AlexN commited on
Commit
8ba698b
1 Parent(s): 0a6506a

pt evaluation

Browse files
Files changed (31) hide show
  1. .ipynb_checkpoints/README-checkpoint.md +116 -0
  2. .ipynb_checkpoints/added_tokens-checkpoint.json +1 -0
  3. .ipynb_checkpoints/all_results-checkpoint.json +14 -0
  4. .ipynb_checkpoints/config-checkpoint.json +107 -0
  5. .ipynb_checkpoints/eval-checkpoint.py +140 -0
  6. .ipynb_checkpoints/eval_results-checkpoint.json +9 -0
  7. .ipynb_checkpoints/log_mozilla-foundation_common_voice_8_0_pt_test_predictions-checkpoint.txt +0 -0
  8. .ipynb_checkpoints/log_mozilla-foundation_common_voice_8_0_pt_test_targets-checkpoint.txt +0 -0
  9. .ipynb_checkpoints/log_speech-recognition-community-v2_dev_data_pt_validation_predictions-checkpoint.txt +0 -0
  10. .ipynb_checkpoints/log_speech-recognition-community-v2_dev_data_pt_validation_targets-checkpoint.txt +0 -0
  11. .ipynb_checkpoints/preprocessor_config-checkpoint.json +9 -0
  12. .ipynb_checkpoints/run-checkpoint.sh +1 -1
  13. .ipynb_checkpoints/tokenizer_config-checkpoint.json +1 -0
  14. .ipynb_checkpoints/validate-checkpoint.sh +3 -3
  15. .ipynb_checkpoints/vocab-checkpoint.json +1 -0
  16. README.md +30 -2
  17. eval.py +0 -13
  18. log_mozilla-foundation_common_voice_8_0_pt_test_predictions.txt +0 -0
  19. log_mozilla-foundation_common_voice_8_0_pt_test_targets.txt +0 -0
  20. log_speech-recognition-community-v2_dev_data_pt_validation_predictions.txt +0 -0
  21. log_speech-recognition-community-v2_dev_data_pt_validation_targets.txt +0 -0
  22. mozilla-foundation_common_voice_8_0_pt_test_eval_results.txt +2 -0
  23. run.sh +1 -1
  24. speech-recognition-community-v2_dev_data_pt_validation_eval_results.txt +2 -0
  25. validate.sh +3 -3
  26. wandb/run-20220201_223624-2b1hcyq3/files/config.yaml +6 -0
  27. wandb/run-20220201_223624-2b1hcyq3/files/output.log +7 -0
  28. wandb/run-20220201_223624-2b1hcyq3/files/wandb-summary.json +0 -0
  29. wandb/run-20220201_223624-2b1hcyq3/logs/debug-internal.log +120 -0
  30. wandb/run-20220201_223624-2b1hcyq3/logs/debug.log +130 -0
  31. wandb/run-20220201_223624-2b1hcyq3/run-2b1hcyq3.wandb +2 -2
.ipynb_checkpoints/README-checkpoint.md ADDED
@@ -0,0 +1,116 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - pt
4
+ license: apache-2.0
5
+ tags:
6
+ - automatic-speech-recognition
7
+ - robust-speech-event
8
+ - mozilla-foundation/common_voice_8_0
9
+ - generated_from_trainer
10
+ datasets:
11
+ - common_voice
12
+ model-index:
13
+ - name: xls-r-300m-pt
14
+ results:
15
+ - task:
16
+ name: Speech Recognition
17
+ type: automatic-speech-recognition
18
+ dataset:
19
+ name: Common Voice 8.0 fr
20
+ type: mozilla-foundation/common_voice_8_0
21
+ args: fr
22
+ metrics:
23
+ - name: Test WER
24
+ type: wer
25
+ value: 21.008
26
+ - name: Test CER
27
+ type: cer
28
+ value: 6.117
29
+ - task:
30
+ name: Speech Recognition
31
+ type: automatic-speech-recognition
32
+ dataset:
33
+ name: Robust Speech Event - Dev Data
34
+ type: speech-recognition-community-v2/dev_data
35
+ args: fr
36
+ metrics:
37
+ - name: Validation WER
38
+ type: wer
39
+ value: 47.812
40
+ - name: Validation CER
41
+ type: cer
42
+ value: 18.805
43
+ ---
44
+
45
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
46
