,
42
+ "u": 39,
43
+ "v": 40,
44
+ "w": 41,
45
+ "x": 42,
46
+ "y": 43,
47
+ "z": 44,
48
+ "|": 0,
49
+ "à": 45,
50
+ "á": 46,
51
+ "ç": 47,
52
+ "è": 48,
53
+ "é": 49,
54
+ "ê": 50,
55
+ "ì": 51,
56
+ "í": 52,
57
+ "ñ": 53,
58
+ "ó": 54,
59
+ "ö": 55,
60
+ "ø": 56,
61
+ "ú": 57,
62
+ "ī": 58,
63
+ "ō": 59,
64
+ "ơ": 60,
65
+ "ư": 61,
66
+ "ǒ": 62,
67
+ "ǔ": 63,
68
+ "˚": 64,
69
+ "а": 65,
70
+ "в": 66,
71
+ "к": 67,
72
+ "о": 68,
73
+ "с": 69,
74
+ "т": 70,
75
+ "у": 71,
76
+ "ы": 72,
77
+ "э": 73,
78
+ "ờ": 74,
79
+ "​": 75,
80
+ "–": 76,
81
+ "—": 77,
82
+ "’": 78,
83
+ "…": 79,
84
+ "⁄": 80,
85
+ "℃": 81,
86
+ "℉": 82,
87
+ "≈": 83,
88
+ "♪": 84,
89
+ "ㄲ": 85,
90
+ "ㄴ": 86,
91
+ "ㅠ": 87,
92
+ "간": 88,
93
+ "구": 89,
94
+ "님": 90,
95
+ "상": 91,
96
+ "설": 92,
97
+ "아": 93,
98
+ "왜": 94,
99
+ "음": 95,
100
+ "이": 96,
101
+ "절": 97,
102
+ "정": 98
103
+ }
config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
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
+ "Wav2Vec2ForCTC"
11
+ ],
12
+ "attention_dropout": 0.1,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 768,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": true,
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": "mean",
46
+ "ctc_zero_infinity": false,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": true,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_dropout": 0.0,
52
+ "feat_extract_norm": "layer",
53
+ "feat_proj_dropout": 0.05,
54
+ "feat_quantizer_dropout": 0.0,
55
+ "final_dropout": 0.0,
56
+ "gradient_checkpointing": false,
57
+ "hidden_act": "gelu",
58
+ "hidden_dropout": 0.05,
59
+ "hidden_size": 1024,
60
+ "initializer_range": 0.02,
61
+ "intermediate_size": 4096,
62
+ "layer_norm_eps": 1e-05,
63
+ "layerdrop": 0.05,
64
+ "mask_channel_length": 10,
65
+ "mask_channel_min_space": 1,
66
+ "mask_channel_other": 0.0,
67
+ "mask_channel_prob": 0.0,
68
+ "mask_channel_selection": "static",
69
+ "mask_feature_length": 10,
70
+ "mask_feature_min_masks": 0,
71
+ "mask_feature_prob": 0.0,
72
+ "mask_time_length": 10,
73
+ "mask_time_min_masks": 2,
74
+ "mask_time_min_space": 1,
75
+ "mask_time_other": 0.0,
76
+ "mask_time_prob": 0.05,
77
+ "mask_time_selection": "static",
78
+ "model_type": "wav2vec2",
79
+ "num_adapter_layers": 3,
80
+ "num_attention_heads": 16,
81
+ "num_codevector_groups": 2,
82
+ "num_codevectors_per_group": 320,
83
+ "num_conv_pos_embedding_groups": 16,
84
+ "num_conv_pos_embeddings": 128,
85
+ "num_feat_extract_layers": 7,
86
+ "num_hidden_layers": 24,
87
+ "num_negatives": 100,
88
+ "output_hidden_size": 1024,
89
+ "pad_token_id": 100,
90
+ "proj_codevector_dim": 768,
91
+ "tdnn_dilation": [
92
+ 1,
93
+ 2,
94
+ 3,
95
+ 1,
96
+ 1
97
+ ],
98
+ "tdnn_dim": [
99
+ 512,
100
+ 512,
101
+ 512,
102
+ 512,
103
+ 1500
104
+ ],
105
+ "tdnn_kernel": [
106
+ 5,
107
+ 3,
108
+ 3,
109
+ 1,
110
+ 1
111
+ ],
112
+ "torch_dtype": "float32",
113
+ "transformers_version": "4.34.1",
114
+ "use_weighted_layer_sum": false,
115
+ "vocab_size": 103,
116
+ "xvector_output_dim": 512
117
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 10.0,
3
+ "eval_loss": 0.3041677474975586,
4
+ "eval_runtime": 10.735,
5
+ "eval_samples": 711,
6
+ "eval_samples_per_second": 66.232,
7
+ "eval_steps_per_second": 8.291,
8
+ "eval_wer": 0.1695852534562212
9
+ }
preprocessor_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0,
7
+ "processor_class": "Wav2Vec2Processor",
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c9bdf0fd4f9034a342a7bdb6cfbf49f649ce26a10a80c8194598687bd004a2f2
3
+ size 1262324650
special_tokens_map.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "[PAD]",
18
+ "lstrip": true,
19
+ "normalized": false,
20
