Hugging Face's logo

marinone94
/
xls-r-300m-sv-robust

Automatic Speech Recognition
Transformers
PyTorch
Swedish
wav2vec2
mozilla-foundation/common_voice_9_0
generated_from_trainer
Eval Results
Inference Endpoints
Model card Files Community
1
marinone94 commited on
Commit
990bfc1
•
1 Parent(s): d84f4ed

Training in progress, step 500

Browse files
Files changed (24) hide show
  1. .ipynb_checkpoints/run-checkpoint.sh +1 -1
  2. .ipynb_checkpoints/run-dummy-sv-gpu-checkpoint.sh +0 -34
  3. added_tokens.json +1 -1
  4. all_results.json +0 -14
  5. checkpoint-10/trainer_state.json +0 -15
  6. {checkpoint-10 → checkpoint-500}/config.json +40 -29
  7. {checkpoint-10 → checkpoint-500}/optimizer.pt +2 -2
  8. {checkpoint-10 → checkpoint-500}/preprocessor_config.json +2 -2
  9. {checkpoint-10 → checkpoint-500}/pytorch_model.bin +2 -2
  10. {checkpoint-10 → checkpoint-500}/rng_state.pth +2 -2
  11. {checkpoint-10 → checkpoint-500}/scaler.pt +1 -1
  12. {checkpoint-10 → checkpoint-500}/scheduler.pt +1 -1
  13. checkpoint-500/trainer_state.json +55 -0
  14. {checkpoint-10 → checkpoint-500}/training_args.bin +1 -1
  15. config.json +40 -29
  16. eval_results.json +0 -9
  17. preprocessor_config.json +2 -2
  18. pytorch_model.bin +2 -2
  19. run-dummy-sv-gpu.sh +0 -34
  20. run.sh +1 -1
  21. train_results.json +0 -8
  22. trainer_state.json +0 -25
  23. training_args.bin +1 -1
  24. vocab.json +1 -1
.ipynb_checkpoints/run-checkpoint.sh CHANGED
@@ -4,7 +4,7 @@ python run_speech_recognition_ctc.py \
4
  --dataset_config_name="sv-SE" \
5
  --output_dir="./" \
6
  --overwrite_output_dir \
7
- --num_train_epochs="2" \
8
  --per_device_train_batch_size="8" \
9
  --per_device_eval_batch_size="8" \
10
  --gradient_accumulation_steps="4" \
 
4
  --dataset_config_name="sv-SE" \
5
  --output_dir="./" \
6
  --overwrite_output_dir \
7
+ --num_train_epochs="5" \
8
  --per_device_train_batch_size="8" \
9
  --per_device_eval_batch_size="8" \
10
  --gradient_accumulation_steps="4" \
.ipynb_checkpoints/run-dummy-sv-gpu-checkpoint.sh DELETED
@@ -1,34 +0,0 @@
1
- python run_speech_recognition_ctc.py \
2
- --dataset_name="mozilla-foundation/common_voice_7_0" \
3
- --model_name_or_path="hf-test/xls-r-dummy" \
4
- --dataset_config_name="ab" \
5
- --output_dir="./" \
6
- --overwrite_output_dir \
7
- --max_steps="10" \
8
- --per_device_train_batch_size="8" \
9
- --per_device_eval_batch_size="8" \
10
- --gradient_accumulation_steps="4" \
11
- --learning_rate="7.5e-3" \
12
- --warmup_steps="2000" \
13
- --length_column_name="input_length" \
14
- --evaluation_strategy="steps" \
15
- --text_column_name="sentence" \
16
- --chars_to_ignore , ? . ! \- \; \: \" “ % ‘ ” � — ’ … – \
17
- --save_steps="5" \
18
- --eval_steps="5" \
19
- --logging_steps="1" \
20
- --layerdrop="0.0" \
21
- --activation_dropout="0.1" \
22
- --save_total_limit="3" \
23
- --freeze_feature_encoder \
24
- --feat_proj_dropout="0.0" \
25
- --mask_time_prob="0.75" \
26
- --mask_time_length="10" \
27
- --mask_feature_prob="0.25" \
28
- --mask_feature_length="64" \
29
- --gradient_checkpointing \
30
- --use_auth_token \
31
- --fp16 \
32
- --group_by_length \
33
- --do_train --do_eval \
34
- --push_to_hub
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
added_tokens.json CHANGED
@@ -1 +1 @@
1
- {"<s>": 51, "</s>": 52}
 