+ should probably proofread and complete it, then remove this comment. -->
47
+
48
+ #
49
+
50
+ This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - PT dataset.
51
+ It achieves the following results on the evaluation set:
52
+ - Loss: 0.2290
53
+ - Wer: 0.2382
54
+
55
+ ## Model description
56
+
57
+ More information needed
58
+
59
+ ## Intended uses & limitations
60
+
61
+ More information needed
62
+
63
+ ## Training and evaluation data
64
+
65
+ More information needed
66
+
67
+ ## Training procedure
68
+
69
+ ### Training hyperparameters
70
+
71
+ The following hyperparameters were used during training:
72
+ - learning_rate: 0.0002
73
+ - train_batch_size: 32
74
+ - eval_batch_size: 32
75
+ - seed: 42
76
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
77
+ - lr_scheduler_type: linear
78
+ - lr_scheduler_warmup_steps: 1500
79
+ - num_epochs: 15.0
80
+ - mixed_precision_training: Native AMP
81
+
82
+ ### Training results
83
+
84
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
85
+ |:-------------:|:-----:|:-----:|:---------------:|:------:|
86
+ | 3.0952 | 0.64 | 500 | 3.0982 | 1.0 |
87
+ | 1.7975 | 1.29 | 1000 | 0.7887 | 0.5651 |
88
+ | 1.4138 | 1.93 | 1500 | 0.5238 | 0.4389 |
89
+ | 1.344 | 2.57 | 2000 | 0.4775 | 0.4318 |
90
+ | 1.2737 | 3.21 | 2500 | 0.4648 | 0.4075 |
91
+ | 1.2554 | 3.86 | 3000 | 0.4069 | 0.3678 |
92
+ | 1.1996 | 4.5 | 3500 | 0.3914 | 0.3668 |
93
+ | 1.1427 | 5.14 | 4000 | 0.3694 | 0.3572 |
94
+ | 1.1372 | 5.78 | 4500 | 0.3568 | 0.3501 |
95
+ | 1.0831 | 6.43 | 5000 | 0.3331 | 0.3253 |
96
+ | 1.1074 | 7.07 | 5500 | 0.3332 | 0.3352 |
97
+ | 1.0536 | 7.71 | 6000 | 0.3131 | 0.3152 |
98
+ | 1.0248 | 8.35 | 6500 | 0.3024 | 0.3023 |
99
+ | 1.0075 | 9.0 | 7000 | 0.2948 | 0.3028 |
100
+ | 0.979 | 9.64 | 7500 | 0.2796 | 0.2853 |
101
+ | 0.9594 | 10.28 | 8000 | 0.2719 | 0.2789 |
102
+ | 0.9172 | 10.93 | 8500 | 0.2620 | 0.2695 |
103
+ | 0.9047 | 11.57 | 9000 | 0.2537 | 0.2596 |
104
+ | 0.8777 | 12.21 | 9500 | 0.2438 | 0.2525 |
105
+ | 0.8629 | 12.85 | 10000 | 0.2409 | 0.2493 |
106
+ | 0.8575 | 13.5 | 10500 | 0.2366 | 0.2440 |
107
+ | 0.8361 | 14.14 | 11000 | 0.2317 | 0.2385 |
108
+ | 0.8126 | 14.78 | 11500 | 0.2290 | 0.2382 |
109
+
110
+
111
+ ### Framework versions
112
+
113
+ - Transformers 4.17.0.dev0
114
+ - Pytorch 1.10.2+cu102
115
+ - Datasets 1.18.2.dev0
116
+ - Tokenizers 0.11.0
.ipynb_checkpoints/added_tokens-checkpoint.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"<s>": 51, "</s>": 52}
.ipynb_checkpoints/all_results-checkpoint.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 15.0,
3
+ "eval_loss": 0.22902172803878784,
4
+ "eval_runtime": 327.6131,
5
+ "eval_samples": 8301,
6
+ "eval_samples_per_second": 25.338,
7
+ "eval_steps_per_second": 0.794,
8
+ "eval_wer": 0.23818860067667216,
9
+ "train_loss": 1.3307904394651542,
10
+ "train_runtime": 23640.5221,
11
+ "train_samples": 24877,
12
+ "train_samples_per_second": 15.785,
13
+ "train_steps_per_second": 0.494
14
+ }
.ipynb_checkpoints/config-checkpoint.json ADDED
@@ -0,0 +1,107 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-xls-r-300m",
3
+ "activation_dropout": 0.05,
4
+ "adapter_kernel_size": 3,
5
+ "adapter_stride": 2,
6
+ "add_adapter": false,
7
+ "apply_spec_augment": true,
8
