Hugging Face's logo

asr-africa
/
w2v-bert-2.0-CV_Fleurs-lg-5hrs-v4

Automatic Speech Recognition
Transformers
Safetensors
wav2vec2-bert
Generated from Trainer
Inference Endpoints
Model card Files Community
sulaimank commited on
Commit
4cfe315
·
verified ·
1 Parent(s): 4662df5

Training in progress, epoch 5, checkpoint

Browse files
Files changed (8) hide show
  1. checkpoint-1290/config.json +82 -0
  2. checkpoint-1290/model.safetensors +3 -0
  3. checkpoint-1290/optimizer.pt +3 -0
  4. checkpoint-1290/preprocessor_config.json +11 -0
  5. checkpoint-1290/rng_state.pth +3 -0
  6. checkpoint-1290/scheduler.pt +3 -0
  7. checkpoint-1290/trainer_state.json +127 -0
  8. checkpoint-1290/training_args.bin +3 -0
checkpoint-1290/config.json ADDED
@@ -0,0 +1,82 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/w2v-bert-2.0",
3
+ "activation_dropout": 0.0,
4
+ "adapter_act": "relu",
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": true,
8
+ "apply_spec_augment": false,
9
+ "architectures": [
10
+ "Wav2Vec2BertForCTC"
11
+ ],
12
+ "attention_dropout": 0.0,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 768,
15
+ "codevector_dim": 768,
16
+ "conformer_conv_dropout": 0.1,
17
+ "contrastive_logits_temperature": 0.1,
18
+ "conv_depthwise_kernel_size": 31,
19
+ "ctc_loss_reduction": "mean",
20
+ "ctc_zero_infinity": true,
21
+ "diversity_loss_weight": 0.1,
22
+ "eos_token_id": 2,
23
+ "feat_proj_dropout": 0.0,
24
+ "feat_quantizer_dropout": 0.0,
25
+ "feature_projection_input_dim": 160,
26
+ "final_dropout": 0.1,
27
+ "hidden_act": "swish",
28
+ "hidden_dropout": 0.0,
29
+ "hidden_size": 1024,
30
+ "initializer_range": 0.02,
31
+ "intermediate_size": 4096,
32
+ "layer_norm_eps": 1e-05,
33
+ "layerdrop": 0.0,
34
+ "left_max_position_embeddings": 64,
35
+ "mask_feature_length": 10,
36
+ "mask_feature_min_masks": 0,
37
+ "mask_feature_prob": 0.0,
38
+ "mask_time_length": 10,
39
+ "mask_time_min_masks": 2,
40
+ "mask_time_prob": 0.0,
41
+ "max_source_positions": 5000,
42
+ "model_type": "wav2vec2-bert",
43
+ "num_adapter_layers": 1,
44
+ "num_attention_heads": 16,
45
+ "num_codevector_groups": 2,
46
+ "num_codevectors_per_group": 320,
47
+ "num_hidden_layers": 24,
48
+ "num_negatives": 100,
49
+ "output_hidden_size": 1024,
50
+ "pad_token_id": 43,
51
+ "position_embeddings_type": "relative_key",
52
+ "proj_codevector_dim": 768,
53
+ "right_max_position_embeddings": 8,
54
+ "rotary_embedding_base": 10000,
55
+ "tdnn_dilation": [
56
+ 1,
57
+ 2,
58
+ 3,
59
+ 1,
60
+ 1
61
+ ],
62
+ "tdnn_dim": [
63
+ 512,
64
+ 512,
65
+ 512,
66
+ 512,
67
+ 1500
68
+ ],
69
+ "tdnn_kernel": [
70
+ 5,
71
+ 3,
72
+ 3,
73
+ 1,
74
+ 1
75
+ ],
76
+ "torch_dtype": "float32",
77
+ "transformers_version": "4.46.0",
78
+ "use_intermediate_ffn_before_adapter": false,
79
+ "use_weighted_layer_sum": false,
80
+ "vocab_size": 46,
81
+ "xvector_output_dim": 512
82
+ }
checkpoint-1290/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9fea90e8ad548ea6b2ae3a249fd42273d20eb6c3e78ab3ca067d0843b4e2ae4a
3
+ size 2423003160
checkpoint-1290/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8261c13a182596054efdb282aedab7e3f592da7e2e354b8875bf0e26e220348d
3
+ size 4846467050
checkpoint-1290/preprocessor_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "feature_extractor_type": "SeamlessM4TFeatureExtractor",
3
+ "feature_size": 80,
4
+ "num_mel_bins": 80,
5
+ "padding_side": "right",
6
+ "padding_value": 1,
7
+ "processor_class": "Wav2Vec2BertProcessor",
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000,
10
+ "stride": 2
11
+ }