1
+ {"<s>": 35, "</s>": 36}
all_results.json DELETED
@@ -1,14 +0,0 @@
1
- {
2
- "epoch": 0.03,
3
- "eval_loss": 156.8788604736328,
4
- "eval_runtime": 10.8705,
5
- "eval_samples": 301,
6
- "eval_samples_per_second": 27.69,
7
- "eval_steps_per_second": 3.496,
8
- "eval_wer": 1.3456221198156681,
9
- "train_loss": 94.46571044921875,
10
- "train_runtime": 13.57,
11
- "train_samples": 704,
12
- "train_samples_per_second": 1.474,
13
- "train_steps_per_second": 0.737
14
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-10/trainer_state.json DELETED
@@ -1,15 +0,0 @@
1
- {
2
- "best_metric": null,
3
- "best_model_checkpoint": null,
4
- "epoch": 0.028409090909090908,
5
- "global_step": 10,
6
- "is_hyper_param_search": false,
7
- "is_local_process_zero": true,
8
- "is_world_process_zero": true,
9
- "log_history": [],
10
- "max_steps": 10,
11
- "num_train_epochs": 1,
12
- "total_flos": 508182128640.0,
13
- "trial_name": null,
14
- "trial_params": null
15
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
{checkpoint-10 → checkpoint-500}/config.json RENAMED
@@ -1,6 +1,6 @@
1
  {
2
- "_name_or_path": "hf-test/xls-r-dummy",
3
- "activation_dropout": 0.0,
4
  "adapter_kernel_size": 3,
5
  "adapter_stride": 2,
6
  "add_adapter": false,
@@ -11,23 +11,35 @@
11
  "attention_dropout": 0.0,
12
  "bos_token_id": 1,
13
  "classifier_proj_size": 256,
14
- "codevector_dim": 256,
15
  "contrastive_logits_temperature": 0.1,
16
- "conv_bias": false,
17
  "conv_dim": [
18
- 32,
19
- 32,
20
- 32
 
 
 
 
21
  ],
22
  "conv_kernel": [
23
- 8,
24
- 8,
25
- 8
 
 
 
 
26
  ],
27
  "conv_stride": [
28
- 4,
29
- 4,
30
- 4
 
 
 
 
31
  ],
32
  "ctc_loss_reduction": "mean",
33
  "ctc_zero_infinity": false,
@@ -42,31 +54,30 @@
42
  "final_dropout": 0.0,
43
  "hidden_act": "gelu",
44
  "hidden_dropout": 0.0,
45
- "hidden_dropout_prob": 0.1,
46
- "hidden_size": 16,
47
  "initializer_range": 0.02,
48
- "intermediate_size": 20,
49
  "layer_norm_eps": 1e-05,
50
  "layerdrop": 0.0,
51
- "mask_feature_length": 10,
52
  "mask_feature_min_masks": 0,
53
- "mask_feature_prob": 0.0,
54
  "mask_time_length": 10,
55
  "mask_time_min_masks": 2,
56
- "mask_time_prob": 0.05,
57
  "model_type": "wav2vec2",
58
  "num_adapter_layers": 3,
59
- "num_attention_heads": 2,
60
  "num_codevector_groups": 2,
61
  "num_codevectors_per_group": 320,
62
- "num_conv_pos_embedding_groups": 2,
63
- "num_conv_pos_embeddings": 16,
64
- "num_feat_extract_layers": 3,
65
- "num_hidden_layers": 4,
66
- "num_negatives": 10,
67
- "output_hidden_size": 16,
68
- "pad_token_id": 50,
69
- "proj_codevector_dim": 256,
70
  "tdnn_dilation": [
71
  1,
72
  2,
@@ -91,6 +102,6 @@
91
  "torch_dtype": "float32",
92
  "transformers_version": "4.16.0.dev0",
93
  "use_weighted_layer_sum": false,
94
- "vocab_size": 53,
95
  "xvector_output_dim": 512
96
  }
 
1
  {
2
+ "_name_or_path": "facebook/wav2vec2-xls-r-300m",
3
+ "activation_dropout": 0.1,
4
  "adapter_kernel_size": 3,
5
  "adapter_stride": 2,
6
  "add_adapter": false,
 