+ "architectures": [
9
+ "Wav2Vec2ForCTC"
10
+ ],
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,
46
+ "diversity_loss_weight": 0.1,
47
+ "do_stable_layer_norm": true,
48
+ "eos_token_id": 2,
49
+ "feat_extract_activation": "gelu",
50
+ "feat_extract_dropout": 0.0,
51
+ "feat_extract_norm": "layer",
52
+ "feat_proj_dropout": 0.0,
53
+ "feat_quantizer_dropout": 0.0,
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": 10,
63
+ "mask_feature_min_masks": 0,
64
+ "mask_feature_prob": 0.3,
65
+ "mask_time_length": 10,
66
+ "mask_time_min_masks": 2,
67
+ "mask_time_prob": 0.65,
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": 50,
80
+ "proj_codevector_dim": 768,
81
+ "tdnn_dilation": [
82
+ 1,
83
+ 2,
84
+ 3,
85
+ 1,
86
+ 1
87
+ ],
88
+ "tdnn_dim": [
89
+ 512,
90
+ 512,
91
+ 512,
92
+ 512,
93
+ 1500
94
+ ],
95
+ "tdnn_kernel": [
96
+ 5,
97
+ 3,
98
+ 3,
99
+ 1,
100
+ 1
101
+ ],
102
+ "torch_dtype": "float32",
103
+ "transformers_version": "4.17.0.dev0",
104
+ "use_weighted_layer_sum": false,
105
+ "vocab_size": 53,
106
+ "xvector_output_dim": 512
107
+ }
.ipynb_checkpoints/eval-checkpoint.py ADDED
@@ -0,0 +1,140 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ import argparse
3
+ import re
4
+ import unicodedata
5
+ from typing import Dict
6
+
7
+ import torch
8
+ from datasets import Audio, Dataset, load_dataset, load_metric
9
+
10
+ from transformers import AutoFeatureExtractor, pipeline
11
+
12
+
13
+ def log_results(result: Dataset, args: Dict[str, str]):
14
+ """DO NOT CHANGE. This function computes and logs the result metrics."""
15
+
16
+ log_outputs = args.log_outputs
17
+ dataset_id = "_".join(args.dataset.split("/") + [args.config, args.split])
18
+
19
+ # load metric
20
+ wer = load_metric("wer")
21
+ cer = load_metric("cer")
22
+
23
+ # compute metrics
24
+ wer_result = wer.compute(references=result["target"], predictions=result["prediction"])
25
+ cer_result = cer.compute(references=result["target"], predictions=result["prediction"])
26
+
27
+ # print & log results
28
+ result_str = f"WER: {wer_result}\n" f"CER: {cer_result}"
29
+ print(result_str)
30
+
31
+ with open(f"{dataset_id}_eval_results.txt", "w") as f:
32
+ f.write(result_str)
33
+
34
+ # log all results in text file. Possibly interesting for analysis
35
+ if log_outputs is not None:
36
+ pred_file = f"log_{dataset_id}_predictions.txt"
37
+ target_file = f"log_{dataset_id}_targets.txt"
38
+
39
+ with open(pred_file, "w") as p, open(target_file, "w") as t:
40
+
41
+ # mapping function to write output
42
+ def write_to_file(batch, i):
43
+ p.write(f"{i}" + "\n")
44
+ p.write(batch["prediction"] + "\n")
45
+ t.write(f"{i}" + "\n")
46
+ t.write(batch["target"] + "\n")
47
+
48
+ result.map(write_to_file, with_indices=True)
49
+
50
+
51
+ chars_to_remove_regex = r'[\,\?\.\!\-\_\;\:\"\“\%\‘\”\�\^]'
52
+
53
+ def remove_special_characters(text):
54
+ text = re.sub(chars_to_remove_regex, '', text).lower()
55
+ return text
56
+
57
+ def normalize_text(text: str) -> str:
58
+ text = remove_special_characters(text)
59
+
60
+ # In addition, we can normalize the target text, e.g. removing new lines characters etc...
61
+ # note that order is important here!
62
+ token_sequences_to_ignore = ["\n\n", "\n", " ", " "]