checkpoint-1290/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ddd03285ae96055b810d9c7b78341a3b36ed0751fefd223255882a65d0023d6
3
+ size 14244
checkpoint-1290/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2360927524bffa7067ff133c29b68dde02206b6045f2b7acdace9d16ab1c7e99
3
+ size 1064
checkpoint-1290/trainer_state.json ADDED
@@ -0,0 +1,127 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.4769304327038011,
3
+ "best_model_checkpoint": "w2v-bert-2.0-CV_Fleurs-lg-5hrs-v4/checkpoint-1290",
4
+ "epoch": 5.0,
5
+ "eval_steps": 500,
6
+ "global_step": 1290,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 1.0,
13
+ "grad_norm": 11.246596336364746,
14
+ "learning_rate": 5.1200000000000004e-05,
15
+ "loss": 1.9959,
16
+ "step": 258
17
+ },
18
+ {
19
+ "epoch": 1.0,
20
+ "eval_cer": 0.13379445701894616,
21
+ "eval_loss": 0.5475470423698425,
22
+ "eval_runtime": 29.6601,
23
+ "eval_samples_per_second": 27.849,
24
+ "eval_steps_per_second": 6.979,
25
+ "eval_wer": 0.6078173559646187,
26
+ "step": 258
27
+ },
28
+ {
29
+ "epoch": 2.0,
30
+ "grad_norm": 1.1338785886764526,
31
+ "learning_rate": 9.994466403162056e-05,
32
+ "loss": 0.5447,
33
+ "step": 516
34
+ },
35
+ {
36
+ "epoch": 2.0,
37
+ "eval_cer": 0.1309612371199291,
38
+ "eval_loss": 0.5649892687797546,
39
+ "eval_runtime": 29.1014,
40
+ "eval_samples_per_second": 28.384,
41
+ "eval_steps_per_second": 7.113,
42
+ "eval_wer": 0.5978962467128854,
43
+ "step": 516
44
+ },
45
+ {
46
+ "epoch": 3.0,
47
+ "grad_norm": 5.585188388824463,
48
+ "learning_rate": 9.892490118577076e-05,
49
+ "loss": 0.4574,
50
+ "step": 774
51
+ },
52
+ {
53
+ "epoch": 3.0,
54
+ "eval_cer": 0.11218917678342487,
55
+ "eval_loss": 0.45441317558288574,
56
+ "eval_runtime": 29.5398,
57
+ "eval_samples_per_second": 27.962,
58
+ "eval_steps_per_second": 7.007,
59
+ "eval_wer": 0.5168539325842697,
60
+ "step": 774
61
+ },
62
+ {
63
+ "epoch": 4.0,
64
+ "grad_norm": 3.5034706592559814,
65
+ "learning_rate": 9.790513833992096e-05,
66
+ "loss": 0.3712,
67
+ "step": 1032
68
+ },
69
+ {
70
+ "epoch": 4.0,
71
+ "eval_cer": 0.10859621076623562,
72
+ "eval_loss": 0.4465671181678772,
73
+ "eval_runtime": 29.2068,
74
+ "eval_samples_per_second": 28.281,
75
+ "eval_steps_per_second": 7.087,
76
+ "eval_wer": 0.5049007889074827,
77
+ "step": 1032
78
+ },
79
+ {
80
+ "epoch": 5.0,
81
+ "grad_norm": 17.74321174621582,
82
+ "learning_rate": 9.688932806324111e-05,
83
+ "loss": 0.3187,
84
+ "step": 1290
85
+ },
86
+ {
87
+ "epoch": 5.0,
88
+ "eval_cer": 0.10036562781937036,
89
+ "eval_loss": 0.44379866123199463,
90
+ "eval_runtime": 29.231,
91
+ "eval_samples_per_second": 28.258,
92
+ "eval_steps_per_second": 7.082,
93
+ "eval_wer": 0.4769304327038011,
94
+ "step": 1290
95
+ }
96
+ ],
97
+ "logging_steps": 500,
98
+ "max_steps": 25800,
99
+ "num_input_tokens_seen": 0,
100
+ "num_train_epochs": 100,
101
+ "save_steps": 500,
102
+ "stateful_callbacks": {
103
+ "EarlyStoppingCallback": {
104
+ "args": {
105
+ "early_stopping_patience": 20,
106
+ "early_stopping_threshold": 0.001
107
+ },
108
+ "attributes": {
109
+ "early_stopping_patience_counter": 0
110
+ }
111
+ },
112
+ "TrainerControl": {
113
+ "args": {
114
+ "should_epoch_stop": false,
115
+ "should_evaluate": false,
116
+ "should_log": false,
117
+ "should_save": true,
118
+ "should_training_stop": false
119
+ },
120
+ "attributes": {}
121
+ }
122
+ },
123
+ "total_flos": 2.728933069480656e+18,
124
+ "train_batch_size": 8,
125
+ "trial_name": null,
126
+ "trial_params": null
127
+ }
checkpoint-1290/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:badc6886eb2995443a897620ab2c36decc3b66898e976799a420ad7fb69a195c
3
+ size 5304