11
  "attention_dropout": 0.0,
12
  "bos_token_id": 1,
13
  "classifier_proj_size": 256,
14
+ "codevector_dim": 768,
15
  "contrastive_logits_temperature": 0.1,
16
+ "conv_bias": true,
17
  "conv_dim": [
18
+ 512,
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512
25
  ],
26
  "conv_kernel": [
27
+ 10,
28
+ 3,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 2,
33
+ 2
34
  ],
35
  "conv_stride": [
36
+ 5,
37
+ 2,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2
43
  ],
44
  "ctc_loss_reduction": "mean",
45
  "ctc_zero_infinity": false,
 
54
  "final_dropout": 0.0,
55
  "hidden_act": "gelu",
56
  "hidden_dropout": 0.0,
57
+ "hidden_size": 1024,
 
58
  "initializer_range": 0.02,
59
+ "intermediate_size": 4096,
60
  "layer_norm_eps": 1e-05,
61
  "layerdrop": 0.0,
62
+ "mask_feature_length": 64,
63
  "mask_feature_min_masks": 0,
64
+ "mask_feature_prob": 0.25,
65
  "mask_time_length": 10,
66
  "mask_time_min_masks": 2,
67
+ "mask_time_prob": 0.75,
68
  "model_type": "wav2vec2",
69
  "num_adapter_layers": 3,
70
+ "num_attention_heads": 16,
71
  "num_codevector_groups": 2,
72
  "num_codevectors_per_group": 320,
73
+ "num_conv_pos_embedding_groups": 16,
74
+ "num_conv_pos_embeddings": 128,
75
+ "num_feat_extract_layers": 7,
76
+ "num_hidden_layers": 24,
77
+ "num_negatives": 100,
78
+ "output_hidden_size": 1024,
79
+ "pad_token_id": 34,
80
+ "proj_codevector_dim": 768,
81
  "tdnn_dilation": [
82
  1,
83
  2,
 
102
  "torch_dtype": "float32",
103
  "transformers_version": "4.16.0.dev0",
104
  "use_weighted_layer_sum": false,
105
+ "vocab_size": 37,
106
  "xvector_output_dim": 512
107
  }
{checkpoint-10 → checkpoint-500}/optimizer.pt RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:72957da22b86d76dde29af1a0b07cc32ca91e90a52ae948848397850f97f112c
3
- size 130385
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dab21b247ec51e4dee7afa84fedd77b13635e30b256cb9caa89e1dfc094123c9
3
+ size 2490362385
{checkpoint-10 → checkpoint-500}/preprocessor_config.json RENAMED
@@ -3,7 +3,7 @@
3
  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
  "feature_size": 1,
5
  "padding_side": "right",
6
- "padding_value": 0.0,
7
- "return_attention_mask": false,
8
  "sampling_rate": 16000
9
  }
 