63
+
64
+ for t in token_sequences_to_ignore:
65
+ text = " ".join(text.split(t))
66
+
67
+ return text
68
+
69
+
70
+ def main(args):
71
+ # load dataset
72
+ dataset = load_dataset(args.dataset, args.config, split=args.split, use_auth_token=True)
73
+
74
+ # for testing: only process the first two examples as a test
75
+ # dataset = dataset.select(range(2))
76
+
77
+ # load processor
78
+ feature_extractor = AutoFeatureExtractor.from_pretrained(args.model_id)
79
+ sampling_rate = feature_extractor.sampling_rate
80
+
81
+ # resample audio
82
+ dataset = dataset.cast_column("audio", Audio(sampling_rate=sampling_rate))
83
+
84
+ # load eval pipeline
85
+ if args.device is None:
86
+ args.device = 0 if torch.cuda.is_available() else -1
87
+ asr = pipeline("automatic-speech-recognition", model=args.model_id, device=args.device)
88
+
89
+ # map function to decode audio
90
+ def map_to_pred(batch):
91
+ prediction = asr(
92
+ batch["audio"]["array"], chunk_length_s=args.chunk_length_s, stride_length_s=args.stride_length_s
93
+ )
94
+
95
+ batch["prediction"] = prediction["text"] # "".join(prediction["text"].split("<s>"))
96
+ batch["target"] = normalize_text(batch["sentence"])
97
+ return batch
98
+
99
+ # run inference on all examples
100
+ result = dataset.map(map_to_pred, remove_columns=dataset.column_names)
101
+
102
+ # compute and log_results
103
+ # do not change function below
104
+ log_results(result, args)
105
+
106
+
107
+ if __name__ == "__main__":
108
+ parser = argparse.ArgumentParser()
109
+
110
+ parser.add_argument(
111
+ "--model_id", type=str, required=True, help="Model identifier. Should be loadable with 🤗 Transformers"
112
+ )
113
+ parser.add_argument(
114
+ "--dataset",
115
+ type=str,
116
+ required=True,
117
+ help="Dataset name to evaluate the `model_id`. Should be loadable with 🤗 Datasets",
118
+ )
119
+ parser.add_argument(
120
+ "--config", type=str, required=True, help="Config of the dataset. *E.g.* `'en'` for Common Voice"
121
+ )
122
+ parser.add_argument("--split", type=str, required=True, help="Split of the dataset. *E.g.* `'test'`")
123
+ parser.add_argument(
124
+ "--chunk_length_s", type=float, default=None, help="Chunk length in seconds. Defaults to 5 seconds."
125
+ )
126
+ parser.add_argument(
127
+ "--stride_length_s", type=float, default=None, help="Stride of the audio chunks. Defaults to 1 second."
128
+ )
129
+ parser.add_argument(
130
+ "--log_outputs", action="store_true", help="If defined, write outputs to log file for analysis."
131
+ )
132
+ parser.add_argument(
133
+ "--device",
134
+ type=int,
135
+ default=None,
136
+ help="The device to run the pipeline on. -1 for CPU (default), 0 for the first GPU and so on.",
137
+ )
138
+ args = parser.parse_args()
139
+
140
+ main(args)
.ipynb_checkpoints/eval_results-checkpoint.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 15.0,
3
+ "eval_loss": 0.22902172803878784,
4
+ "eval_runtime": 327.6131,
5
+ "eval_samples": 8301,
6
+ "eval_samples_per_second": 25.338,
7
+ "eval_steps_per_second": 0.794,
8
+ "eval_wer": 0.23818860067667216
9
+ }
.ipynb_checkpoints/log_mozilla-foundation_common_voice_8_0_pt_test_predictions-checkpoint.txt ADDED
The diff for this file is too large to render. See raw diff
 