+ "rstrip": true,
21
+ "single_word": false
22
+ },
23
+ "unk_token": {
24
+ "content": "[UNK]",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": true,
28
+ "single_word": false
29
+ }
30
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "99": {
4
+ "content": "[UNK]",
5
+ "lstrip": true,
6
+ "normalized": false,
7
+ "rstrip": true,
8
+ "single_word": false,
9
+ "special": false
10
+ },
11
+ "100": {
12
+ "content": "[PAD]",
13
+ "lstrip": true,
14
+ "normalized": false,
15
+ "rstrip": true,
16
+ "single_word": false,
17
+ "special": false
18
+ },
19
+ "101": {
20
+ "content": "<s>",
21
+ "lstrip": false,
22
+ "normalized": true,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "102": {
28
+ "content": "</s>",
29
+ "lstrip": false,
30
+ "normalized": true,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ }
35
+ },
36
+ "bos_token": "<s>",
37
+ "clean_up_tokenization_spaces": true,
38
+ "config": null,
39
+ "do_lower_case": false,
40
+ "eos_token": "</s>",
41
+ "model_max_length": 1000000000000000019884624838656,
42
+ "pad_token": "[PAD]",
43
+ "processor_class": "Wav2Vec2Processor",
44
+ "replace_word_delimiter_char": " ",
45
+ "target_lang": null,
46
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
47
+ "tokenizer_type": "wav2vec2",
48
+ "trust_remote_code": false,
49
+ "unk_token": "[UNK]",
50
+ "word_delimiter_token": "|"
51
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 10.0,
3
+ "train_loss": 0.4655882824058503,
4
+ "train_runtime": 41869.3879,
5
+ "train_samples": 123225,
6
+ "train_samples_per_second": 29.431,
7
+ "train_steps_per_second": 1.84
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1042 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.3040466010570526,
3
+ "best_model_checkpoint": "./enko_xlsr_13p_run1/checkpoint-77020",
4
+ "epoch": 10.0,
5
+ "eval_steps": 500,
6
+ "global_step": 77020,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.06,
13
+ "learning_rate": 0.0003,
14
+ "loss": 4.7797,
15
+ "step": 500
16
+ },
17
+ {
18
+ "epoch": 0.13,
19
+ "learning_rate": 0.00029803972817564033,
20
+ "loss": 1.2377,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.19,
25
+ "learning_rate": 0.0002960794563512807,
26
+ "loss": 0.9452,
27
+ "step": 1500
28
+ },
29
+ {
30
+ "epoch": 0.26,
31
+ "learning_rate": 0.000294119184526921,
32
+ "loss": 0.8622,
33
+ "step": 2000
34
+ },
35
+ {
36
+ "epoch": 0.32,
37
+ "learning_rate": 0.0002921589127025614,
38
+ "loss": 0.82,
39
+ "step": 2500
40
+ },
41
+ {
42
+ "epoch": 0.39,
43
+ "learning_rate": 0.0002901986408782018,
44
+ "loss": 0.7742,
45
+ "step": 3000
46
+ },
47
+ {
48
+ "epoch": 0.45,
49
+ "learning_rate": 0.00028823836905384213,
50
+ "loss": 0.75,
51
+ "step": 3500
52
+ },
53
+ {
54
+ "epoch": 0.52,
55
+ "learning_rate": 0.0002862780972294825,
56
+ "loss": 0.7459,
57
+ "step": 4000
58
+ },
59
+ {
60
+ "epoch": 0.58,
61
+ "learning_rate": 0.00028431782540512285,
62
+ "loss": 0.7282,
63
+ "step": 4500
64
+ },
65
+ {
66
+ "epoch": 0.65,
67
+ "learning_rate": 0.00028235755358076316,
68
+ "loss": 0.7104,
69
+ "step": 5000
70
+ },
71
+ {
72
+ "epoch": 0.71,
73
+ "learning_rate": 0.0002803972817564035,
74
+ "loss": 0.692,
75
+ "step": 5500
76
+ },
77
+ {
78
+ "epoch": 0.78,
79
+ "learning_rate": 0.0002784370099320439,
80
+ "loss": 0.6954,
81
+ "step": 6000
82
+ },
83
+ {
84
+ "epoch": 0.84,
85
+ "learning_rate": 0.00027647673810768424,
86
+ "loss": 0.6844,
87
+ "step": 6500
88
+ },
89
+ {
90
+ "epoch": 0.91,
91
+ "learning_rate": 0.0002745164662833246,
92
+ "loss": 0.6779,
93
+ "step": 7000
94
+ },
95
+ {
96
+ "epoch": 0.97,
97
+ "learning_rate": 0.00027255619445896496,
98
+ "loss": 0.6595,
99
+ "step": 7500
100
+ },
101
+ {
102