3
  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
  "feature_size": 1,
5
  "padding_side": "right",
6
+ "padding_value": 0,
7
+ "return_attention_mask": true,
8
  "sampling_rate": 16000
9
  }
{checkpoint-10 → checkpoint-500}/pytorch_model.bin RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:bc977003f835439fa30e8deda90e702230cee93cf93cdfa254b2d5d7ab36a82c
3
- size 143910
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:401051ca4bb075cafe2856a328c8f6a5e3e87fa0c13df1de35ee6e7873004e73
3
+ size 1262075377
{checkpoint-10 → checkpoint-500}/rng_state.pth RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a3d9c16ce801cec7b410e0af5e5a6cda8ecea7b42faa636c085ace1ffb6e20ad
3
- size 14567
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32b82ac75b07e68dc3bc90e76f55a339f64dce724d87a9ae3c69ee46df441867
3
+ size 14503
{checkpoint-10 → checkpoint-500}/scaler.pt RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ead3ef1a5eb5dac2a9c2364b993bfd84efcde1506028e7ce7af6b536ce44f59c
3
  size 559
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c8bceadebe118ff459b01a775a8e9b38a6b8302c162d022f78d3646163e6486
3
  size 559
{checkpoint-10 → checkpoint-500}/scheduler.pt RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:59216c25b5b10e66c1be135b3804bb7b79f89ad700aea4e78300a0b0e97a9507
3
  size 623
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3757aee7e707dcd124ce47058f7058fcb392f4c693c82774d9b2b5dcccf35b49
3
  size 623
checkpoint-500/trainer_state.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.4525018129079044,
5
+ "global_step": 500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.29,
12
+ "learning_rate": 7.151470588235293e-05,
13
+ "loss": 5.1135,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 0.58,
18
+ "learning_rate": 6.710294117647058e-05,
19
+ "loss": 3.0957,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 0.87,
24
+ "learning_rate": 6.269117647058824e-05,
25
+ "loss": 3.0078,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 1.16,
30
+ "learning_rate": 5.827941176470588e-05,
31
+ "loss": 2.9785,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 1.45,
36
+ "learning_rate": 5.3867647058823525e-05,
37
+ "loss": 2.9069,
38
+ "step": 500
39
+ },
40
+ {
41
+ "epoch": 1.45,
42
+ "eval_loss": 2.9046826362609863,
43
+ "eval_runtime": 128.7688,
44
+ "eval_samples_per_second": 35.878,
45
+ "eval_steps_per_second": 4.489,
46
+ "eval_wer": 1.0,
47
+ "step": 500
48
+ }
49
+ ],
50
+ "max_steps": 1720,
51
+ "num_train_epochs": 5,
52
+ "total_flos": 1.4827194756605722e+18,
53
+ "trial_name": null,
54
+ "trial_params": null
55
+ }
{checkpoint-10 → checkpoint-500}/training_args.bin RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:13f03d8c8ae9bd8496ddb1b6ef8d282d15b54e9aee2f515f000a57e9ac24d7c2
3
  size 2991
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1208f22bb7e06e1e9a51692db0520fbddfc3640941d51dfe45ba3188ada2ecbf
3
  size 2991
config.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
- "_name_or_path": "hf-test/xls-r-dummy",
3
- "activation_dropout": 0.0,
4
  "adapter_kernel_size": 3,
5
  "adapter_stride": 2,
6
  "add_adapter": false,
@@ -11,23 +11,35 @@
11
  "attention_dropout": 0.0,
12
  "bos_token_id": 1,
13
  "classifier_proj_size": 256,
14
- "codevector_dim": 256,
15
  "contrastive_logits_temperature": 0.1,
16
- "conv_bias": false,
17
  "conv_dim": [
18
- 32,
19
- 32,
20
- 32
 
 
 
 
21
  ],
22
  "conv_kernel": [
23
- 8,
24
- 8,
25
- 8
 
 
 
 
26
  ],
27
  "conv_stride": [
28
- 4,
29
- 4,
30
- 4
 
 
 
 
31
  ],
32
  "ctc_loss_reduction": "mean",
33
  "ctc_zero_infinity": false,
@@ -42,31 +54,30 @@
42
  "final_dropout": 0.0,
43
  "hidden_act": "gelu",
44
  "hidden_dropout": 0.0,
45
- "hidden_dropout_prob": 0.1,
46
- "hidden_size": 16,
47
  "initializer_range": 0.02,
48
- "intermediate_size": 20,
49
  "layer_norm_eps": 1e-05,
50
  "layerdrop": 0.0,
51
- "mask_feature_length": 10,
52
  "mask_feature_min_masks": 0,
53
- "mask_feature_prob": 0.0,
54
  "mask_time_length": 10,
55
  "mask_time_min_masks": 2,
56
- "mask_time_prob": 0.05,
57
  "model_type": "wav2vec2",
58
  "num_adapter_layers": 3,
59
- "num_attention_heads": 2,
60
  "num_codevector_groups": 2,
61
  "num_codevectors_per_group": 320,
62
- "num_conv_pos_embedding_groups": 2,
63
- "num_conv_pos_embeddings": 16,
64
- "num_feat_extract_layers": 3,
65
- "num_hidden_layers": 4,
66
- "num_negatives": 10,
67
- "output_hidden_size": 16,
68
- "pad_token_id": 50,
69
- "proj_codevector_dim": 256,
70
  "tdnn_dilation": [
71
  1,
72
  2,
@@ -91,6 +102,6 @@
91
  "torch_dtype": "float32",
92
  "transformers_version": "4.16.0.dev0",
93
  "use_weighted_layer_sum": false,
94
- "vocab_size": 53,
95
  "xvector_output_dim": 512
96
  }
 
1
  {
2
+ "_name_or_path": "facebook/wav2vec2-xls-r-300m",
3
+ "activation_dropout": 0.1,
4
  "adapter_kernel_size": 3,
5
  "adapter_stride": 2,
6
  "add_adapter": false,
 