.ipynb_checkpoints/log_mozilla-foundation_common_voice_8_0_pt_test_targets-checkpoint.txt ADDED
The diff for this file is too large to render. See raw diff
 
.ipynb_checkpoints/log_speech-recognition-community-v2_dev_data_pt_validation_predictions-checkpoint.txt ADDED
The diff for this file is too large to render. See raw diff
 
.ipynb_checkpoints/log_speech-recognition-community-v2_dev_data_pt_validation_targets-checkpoint.txt ADDED
The diff for this file is too large to render. See raw diff
 
.ipynb_checkpoints/preprocessor_config-checkpoint.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
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
+ }
.ipynb_checkpoints/run-checkpoint.sh CHANGED
@@ -28,7 +28,7 @@ python run_speech_recognition_ctc.py \
28
  --mask_feature_length="10" \
29
  --gradient_checkpointing \
30
  --report_to="wandb" \
31
- --run_name="xls-r-300m-fr" \
32
  --max_duration_in_seconds="20" \
33
  --use_auth_token \
34
  --fp16 \
 
28
  --mask_feature_length="10" \
29
  --gradient_checkpointing \
30
  --report_to="wandb" \
31
+ --run_name="xls-r-300m-pt" \
32
  --max_duration_in_seconds="20" \
33
  --use_auth_token \
34
  --fp16 \
.ipynb_checkpoints/tokenizer_config-checkpoint.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"unk_token": "[UNK]", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "[PAD]", "do_lower_case": false, "word_delimiter_token": "|", "config": null, "tokenizer_type": "wav2vec2", "special_tokens_map_file": null, "tokenizer_file": null, "name_or_path": "./", "tokenizer_class": "Wav2Vec2CTCTokenizer"}
.ipynb_checkpoints/validate-checkpoint.sh CHANGED
@@ -1,3 +1,3 @@
1
- python eval.py --model_id="./" --dataset="mozilla-foundation/common_voice_8_0" --config="fr" --split="test" --log_outputs
2
- python eval.py --model_id="./" --dataset="speech-recognition-community-v2/dev_data" --config="fr" \
3
- --split="validation" --chunk_length_s="5.0" --stride_length_s="1.0" --log_outputs
 
1
+ python eval.py --model_id="./" --dataset="mozilla-foundation/common_voice_8_0" --config="pt" --split="test" --log_outputs
2
+ python eval.py --model_id="./" --dataset="speech-recognition-community-v2/dev_data" --config="pt" \
3
+ --split="validation" --chunk_length_s="5.0" --stride_length_s="1.0" --log_outputs --device="-1"
.ipynb_checkpoints/vocab-checkpoint.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"&": 1, "'": 2, "a": 3, "b": 4, "c": 5, "d": 6, "e": 7, "f": 8, "g": 9, "h": 10, "i": 11, "j": 12, "k": 13, "l": 14, "m": 15, "n": 16, "o": 17, "p": 18, "q": 19, "r": 20, "s": 21, "t": 22, "u": 23, "v": 24, "w": 25, "x": 26, "y": 27, "z": 28, "«": 29, "´": 30, "»": 31, "à": 32, "á": 33, "â": 34, "ã": 35, "ç": 36, "è": 37, "é": 38, "ê": 39, "í": 40, "ñ": 41, "ó": 42, "ô": 43, "õ": 44, "ú": 45, "ü": 46, "š": 47, "ž": 48, "|": 0, "[UNK]": 49, "[PAD]": 50}
README.md CHANGED
@@ -10,8 +10,36 @@ tags:
10
  datasets:
11
  - common_voice
12
  model-index:
13
- - name: ''
14
- results: []
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
15
  ---
16
 
