ivangtorre commited on
Commit
d35ad98
1 Parent(s): cd3386c

Add model files

Browse files

Files changed (5) hide show
  1. README.md +3 -3
  2. pytorch_model.bin +1 -1
  3. trainer_state.json +12 -12
  4. training_args.bin +1 -1
  5. vocab.json +1 -1
README.md CHANGED
@@ -21,7 +21,7 @@ model-index:
21
  metrics:
22
  - name: Test WER
23
  type: wer
24
- value:
25
  ---
26
 
27
  # Wav2Vec2-Large-XLSR-53-euskera
@@ -84,7 +84,7 @@ processor = Wav2Vec2Processor.from_pretrained("ivangtorre/wav2vec2-large-xlsr-53
84
  model = Wav2Vec2ForCTC.from_pretrained("ivangtorre/wav2vec2-large-xlsr-53-basque")
85
  model.to("cuda")
86
 
87
- chars_to_ignore_regex = 'default=[",", "?", ".", "!", "-", ";", ":", '""', "%", "'", '"', "�"]'
88
  resampler = torchaudio.transforms.Resample(48_000, 16_000)
89
 
90
  # Preprocessing the datasets.
@@ -114,7 +114,7 @@ result = test_dataset.map(evaluate, batched=True, batch_size=8)
114
  print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))
115
  ```
116
 
117
- **Test Result**:
118
 
119
 
120
  ## Training
21
  metrics:
22
  - name: Test WER
23
  type: wer
24
+ value: 44.48
25
  ---
26
 
27
  # Wav2Vec2-Large-XLSR-53-euskera
84
  model = Wav2Vec2ForCTC.from_pretrained("ivangtorre/wav2vec2-large-xlsr-53-basque")
85
  model.to("cuda")
86
 
87
+ chars_to_ignore_regex = '[",", "?", ".", "!", "-", ";", ":", '""', "%", "'", '"', "�"]'
88
  resampler = torchaudio.transforms.Resample(48_000, 16_000)
89
 
90
  # Preprocessing the datasets.
114
  print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))
115
  ```
116
 
117
+ **Test Result**: 44.48 %
118
 
119
 
120
  ## Training
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2e816e78ee7fd0e498f93f933dd0b18034841e0ba38042639fceb7d006be4270
3
  size 1262055847
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:509588dea56f1c5e67678b30228db8d4070585445da2cfab49c13ee8fc134dc5
3
  size 1262055847
trainer_state.json CHANGED
@@ -1,30 +1,30 @@
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
- "epoch": 1.1845425867507886,
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": 1.18,
12
- "learning_rate": 0.0003470916966876807,
13
- "loss": 1.587,
14
  "step": 500
15
  },
16
  {
17
- "epoch": 1.18,
18
- "eval_loss": 0.2940497398376465,
19
- "eval_runtime": 783.4997,
20
- "eval_samples_per_second": 6.601,
21
- "eval_wer": 0.5440472766257878,
22
  "step": 500
23
  }
24
  ],
25
- "max_steps": 2110,
26
- "num_train_epochs": 5,
27
- "total_flos": 2.4510852111832704e+18,
28
  "trial_name": null,
29
  "trial_params": null
30
  }
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
+ "epoch": 3.785894206549118,
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": 3.79,
12
+ "learning_rate": 0.00039436231366470835,
13
+ "loss": 1.5121,
14
  "step": 500
15
  },
16
  {
17
+ "epoch": 3.79,
18
+ "eval_loss": 0.186998650431633,
19
+ "eval_runtime": 714.138,
20
+ "eval_samples_per_second": 7.242,
21
+ "eval_wer": 0.39502654080485317,
22
  "step": 500
23
  }
24
  ],
25
+ "max_steps": 6600,
26
+ "num_train_epochs": 50,
27
+ "total_flos": 7.843188479774457e+18,
28
  "trial_name": null,
29
  "trial_params": null
30
  }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:474caa22f1c7f0c4e89880f0d02b07015187ad5c2b1a354e1f8340ea7bbd6181
3
  size 2287
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:38479a919eda7dd28d3ae5f9cab083881c8741ef35462dbc9ab73b870d8c6d48
3
  size 2287
vocab.json CHANGED
@@ -1 +1 @@
1
- {"r": 0, "w": 1, "e": 2, "a": 3, "b": 4, "p": 5, "m": 7, "x": 8, "q": 9, "d": 10, "g": 11, "h": 12, "f": 13, "\u00ed": 14, "v": 15, "c": 16, "\u00f1": 17, "i": 18, "t": 19, "y": 20, "n": 21, "z": 22, "o": 23, "l": 24, "s": 25, "k": 26, "j": 27, "u": 28, "|": 6, "[UNK]": 29, "[PAD]": 30}
1
+ {"m": 0, "d": 1, "c": 2, "l": 3, "b": 4, "i": 5, "y": 6, "w": 7, "k": 8, "n": 10, "\u00f1": 11, "o": 12, "a": 13, "x": 14, "z": 15, "u": 16, "r": 17, "q": 18, "h": 19, "t": 20, "g": 21, "p": 22, "v": 23, "e": 24, "j": 25, "f": 26, "s": 27, "\u00ed": 28, "|": 9, "[UNK]": 29, "[PAD]": 30}