11
  "attention_dropout": 0.0,
12
  "bos_token_id": 1,
13
  "classifier_proj_size": 256,
14
+ "codevector_dim": 768,
15
  "contrastive_logits_temperature": 0.1,
16
+ "conv_bias": true,
17
  "conv_dim": [
18
+ 512,
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512
25
  ],
26
  "conv_kernel": [
27
+ 10,
28
+ 3,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 2,
33
+ 2
34
  ],
35
  "conv_stride": [
36
+ 5,
37
+ 2,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2
43
  ],
44
  "ctc_loss_reduction": "mean",
45
  "ctc_zero_infinity": false,
 
54
  "final_dropout": 0.0,
55
  "hidden_act": "gelu",
56
  "hidden_dropout": 0.0,
57
+ "hidden_size": 1024,
 
58
  "initializer_range": 0.02,
59
+ "intermediate_size": 4096,
60
  "layer_norm_eps": 1e-05,
61
  "layerdrop": 0.0,
62
+ "mask_feature_length": 64,
63
  "mask_feature_min_masks": 0,
64
+ "mask_feature_prob": 0.25,
65
  "mask_time_length": 10,
66
  "mask_time_min_masks": 2,
67
+ "mask_time_prob": 0.75,
68
  "model_type": "wav2vec2",
69
  "num_adapter_layers": 3,
70
+ "num_attention_heads": 16,
71
  "num_codevector_groups": 2,
72
  "num_codevectors_per_group": 320,
73
+ "num_conv_pos_embedding_groups": 16,
74
+ "num_conv_pos_embeddings": 128,
75
+ "num_feat_extract_layers": 7,
76
+ "num_hidden_layers": 24,
77
+ "num_negatives": 100,
78
+ "output_hidden_size": 1024,
79
+ "pad_token_id": 34,
80
+ "proj_codevector_dim": 768,
81
  "tdnn_dilation": [
82
  1,
83
  2,
 
102
  "torch_dtype": "float32",
103
  "transformers_version": "4.16.0.dev0",
104
  "use_weighted_layer_sum": false,
105
+ "vocab_size": 37,
106
  "xvector_output_dim": 512
107
  }
eval_results.json DELETED
@@ -1,9 +0,0 @@
1
- {
2
- "epoch": 0.03,
3
- "eval_loss": 156.8788604736328,
4
- "eval_runtime": 10.8705,
5
- "eval_samples": 301,
6
- "eval_samples_per_second": 27.69,
7
- "eval_steps_per_second": 3.496,
8
- "eval_wer": 1.3456221198156681
9
- }
 
 
 
 
 
 
 
 
 
 
preprocessor_config.json CHANGED
@@ -3,7 +3,7 @@
3
  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
  "feature_size": 1,
5
  "padding_side": "right",
6
- "padding_value": 0.0,
7
- "return_attention_mask": false,
8
  "sampling_rate": 16000
9
  }
 
3
  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
  "feature_size": 1,
5
  "padding_side": "right",
6
+ "padding_value": 0,
7
+ "return_attention_mask": true,
8
  "sampling_rate": 16000
9
  }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:bc977003f835439fa30e8deda90e702230cee93cf93cdfa254b2d5d7ab36a82c
3
- size 143910
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:401051ca4bb075cafe2856a328c8f6a5e3e87fa0c13df1de35ee6e7873004e73
3
+ size 1262075377
run-dummy-sv-gpu.sh DELETED
@@ -1,34 +0,0 @@
1
- python run_speech_recognition_ctc.py \
2
- --dataset_name="mozilla-foundation/common_voice_7_0" \
3
- --model_name_or_path="hf-test/xls-r-dummy" \
4
- --dataset_config_name="ab" \
5
- --output_dir="./" \
6
- --overwrite_output_dir \
7
- --max_steps="10" \
8
- --per_device_train_batch_size="8" \
9
- --per_device_eval_batch_size="8" \
10
- --gradient_accumulation_steps="4" \
11
- --learning_rate="7.5e-3" \
12
- --warmup_steps="2000" \
13
- --length_column_name="input_length" \
14
- --evaluation_strategy="steps" \
15
- --text_column_name="sentence" \
16
- --chars_to_ignore , ? . ! \- \; \: \" “ % ‘ ” � — ’ … – \
17
- --save_steps="5" \
18
- --eval_steps="5" \
19
- --logging_steps="1" \
20
- --layerdrop="0.0" \
21
- --activation_dropout="0.1" \
22
- --save_total_limit="3" \
23
- --freeze_feature_encoder \
24
- --feat_proj_dropout="0.0" \
25
- --mask_time_prob="0.75" \
26
- --mask_time_length="10" \
27
- --mask_feature_prob="0.25" \
28
- --mask_feature_length="64" \
29
- --gradient_checkpointing \
30
- --use_auth_token \
31
- --fp16 \
32
- --group_by_length \
33
- --do_train --do_eval \
34
- --push_to_hub
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
run.sh CHANGED
@@ -4,7 +4,7 @@ python run_speech_recognition_ctc.py \
4
  --dataset_config_name="sv-SE" \
5
  --output_dir="./" \
6
  --overwrite_output_dir \
7
- --num_train_epochs="2" \
8
  --per_device_train_batch_size="8" \
9
  --per_device_eval_batch_size="8" \
10
  --gradient_accumulation_steps="4" \
 