17
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
10
  datasets:
11
  - common_voice
12
  model-index:
13
+ - name: xls-r-300m-pt
14
+ results:
15
+ - task:
16
+ name: Speech Recognition
17
+ type: automatic-speech-recognition
18
+ dataset:
19
+ name: Common Voice 8.0 fr
20
+ type: mozilla-foundation/common_voice_8_0
21
+ args: fr
22
+ metrics:
23
+ - name: Test WER
24
+ type: wer
25
+ value: 21.008
26
+ - name: Test CER
27
+ type: cer
28
+ value: 6.117
29
+ - task:
30
+ name: Speech Recognition
31
+ type: automatic-speech-recognition
32
+ dataset:
33
+ name: Robust Speech Event - Dev Data
34
+ type: speech-recognition-community-v2/dev_data
35
+ args: fr
36
+ metrics:
37
+ - name: Validation WER
38
+ type: wer
39
+ value: 47.812
40
+ - name: Validation CER
41
+ type: cer
42
+ value: 18.805
43
  ---
44
 
45
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
eval.py CHANGED
@@ -50,21 +50,8 @@ def log_results(result: Dataset, args: Dict[str, str]):
50
 
51
  chars_to_remove_regex = r'[\,\?\.\!\-\_\;\:\"\“\%\‘\”\�\^]'
52
 
53
- def remove_accents(text):
54
- nfkd_form = unicodedata.normalize('NFKD', text)
55
- return u"".join([c for c in nfkd_form if not unicodedata.combining(c)])
56
-
57
  def remove_special_characters(text):
58
  text = re.sub(chars_to_remove_regex, '', text).lower()
59
- text = re.sub("ç", r'[cedille]', text)
60
- text = re.sub("&", r'et', text)
61
- text = re.sub("%", r' pourcents', text)
62
- text = re.sub("([0-9]+)(,|.)([0-9+])", r'\1 virgule \3', text)
63
- text = re.sub("\$", r'dollar', text)
64
- text = re.sub("\£", r'livre', text)
65
- text = re.sub("\€", r'euro', text)
66
- text = remove_accents(text)
67
- text = re.sub(r"\[cedille\]", 'ç', text) + " "
68
  return text
69
 
70
  def normalize_text(text: str) -> str:
 
50
 
51
  chars_to_remove_regex = r'[\,\?\.\!\-\_\;\:\"\“\%\‘\”\�\^]'
52
 
 
 
 
 
53
  def remove_special_characters(text):
54
  text = re.sub(chars_to_remove_regex, '', text).lower()
 
 
 
 
 
 
 
 
 
55
  return text
56
 
57
  def normalize_text(text: str) -> str:
log_mozilla-foundation_common_voice_8_0_pt_test_predictions.txt ADDED
The diff for this file is too large to render. See raw diff
 
log_mozilla-foundation_common_voice_8_0_pt_test_targets.txt ADDED
The diff for this file is too large to render. See raw diff
 
log_speech-recognition-community-v2_dev_data_pt_validation_predictions.txt ADDED
The diff for this file is too large to render. See raw diff
 
log_speech-recognition-community-v2_dev_data_pt_validation_targets.txt ADDED
The diff for this file is too large to render. See raw diff
 
mozilla-foundation_common_voice_8_0_pt_test_eval_results.txt ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ WER: 0.21008068014227466
2
+ CER: 0.06117429121735184
run.sh CHANGED
@@ -28,7 +28,7 @@ python run_speech_recognition_ctc.py \
28
  --mask_feature_length="10" \
29
  --gradient_checkpointing \
30
  --report_to="wandb" \
31
- --run_name="xls-r-300m-fr" \
32
  --max_duration_in_seconds="20" \
33
  --use_auth_token \
34
  --fp16 \
 