+ "epoch": 1.0,
103
+ "eval_loss": 0.44951537251472473,
104
+ "eval_runtime": 11.7967,
105
+ "eval_samples_per_second": 60.271,
106
+ "eval_steps_per_second": 7.544,
107
+ "eval_wer": 0.297432521395655,
108
+ "step": 7702
109
+ },
110
+ {
111
+ "epoch": 1.04,
112
+ "learning_rate": 0.0002705959226346053,
113
+ "loss": 0.6426,
114
+ "step": 8000
115
+ },
116
+ {
117
+ "epoch": 1.1,
118
+ "learning_rate": 0.0002686356508102457,
119
+ "loss": 0.6215,
120
+ "step": 8500
121
+ },
122
+ {
123
+ "epoch": 1.17,
124
+ "learning_rate": 0.00026667537898588604,
125
+ "loss": 0.6209,
126
+ "step": 9000
127
+ },
128
+ {
129
+ "epoch": 1.23,
130
+ "learning_rate": 0.0002647151071615264,
131
+ "loss": 0.6134,
132
+ "step": 9500
133
+ },
134
+ {
135
+ "epoch": 1.3,
136
+ "learning_rate": 0.00026275483533716676,
137
+ "loss": 0.6127,
138
+ "step": 10000
139
+ },
140
+ {
141
+ "epoch": 1.36,
142
+ "learning_rate": 0.00026079456351280707,
143
+ "loss": 0.5987,
144
+ "step": 10500
145
+ },
146
+ {
147
+ "epoch": 1.43,
148
+ "learning_rate": 0.0002588342916884474,
149
+ "loss": 0.6028,
150
+ "step": 11000
151
+ },
152
+ {
153
+ "epoch": 1.49,
154
+ "learning_rate": 0.0002568740198640878,
155
+ "loss": 0.5893,
156
+ "step": 11500
157
+ },
158
+ {
159
+ "epoch": 1.56,
160
+ "learning_rate": 0.00025491374803972815,
161
+ "loss": 0.59,
162
+ "step": 12000
163
+ },
164
+ {
165
+ "epoch": 1.62,
166
+ "learning_rate": 0.0002529534762153685,
167
+ "loss": 0.594,
168
+ "step": 12500
169
+ },
170
+ {
171
+ "epoch": 1.69,
172
+ "learning_rate": 0.00025099320439100887,
173
+ "loss": 0.5903,
174
+ "step": 13000
175
+ },
176
+ {
177
+ "epoch": 1.75,
178
+ "learning_rate": 0.0002490329325666492,
179
+ "loss": 0.5779,
180
+ "step": 13500
181
+ },
182
+ {
183
+ "epoch": 1.82,
184
+ "learning_rate": 0.0002470726607422896,
185
+ "loss": 0.582,
186
+ "step": 14000
187
+ },
188
+ {
189
+ "epoch": 1.88,
190
+ "learning_rate": 0.00024511238891792995,
191
+ "loss": 0.5781,
192
+ "step": 14500
193
+ },
194
+ {
195
+ "epoch": 1.95,
196
+ "learning_rate": 0.0002431521170935703,
197
+ "loss": 0.5717,
198
+ "step": 15000
199
+ },
200
+ {
201
+ "epoch": 2.0,
202
+ "eval_loss": 0.39816993474960327,
203
+ "eval_runtime": 11.9909,
204
+ "eval_samples_per_second": 59.295,
205
+ "eval_steps_per_second": 7.422,
206
+ "eval_wer": 0.25622119815668204,
207
+ "step": 15404
208
+ },
209
+ {
210
+ "epoch": 2.01,
211
+ "learning_rate": 0.00024119184526921064,
212
+ "loss": 0.5632,
213
+ "step": 15500
214
+ },
215
+ {
216
+ "epoch": 2.08,
217
+ "learning_rate": 0.000239231573444851,
218
+ "loss": 0.5298,
219
+ "step": 16000
220
+ },
221
+ {
222
+ "epoch": 2.14,
223
+ "learning_rate": 0.00023727130162049136,
224
+ "loss": 0.5391,
225
+ "step": 16500
226
+ },
227
+ {
228
+ "epoch": 2.21,
229
+ "learning_rate": 0.00023531102979613172,
230
+ "loss": 0.5274,
231
+ "step": 17000
232
+ },
233
+ {
234
+ "epoch": 2.27,
235
+ "learning_rate": 0.00023335075797177208,
236
+ "loss": 0.5368,
237
+ "step": 17500
238
+ },
239
+ {
240
+ "epoch": 2.34,
241
+ "learning_rate": 0.00023139048614741244,
242
+ "loss": 0.5191,
243
+ "step": 18000
244
+ },
245
+ {
246
+ "epoch": 2.4,
247
+ "learning_rate": 0.00022943021432305277,
248
+ "loss": 0.523,
249
+ "step": 18500
250
+ },
251
+ {
252
+ "epoch": 2.47,
253
+ "learning_rate": 0.00022746994249869313,
254
+ "loss": 0.529,
255
+ "step": 19000
256
+ },
257
+ {
258
+ "epoch": 2.53,
259
+ "learning_rate": 0.0002255096706743335,
260
+ "loss": 0.5293,
261
+ "step": 19500
262
+ },
263
+ {
264
+ "epoch": 2.6,
265
+ "learning_rate": 0.00022354939884997385,
266
+ "loss": 0.5091,
267
+ "step": 20000
268
+ },
269
+ {
270
+ "epoch": 2.66,
271
+ "learning_rate": 0.0002215891270256142,