4
  --dataset_config_name="sv-SE" \
5
  --output_dir="./" \
6
  --overwrite_output_dir \
7
+ --num_train_epochs="5" \
8
  --per_device_train_batch_size="8" \
9
  --per_device_eval_batch_size="8" \
10
  --gradient_accumulation_steps="4" \
train_results.json DELETED
@@ -1,8 +0,0 @@
1
- {
2
- "epoch": 0.03,
3
- "train_loss": 94.46571044921875,
4
- "train_runtime": 13.57,
5
- "train_samples": 704,
6
- "train_samples_per_second": 1.474,
7
- "train_steps_per_second": 0.737
8
- }
 
 
 
 
 
 
 
 
 
trainer_state.json DELETED
@@ -1,25 +0,0 @@
1
- {
2
- "best_metric": null,
3
- "best_model_checkpoint": null,
4
- "epoch": 0.028409090909090908,
5
- "global_step": 10,
6
- "is_hyper_param_search": false,
7
- "is_local_process_zero": true,
8
- "is_world_process_zero": true,
9
- "log_history": [
10
- {
11
- "epoch": 0.03,
12
- "step": 10,
13
- "total_flos": 508182128640.0,
14
- "train_loss": 94.46571044921875,
15
- "train_runtime": 13.57,
16
- "train_samples_per_second": 1.474,
17
- "train_steps_per_second": 0.737
18
- }
19
- ],
20
- "max_steps": 10,
21
- "num_train_epochs": 1,
22
- "total_flos": 508182128640.0,
23
- "trial_name": null,
24
- "trial_params": null
25
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:13f03d8c8ae9bd8496ddb1b6ef8d282d15b54e9aee2f515f000a57e9ac24d7c2
3
  size 2991
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1208f22bb7e06e1e9a51692db0520fbddfc3640941d51dfe45ba3188ada2ecbf
3
  size 2991
vocab.json CHANGED
@@ -1 +1 @@
1
- {"!": 1, ",": 2, "-": 3, ".": 4, ":": 5, ";": 6, "?": 7, "а": 8, "б": 9, "в": 10, "г": 11, "д": 12, "е": 13, "ж": 14, "з": 15, "и": 16, "к": 17, "л": 18, "м": 19, "н": 20, "о": 21, "п": 22, "р": 23, "с": 24, "т": 25, "у": 26, "ф": 27, "х": 28, "ц": 29, "ч": 30, "ш": 31, "ы": 32, "ь": 33, "џ": 34, "қ": 35, "ҟ": 36, "ҩ": 37, "ҭ": 38, "ҳ": 39, "ҵ": 40, "ҷ": 41, "ҽ": 42, "ҿ": 43, "ә": 44, "ӡ": 45, "ӷ": 46, "ԥ": 47, "–": 48, "|": 0, "[UNK]": 49, "[PAD]": 50}
 
1
+ {"a": 1, "b": 2, "c": 3, "d": 4, "e": 5, "f": 6, "g": 7, "h": 8, "i": 9, "j": 10, "k": 11, "l": 12, "m": 13, "n": 14, "o": 15, "p": 16, "q": 17, "r": 18, "s": 19, "t": 20, "u": 21, "v": 22, "w": 23, "x": 24, "y": 25, "z": 26, "ä": 27, "å": 28, "é": 29, "ô": 30, "ö": 31, "ü": 32, "|": 0, "[UNK]": 33, "[PAD]": 34}