28
  --mask_feature_length="10" \
29
  --gradient_checkpointing \
30
  --report_to="wandb" \
31
+ --run_name="xls-r-300m-pt" \
32
  --max_duration_in_seconds="20" \
33
  --use_auth_token \
34
  --fp16 \
speech-recognition-community-v2_dev_data_pt_validation_eval_results.txt ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ WER: 0.4781234251156824
2
+ CER: 0.1880587725716785
validate.sh CHANGED
@@ -1,3 +1,3 @@
1
- python eval.py --model_id="./" --dataset="mozilla-foundation/common_voice_8_0" --config="fr" --split="test" --log_outputs
2
- python eval.py --model_id="./" --dataset="speech-recognition-community-v2/dev_data" --config="fr" \
3
- --split="validation" --chunk_length_s="5.0" --stride_length_s="1.0" --log_outputs
 
1
+ python eval.py --model_id="./" --dataset="mozilla-foundation/common_voice_8_0" --config="pt" --split="test" --log_outputs
2
+ python eval.py --model_id="./" --dataset="speech-recognition-community-v2/dev_data" --config="pt" \
3
+ --split="validation" --chunk_length_s="5.0" --stride_length_s="1.0" --log_outputs --device="-1"
wandb/run-20220201_223624-2b1hcyq3/files/config.yaml CHANGED
@@ -4829,7 +4829,13 @@ _wandb:
4829
  - 1
4830
  - 5
4831
  - 11
 
 
 
 
4832
  3:
 
 
4833
  - 13
4834
  4: 3.8.8
4835
  5: 0.12.9
 
4829
  - 1
4830
  - 5
4831
  - 11
4832
+ 2:
4833
+ - 1
4834
+ - 5
4835
+ - 11
4836
  3:
4837
+ - 1
4838
+ - 7
4839
  - 13
4840
  4: 3.8.8
4841
  5: 0.12.9
wandb/run-20220201_223624-2b1hcyq3/files/output.log CHANGED
@@ -10238,3 +10238,10 @@ Deleting older checkpoint [checkpoint-4500] due to args.save_total_limit
10238
  02/02/2022 05:17:09 - WARNING - huggingface_hub.repository - To https://huggingface.co/AlexN/xls-r-300m-pt
10239
  1f92738..6db3b99 main -> main
10240
  0%| | 0/260 [00:00<?, ?it/s]
 
 
 
 
 
 
 
 
10238
  02/02/2022 05:17:09 - WARNING - huggingface_hub.repository - To https://huggingface.co/AlexN/xls-r-300m-pt
10239
  1f92738..6db3b99 main -> main
10240
  0%| | 0/260 [00:00<?, ?it/s]
10241
+ 0%| | 0/260 [00:00<?, ?it/s]
10242
+ 0%| | 0/260 [00:00<?, ?it/s]
10243
+ 0%| | 0/260 [00:00<?, ?it/s]
10244
+ 02/02/2022 05:17:23 - WARNING - huggingface_hub.repository - To https://huggingface.co/AlexN/xls-r-300m-pt
10245
+ 6db3b99..0a6506a main -> main
10246
+ 0%| | 0/260 [00:00<?, ?it/s]
10247
+ 0%| | 0/260 [00:00<?, ?it/s]
wandb/run-20220201_223624-2b1hcyq3/files/wandb-summary.json CHANGED
The diff for this file is too large to render. See raw diff
 