272
+ "loss": 0.5258,
273
+ "step": 20500
274
+ },
275
+ {
276
+ "epoch": 2.73,
277
+ "learning_rate": 0.00021962885520125454,
278
+ "loss": 0.5107,
279
+ "step": 21000
280
+ },
281
+ {
282
+ "epoch": 2.79,
283
+ "learning_rate": 0.0002176685833768949,
284
+ "loss": 0.5197,
285
+ "step": 21500
286
+ },
287
+ {
288
+ "epoch": 2.86,
289
+ "learning_rate": 0.00021570831155253526,
290
+ "loss": 0.5162,
291
+ "step": 22000
292
+ },
293
+ {
294
+ "epoch": 2.92,
295
+ "learning_rate": 0.00021374803972817562,
296
+ "loss": 0.5162,
297
+ "step": 22500
298
+ },
299
+ {
300
+ "epoch": 2.99,
301
+ "learning_rate": 0.00021178776790381598,
302
+ "loss": 0.5134,
303
+ "step": 23000
304
+ },
305
+ {
306
+ "epoch": 3.0,
307
+ "eval_loss": 0.3769395649433136,
308
+ "eval_runtime": 12.4234,
309
+ "eval_samples_per_second": 57.231,
310
+ "eval_steps_per_second": 7.164,
311
+ "eval_wer": 0.23647136273864383,
312
+ "step": 23106
313
+ },
314
+ {
315
+ "epoch": 3.05,
316
+ "learning_rate": 0.00020982749607945634,
317
+ "loss": 0.4797,
318
+ "step": 23500
319
+ },
320
+ {
321
+ "epoch": 3.12,
322
+ "learning_rate": 0.00020786722425509668,
323
+ "loss": 0.4777,
324
+ "step": 24000
325
+ },
326
+ {
327
+ "epoch": 3.18,
328
+ "learning_rate": 0.00020590695243073704,
329
+ "loss": 0.4729,
330
+ "step": 24500
331
+ },
332
+ {
333
+ "epoch": 3.25,
334
+ "learning_rate": 0.0002039466806063774,
335
+ "loss": 0.4749,
336
+ "step": 25000
337
+ },
338
+ {
339
+ "epoch": 3.31,
340
+ "learning_rate": 0.00020198640878201776,
341
+ "loss": 0.4803,
342
+ "step": 25500
343
+ },
344
+ {
345
+ "epoch": 3.38,
346
+ "learning_rate": 0.00020002613695765812,
347
+ "loss": 0.4738,
348
+ "step": 26000
349
+ },
350
+ {
351
+ "epoch": 3.44,
352
+ "learning_rate": 0.00019806586513329845,
353
+ "loss": 0.4782,
354
+ "step": 26500
355
+ },
356
+ {
357
+ "epoch": 3.51,
358
+ "learning_rate": 0.0001961055933089388,
359
+ "loss": 0.4994,
360
+ "step": 27000
361
+ },
362
+ {
363
+ "epoch": 3.57,
364
+ "learning_rate": 0.00019414532148457917,
365
+ "loss": 0.4676,
366
+ "step": 27500
367
+ },
368
+ {
369
+ "epoch": 3.64,
370
+ "learning_rate": 0.00019218504966021953,
371
+ "loss": 0.457,
372
+ "step": 28000
373
+ },
374
+ {
375
+ "epoch": 3.7,
376
+ "learning_rate": 0.0001902247778358599,
377
+ "loss": 0.469,
378
+ "step": 28500
379
+ },
380
+ {
381
+ "epoch": 3.77,
382
+ "learning_rate": 0.00018826450601150025,
383
+ "loss": 0.4573,
384
+ "step": 29000
385
+ },
386
+ {
387
+ "epoch": 3.83,
388
+ "learning_rate": 0.00018630423418714058,
389
+ "loss": 0.4672,
390
+ "step": 29500
391
+ },
392
+ {
393
+ "epoch": 3.9,
394
+ "learning_rate": 0.00018434396236278094,
395
+ "loss": 0.4736,
396
+ "step": 30000
397
+ },
398
+ {
399
+ "epoch": 3.96,
400
+ "learning_rate": 0.0001823836905384213,
401
+ "loss": 0.467,
402
+ "step": 30500
403
+ },
404
+ {
405
+ "epoch": 4.0,
406
+ "eval_loss": 0.34988921880722046,
407
+ "eval_runtime": 12.3506,
408
+ "eval_samples_per_second": 57.568,
409
+ "eval_steps_per_second": 7.206,
410
+ "eval_wer": 0.22027649769585253,
411
+ "step": 30808
412
+ },
413
+ {
414
+ "epoch": 4.02,
415
+ "learning_rate": 0.00018042341871406166,
416
+ "loss": 0.4541,
417
+ "step": 31000
418
+ },
419
+ {
420
+ "epoch": 4.09,
421
+ "learning_rate": 0.00017846314688970202,
422
+ "loss": 0.4273,
423
+ "step": 31500
424
+ },
425
+ {
426
+ "epoch": 4.15,
427
+ "learning_rate": 0.00017650287506534236,
428
+ "loss": 0.4334,
429
+ "step": 32000
430
+ },
431
+ {
432
+ "epoch": 4.22,
433
+ "learning_rate": 0.00017454260324098272,
434
+ "loss": 0.4203,
435
+ "step": 32500
436
+ },
437
+ {
438
+ "epoch": 4.28,
439
+ "learning_rate": 0.00017258233141662308,
440
+ "loss": 0.4331,
441
+ "step": 33000
442
+ },
443
+ {
444
+ "epoch": 4.35,
445
+ "learning_rate": 0.00017062205959226344,
446
+ "loss": 0.4306,
447
+ "step": 33500
448
+ },
449
+ {
450
+ "epoch": 4.41,
451
+ "learning_rate": 0.00016866178776790382,
452
+ "loss": 0.4251,
453
+ "step": 34000
454
+ },
455
+ {
456
+ "epoch": 4.48,
457
+ "learning_rate": 0.00016670151594354418,
458
+ "loss": 0.4231,
459
+ "step": 34500
460
+ },
461
+ {
462
+ "epoch": 4.54,
463
+ "learning_rate": 0.0001647412441191845,
464
+ "loss": 0.4256,
465
+ "step": 35000
466
+ },
467
+ {
468
+ "epoch": 4.61,
469
+ "learning_rate": 0.00016278097229482488,
470
+ "loss": 0.4332,
471
+ "step": 35500
472
+ },
473
+ {
474
+ "epoch": 4.67,
475
+ "learning_rate": 0.00016082070047046524,
476
+ "loss": 0.4238,
477
+ "step": 36000
478
+ },
479
+ {
480
+ "epoch": 4.74,
481
+ "learning_rate": 0.0001588604286461056,
482
+ "loss": 0.4251,
483
+ "step": 36500
484
+ },
485
+ {
486
+ "epoch": 4.8,
487
+ "learning_rate": 0.00015690015682174596,
488
+ "loss": 0.425,
489
+ "step": 37000
490
+ },
491
+ {
492
+ "epoch": 4.87,
493
+ "learning_rate": 0.0001549398849973863,
494
+ "loss": 0.4208,
495
+ "step": 37500
496
+ },
497
+ {
498
+ "epoch": 4.93,
499
+ "learning_rate": 0.00015297961317302665,
500
+ "loss": 0.4245,
501
+ "step": 38000
502
+ },
503
+ {
504
+ "epoch": 5.0,
505
+ "learning_rate": 0.000151019341348667,
506
+ "loss": 0.4156,
507
+ "step": 38500
508
+ },
509
+ {
510
+ "epoch": 5.0,
511
+ "eval_loss": 0.3391418159008026,
512
+ "eval_runtime": 12.5768,
513
+ "eval_samples_per_second": 56.533,
514
+ "eval_steps_per_second": 7.077,
515
+ "eval_wer": 0.21158657011191573,
516
+ "step": 38510
517
+ },
518
+ {
519
+ "epoch": 5.06,
520
+ "learning_rate": 0.00014905906952430737,
521
+ "loss": 0.3845,
522
+ "step": 39000
523
+ },
524
+ {
525
+ "epoch": 5.13,
526
+ "learning_rate": 0.0001470987976999477,
527
+ "loss": 0.393,
528
+ "step": 39500
529
+ },
530
+ {
531
+ "epoch": 5.19,
532
+ "learning_rate": 0.00014513852587558806,
533
+ "loss": 0.3938,
534
+ "step": 40000
535
+ },
536
+ {
537
+ "epoch": 5.26,
538
+ "learning_rate": 0.00014317825405122842,
539
+ "loss": 0.3902,
540
+ "step": 40500
541
+ },
542
+ {
543
+ "epoch": 5.32,
544
+ "learning_rate": 0.00014121798222686878,
545
+ "loss": 0.3817,
546
+ "step": 41000
547
+ },
548
+ {
549
+ "epoch": 5.39,
550
+ "learning_rate": 0.00013925771040250914,
551
+ "loss": 0.3937,
552
+ "step": 41500
553
+ },
554
+ {
555
+ "epoch": 5.45,
556
+ "learning_rate": 0.0001372974385781495,
557
+ "loss": 0.3821,
558
+ "step": 42000
559
+ },
560
+ {
561
+ "epoch": 5.52,
562
+ "learning_rate": 0.00013533716675378984,
563
+ "loss": 0.3903,
564
+ "step": 42500
565
+ },
566
+ {
567
+ "epoch": 5.58,
568
+ "learning_rate": 0.0001333768949294302,
569
+ "loss": 0.381,
570
+ "step": 43000
571
+ },
572
+ {
573
+ "epoch": 5.65,
574
+ "learning_rate": 0.00013141662310507056,
575
+ "loss": 0.3866,
576
+ "step": 43500
577
+ },
578
+ {
579
+ "epoch": 5.71,
580
+ "learning_rate": 0.00012945635128071092,
581
+ "loss": 0.3816,
582
+ "step": 44000
583
+ },
584
+ {
585
+ "epoch": 5.78,
586
+ "learning_rate": 0.00012749607945635128,
587
+ "loss": 0.3845,
588
+ "step": 44500
589
+ },
590
+ {
591
+ "epoch": 5.84,
592
+ "learning_rate": 0.0001255358076319916,
593
+ "loss": 0.3777,
594
+ "step": 45000
595
+ },
596
+ {
597
+ "epoch": 5.91,
598
+ "learning_rate": 0.00012357553580763197,
599
+ "loss": 0.3777,
600
+ "step": 45500
601
+ },
602
+ {
603
+ "epoch": 5.97,
604
+ "learning_rate": 0.00012161526398327234,
605
+ "loss": 0.379,
606
+ "step": 46000
607
+ },
608
+ {
609
+ "epoch": 6.0,
610
+ "eval_loss": 0.3327275514602661,
611
+ "eval_runtime": 12.31,
612
+ "eval_samples_per_second": 57.758,
613
+ "eval_steps_per_second": 7.23,
614
+ "eval_wer": 0.1998683344305464,
615
+ "step": 46212
616
+ },
617
+ {
618
+ "epoch": 6.04,
619
+ "learning_rate": 0.00011965499215891269,
620
+ "loss": 0.3657,
621
+ "step": 46500
622
+ },
623
+ {
624
+ "epoch": 6.1,
625
+ "learning_rate": 0.00011769472033455305,
626
+ "loss": 0.3527,
627
+ "step": 47000
628
+ },
629
+ {
630
+ "epoch": 6.17,
631
+ "learning_rate": 0.00011573444851019341,
632
+ "loss": 0.3505,
633
+ "step": 47500
634
+ },
635
+ {
636
+ "epoch": 6.23,
637
+ "learning_rate": 0.00011377417668583375,
638
+ "loss": 0.3472,
639
+ "step": 48000
640
+ },
641
+ {
642
+ "epoch": 6.3,
643
+ "learning_rate": 0.00011181390486147411,
644
+ "loss": 0.3497,
645
+ "step": 48500
646
+ },
647
+ {
648
+ "epoch": 6.36,
649
+ "learning_rate": 0.00010985363303711446,
650
+ "loss": 0.349,
651
+ "step": 49000
652
+ },
653
+ {
654
+ "epoch": 6.43,
655
+ "learning_rate": 0.00010789336121275482,
656
+ "loss": 0.3545,
657
+ "step": 49500
658
+ },
659
+ {
660
+ "epoch": 6.49,
661
+ "learning_rate": 0.00010593308938839518,
662
+ "loss": 0.3453,
663
+ "step": 50000
664
+ },
665
+ {
666
+ "epoch": 6.56,
667
+ "learning_rate": 0.00010397281756403553,
668
+ "loss": 0.3501,
669
+ "step": 50500
670
+ },
671
+ {
672
+ "epoch": 6.62,
673
+ "learning_rate": 0.00010201254573967589,
674
+ "loss": 0.3473,
675
+ "step": 51000
676
+ },
677
+ {
678
+ "epoch": 6.69,
679
+ "learning_rate": 0.00010005227391531626,
680
+ "loss": 0.3462,
681
+ "step": 51500
682
+ },
683
+ {
684
+ "epoch": 6.75,
685
+ "learning_rate": 9.80920020909566e-05,
686
+ "loss": 0.3385,
687
+ "step": 52000
688
+ },
689
+ {
690
+ "epoch": 6.82,
691
+ "learning_rate": 9.613173026659697e-05,
692
+ "loss": 0.3459,
693
+ "step": 52500
694
+ },
695
+ {
696
+ "epoch": 6.88,
697
+ "learning_rate": 9.417145844223731e-05,
698
+ "loss": 0.3441,
699
+ "step": 53000
700
+ },
701
+ {
702
+ "epoch": 6.95,
703
+ "learning_rate": 9.221118661787767e-05,
704
+ "loss": 0.3475,
705
+ "step": 53500
706
+ },
707
+ {
708
+ "epoch": 7.0,
709
+ "eval_loss": 0.31270119547843933,
710
+ "eval_runtime": 12.4561,
711
+ "eval_samples_per_second": 57.08,
712
+ "eval_steps_per_second": 7.145,
713
+ "eval_wer": 0.1947333772218565,
714
+ "step": 53914
715
+ },
716
+ {
717
+ "epoch": 7.01,
718
+ "learning_rate": 9.025091479351803e-05,
719
+ "loss": 0.3379,
720
+ "step": 54000
721
+ },
722
+ {
723
+ "epoch": 7.08,
724
+ "learning_rate": 8.829064296915838e-05,
725
+ "loss": 0.3159,
726
+ "step": 54500
727
+ },
728
+ {
729
+ "epoch": 7.14,
730
+ "learning_rate": 8.633037114479874e-05,
731
+ "loss": 0.3254,
732
+ "step": 55000
733
+ },
734
+ {
735
+ "epoch": 7.21,
736
+ "learning_rate": 8.43700993204391e-05,
737
+ "loss": 0.3195,
738
+ "step": 55500
739
+ },
740
+ {
741
+ "epoch": 7.27,
742
+ "learning_rate": 8.240982749607945e-05,
743
+ "loss": 0.3225,
744
+ "step": 56000
745
+ },
746
+ {
747
+ "epoch": 7.34,
748
+ "learning_rate": 8.044955567171981e-05,
749
+ "loss": 0.3128,
750
+ "step": 56500
751
+ },
752
+ {
753
+ "epoch": 7.4,
754
+ "learning_rate": 7.848928384736017e-05,
755
+ "loss": 0.318,
756
+ "step": 57000
757
+ },
758
+ {
759
+ "epoch": 7.47,
760
+ "learning_rate": 7.652901202300051e-05,
761
+ "loss": 0.3165,
762
+ "step": 57500
763
+ },
764
+ {
765
+ "epoch": 7.53,
766
+ "learning_rate": 7.456874019864087e-05,
767
+ "loss": 0.3191,
768
+ "step": 58000
769
+ },
770
+ {
771
+ "epoch": 7.6,
772
+ "learning_rate": 7.260846837428123e-05,
773
+ "loss": 0.3169,
774
+ "step": 58500
775
+ },
776
+ {
777
+ "epoch": 7.66,
778
+ "learning_rate": 7.064819654992158e-05,
779
+ "loss": 0.3118,
780
+ "step": 59000
781
+ },
782
+ {
783
+ "epoch": 7.73,
784
+ "learning_rate": 6.868792472556194e-05,
785
+ "loss": 0.3192,
786
+ "step": 59500
787
+ },
788
+ {
789
+ "epoch": 7.79,
790
+ "learning_rate": 6.67276529012023e-05,
791
+ "loss": 0.3055,
792
+ "step": 60000
793
+ },
794
+ {
795
+ "epoch": 7.86,
796
+ "learning_rate": 6.476738107684265e-05,
797
+ "loss": 0.3134,
798
+ "step": 60500
799
+ },
800
+ {
801
+ "epoch": 7.92,
802
+ "learning_rate": 6.2807109252483e-05,
803
+ "loss": 0.3081,
804
+ "step": 61000
805
+ },
806
+ {
807
+ "epoch": 7.98,
808
+ "learning_rate": 6.084683742812336e-05,
809
+ "loss": 0.3105,
810
+ "step": 61500
811
+ },
812
+ {
813
+ "epoch": 8.0,
814
+ "eval_loss": 0.30813026428222656,
815
+ "eval_runtime": 12.2002,
816
+ "eval_samples_per_second": 58.278,
817
+ "eval_steps_per_second": 7.295,
818
+ "eval_wer": 0.1814351547070441,
819
+ "step": 61616
820
+ },
821
+ {
822
+ "epoch": 8.05,
823
+ "learning_rate": 5.888656560376372e-05,
824
+ "loss": 0.2986,
825
+ "step": 62000
826
+ },
827
+ {
828
+ "epoch": 8.11,
829
+ "learning_rate": 5.692629377940407e-05,
830
+ "loss": 0.2908,
831
+ "step": 62500
832
+ },
833
+ {
834
+ "epoch": 8.18,
835
+ "learning_rate": 5.4966021955044426e-05,
836
+ "loss": 0.2835,
837
+ "step": 63000
838
+ },
839
+ {
840
+ "epoch": 8.24,
841
+ "learning_rate": 5.3005750130684786e-05,
842
+ "loss": 0.2904,
843
+ "step": 63500
844
+ },
845
+ {
846
+ "epoch": 8.31,
847
+ "learning_rate": 5.104547830632514e-05,
848
+ "loss": 0.289,
849
+ "step": 64000
850
+ },
851
+ {
852
+ "epoch": 8.37,
853
+ "learning_rate": 4.908520648196549e-05,
854
+ "loss": 0.2905,
855
+ "step": 64500
856
+ },
857
+ {
858
+ "epoch": 8.44,
859
+ "learning_rate": 4.7124934657605846e-05,
860
+ "loss": 0.288,
861
+ "step": 65000
862
+ },
863
+ {
864
+ "epoch": 8.5,
865
+ "learning_rate": 4.516466283324621e-05,
866
+ "loss": 0.2813,
867
+ "step": 65500
868
+ },
869
+ {
870
+ "epoch": 8.57,
871
+ "learning_rate": 4.3204391008886566e-05,
872
+ "loss": 0.2831,
873
+ "step": 66000
874
+ },
875
+ {
876
+ "epoch": 8.63,
877
+ "learning_rate": 4.124411918452692e-05,
878
+ "loss": 0.2914,
879
+ "step": 66500
880
+ },
881
+ {
882
+ "epoch": 8.7,
883
+ "learning_rate": 3.928384736016727e-05,
884
+ "loss": 0.2807,
885
+ "step": 67000
886
+ },
887
+ {
888
+ "epoch": 8.76,
889
+ "learning_rate": 3.7323575535807626e-05,
890
+ "loss": 0.2833,
891
+ "step": 67500
892
+ },
893
+ {
894
+ "epoch": 8.83,
895
+ "learning_rate": 3.5363303711447986e-05,
896
+ "loss": 0.2825,
897
+ "step": 68000
898
+ },
899
+ {
900
+ "epoch": 8.89,
901
+ "learning_rate": 3.340303188708834e-05,
902
+ "loss": 0.2824,
903
+ "step": 68500
904
+ },
905
+ {
906
+ "epoch": 8.96,
907
+ "learning_rate": 3.14427600627287e-05,
908
+ "loss": 0.281,
909
+ "step": 69000
910
+ },
911
+ {
912
+ "epoch": 9.0,
913
+ "eval_loss": 0.3067900538444519,
914
+ "eval_runtime": 12.6252,
915
+ "eval_samples_per_second": 56.316,
916
+ "eval_steps_per_second": 7.049,
917
+ "eval_wer": 0.17419354838709677,
918
+ "step": 69318
919
+ },
920
+ {
921
+ "epoch": 9.02,
922
+ "learning_rate": 2.948248823836905e-05,
923
+ "loss": 0.2725,
924
+ "step": 69500
925
+ },
926
+ {
927
+ "epoch": 9.09,
928
+ "learning_rate": 2.7522216414009406e-05,
929
+ "loss": 0.2696,
930
+ "step": 70000
931
+ },
932
+ {
933
+ "epoch": 9.15,
934
+ "learning_rate": 2.5561944589649766e-05,
935
+ "loss": 0.2713,
936
+ "step": 70500
937
+ },
938
+ {
939
+ "epoch": 9.22,
940
+ "learning_rate": 2.360167276529012e-05,
941
+ "loss": 0.2634,
942
+ "step": 71000
943
+ },
944
+ {
945
+ "epoch": 9.28,
946
+ "learning_rate": 2.1641400940930475e-05,
947
+ "loss": 0.2635,
948
+ "step": 71500
949
+ },
950
+ {
951
+ "epoch": 9.35,
952
+ "learning_rate": 1.968112911657083e-05,
953
+ "loss": 0.2658,
954
+ "step": 72000
955
+ },
956
+ {
957
+ "epoch": 9.41,
958
+ "learning_rate": 1.7720857292211185e-05,
959
+ "loss": 0.2648,
960
+ "step": 72500
961
+ },
962
+ {
963
+ "epoch": 9.48,
964
+ "learning_rate": 1.576058546785154e-05,
965
+ "loss": 0.273,
966
+ "step": 73000
967
+ },
968
+ {
969
+ "epoch": 9.54,
970
+ "learning_rate": 1.3800313643491895e-05,
971
+ "loss": 0.2601,
972
+ "step": 73500
973
+ },
974
+ {
975
+ "epoch": 9.61,
976
+ "learning_rate": 1.1840041819132252e-05,
977
+ "loss": 0.2598,
978
+ "step": 74000
979
+ },
980
+ {
981
+ "epoch": 9.67,
982
+ "learning_rate": 9.879769994772608e-06,
983
+ "loss": 0.2698,
984
+ "step": 74500
985
+ },
986
+ {
987
+ "epoch": 9.74,
988
+ "learning_rate": 7.919498170412963e-06,
989
+ "loss": 0.2574,
990
+ "step": 75000
991
+ },
992
+ {
993
+ "epoch": 9.8,
994
+ "learning_rate": 5.959226346053318e-06,
995
+ "loss": 0.2646,
996
+ "step": 75500
997
+ },
998
+ {
999
+ "epoch": 9.87,
1000
+ "learning_rate": 3.998954521693675e-06,
1001
+ "loss": 0.2584,
1002
+ "step": 76000
1003
+ },
1004
+ {
1005
+ "epoch": 9.93,
1006
+ "learning_rate": 2.0386826973340303e-06,
1007
+ "loss": 0.2623,
1008
+ "step": 76500
1009
+ },
1010
+ {
1011
+ "epoch": 10.0,
1012
+ "learning_rate": 7.841087297438577e-08,
1013
+ "loss": 0.2584,
1014
+ "step": 77000
1015
+ },
1016
+ {
1017
+ "epoch": 10.0,
1018
+ "eval_loss": 0.3040466010570526,
1019
+ "eval_runtime": 12.4236,
1020
+ "eval_samples_per_second": 57.23,
1021
+ "eval_steps_per_second": 7.164,
1022
+ "eval_wer": 0.17129690585911783,
1023
+ "step": 77020
1024
+ },
1025
+ {
1026
+ "epoch": 10.0,
1027
+ "step": 77020,
1028
+ "total_flos": 1.409382714117005e+20,
1029
+ "train_loss": 0.4655882824058503,
1030
+ "train_runtime": 41869.3879,
1031
+ "train_samples_per_second": 29.431,
1032
+ "train_steps_per_second": 1.84
1033
+ }
1034
+ ],
1035
+ "logging_steps": 500,
1036
+ "max_steps": 77020,
1037
+ "num_train_epochs": 10,
1038
+ "save_steps": 500,
1039
+ "total_flos": 1.409382714117005e+20,
1040
+ "trial_name": null,
1041
+ "trial_params": null
1042
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:57454b8ef61e73ee41e0231ebd404fc4c44e52d4e9fbcfd019aea1cdd0b25721
3
+ size 4536
vocab.json ADDED
@@ -0,0 +1,103 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "\"": 1,
3
+ "#": 2,
4
+ "$": 3,
5
+ "&": 4,
6
+ "0": 5,
7
+ "1": 6,
8
+ "2": 7,
9
+ "3": 8,
10
+ "4": 9,
11
+ "5": 10,
12
+ "6": 11,
13
+ "7": 12,
14
+ "8": 13,
15
+ "9": 14,
16
+ ";": 15,
17
+ "[PAD]": 100,
18
+ "[UNK]": 99,
19
+ "\\": 16,
20
+ "_": 17,
21
+ "`": 18,
22
+ "a": 19,
23
+ "b": 20,
24
+ "c": 21,
25
+ "d": 22,
26
+ "e": 23,
27
+ "f": 24,
28
+ "g": 25,
29
+ "h": 26,
30
+ "i": 27,
31
+ "j": 28,
32
+ "k": 29,
33
+ "l": 30,
34
+ "m": 31,
35
+ "n": 32,
36
+ "o": 33,
37
+ "p": 34,
38
+ "q": 35,
39
+ "r": 36,
40
+ "s": 37,
41
+ "t": 38,
42
+ "u": 39,
43
+ "v": 40,
44
+ "w": 41,
45
+ "x": 42,
46
+ "y": 43,
47
+ "z": 44,
48
+ "|": 0,
49
+ "à": 45,
50
+ "á": 46,
51
+ "ç": 47,
52
+ "è": 48,
53
+ "é": 49,
54
+ "ê": 50,
55
+ "ì": 51,
56
+ "í": 52,
57
+ "ñ": 53,
58
+ "ó": 54,
59
+ "ö": 55,
60
+ "ø": 56,
61
+ "ú": 57,
62
+ "ī": 58,
63
+ "ō": 59,
64
+ "ơ": 60,
65
+ "ư": 61,
66
+ "ǒ": 62,
67
+ "ǔ": 63,
68
+ "˚": 64,
69
+ "а": 65,
70
+ "в": 66,
71
+ "к": 67,
72
+ "о": 68,
73
+ "с": 69,
74
+ "т": 70,
75
+ "у": 71,
76
+ "ы": 72,
77
+ "э": 73,
78
+ "ờ": 74,
79
+ "​": 75,
80
+ "–": 76,
81
+ "—": 77,
82
+ "’": 78,
83
+ "…": 79,
84
+ "⁄": 80,
85
+ "℃": 81,
86
+ "℉": 82,
87
+ "≈": 83,
88
+ "♪": 84,
89
+ "ㄲ": 85,
90
+ "ㄴ": 86,
91
+ "ㅠ": 87,
92
+ "간": 88,
93
+ "구": 89,
94
+ "님": 90,
95
+ "상": 91,
96
+ "설": 92,
97
+ "아": 93,
98
+ "왜": 94,
99
+ "음": 95,
100
+ "이": 96,
101
+ "절": 97,
102
+ "정": 98
103
+ }