wandb/run-20220201_223624-2b1hcyq3/logs/debug-internal.log CHANGED
@@ -15864,3 +15864,123 @@
15864
  2022-02-02 05:17:12,233 INFO Thread-8 :38198 [dir_watcher.py:_on_file_modified():230] file/dir modified: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/output.log
15865
  2022-02-02 05:17:13,234 INFO Thread-8 :38198 [dir_watcher.py:_on_file_modified():230] file/dir modified: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/output.log
15866
  2022-02-02 05:17:15,236 INFO Thread-8 :38198 [dir_watcher.py:_on_file_modified():230] file/dir modified: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/output.log
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
15864
  2022-02-02 05:17:12,233 INFO Thread-8 :38198 [dir_watcher.py:_on_file_modified():230] file/dir modified: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/output.log
15865
  2022-02-02 05:17:13,234 INFO Thread-8 :38198 [dir_watcher.py:_on_file_modified():230] file/dir modified: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/output.log
15866
  2022-02-02 05:17:15,236 INFO Thread-8 :38198 [dir_watcher.py:_on_file_modified():230] file/dir modified: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/output.log
15867
+ 2022-02-02 05:17:21,239 INFO Thread-8 :38198 [dir_watcher.py:_on_file_modified():230] file/dir modified: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/output.log
15868
+ 2022-02-02 05:17:23,240 INFO Thread-8 :38198 [dir_watcher.py:_on_file_modified():230] file/dir modified: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/output.log
15869
+ 2022-02-02 05:17:24,240 INFO Thread-8 :38198 [dir_watcher.py:_on_file_modified():230] file/dir modified: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/output.log
15870
+ 2022-02-02 05:17:25,241 INFO Thread-8 :38198 [dir_watcher.py:_on_file_modified():230] file/dir modified: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/output.log
15871
+ 2022-02-02 05:17:28,344 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15872
+ 2022-02-02 05:17:28,344 DEBUG SenderThread:38198 [sender.py:send():234] send: telemetry
15873
+ 2022-02-02 05:17:28,345 DEBUG SenderThread:38198 [sender.py:send():234] send: exit
15874
+ 2022-02-02 05:17:28,345 INFO SenderThread:38198 [sender.py:send_exit():366] handling exit code: 0
15875
+ 2022-02-02 05:17:28,345 INFO SenderThread:38198 [sender.py:send_exit():368] handling runtime: 24062
15876
+ 2022-02-02 05:17:28,377 INFO SenderThread:38198 [sender.py:_save_file():939] saving file wandb-summary.json with policy end
15877
+ 2022-02-02 05:17:28,377 INFO SenderThread:38198 [sender.py:send_exit():374] send defer
15878
+ 2022-02-02 05:17:28,377 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15879
+ 2022-02-02 05:17:28,378 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: defer
15880
+ 2022-02-02 05:17:28,379 INFO HandlerThread:38198 [handler.py:handle_request_defer():147] handle defer: 0
15881
+ 2022-02-02 05:17:28,379 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: defer
15882
+ 2022-02-02 05:17:28,379 INFO SenderThread:38198 [sender.py:send_request_defer():383] handle sender defer: 0
15883
+ 2022-02-02 05:17:28,379 INFO SenderThread:38198 [sender.py:transition_state():387] send defer: 1
15884
+ 2022-02-02 05:17:28,380 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: defer
15885
+ 2022-02-02 05:17:28,380 INFO HandlerThread:38198 [handler.py:handle_request_defer():147] handle defer: 1
15886
+ 2022-02-02 05:17:28,486 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: defer
15887
+ 2022-02-02 05:17:28,486 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15888
+ 2022-02-02 05:17:28,486 INFO SenderThread:38198 [sender.py:send_request_defer():383] handle sender defer: 1
15889
+ 2022-02-02 05:17:28,487 INFO SenderThread:38198 [sender.py:transition_state():387] send defer: 2
15890
+ 2022-02-02 05:17:28,487 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15891
+ 2022-02-02 05:17:28,488 DEBUG SenderThread:38198 [sender.py:send():234] send: stats
15892
+ 2022-02-02 05:17:28,489 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: defer
15893
+ 2022-02-02 05:17:28,489 INFO HandlerThread:38198 [handler.py:handle_request_defer():147] handle defer: 2
15894
+ 2022-02-02 05:17:28,489 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: defer
15895
+ 2022-02-02 05:17:28,490 INFO SenderThread:38198 [sender.py:send_request_defer():383] handle sender defer: 2
15896
+ 2022-02-02 05:17:28,490 INFO SenderThread:38198 [sender.py:transition_state():387] send defer: 3
15897
+ 2022-02-02 05:17:28,490 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: defer
15898
+ 2022-02-02 05:17:28,490 INFO HandlerThread:38198 [handler.py:handle_request_defer():147] handle defer: 3
15899
+ 2022-02-02 05:17:28,530 DEBUG SenderThread:38198 [sender.py:send():234] send: summary
15900
+ 2022-02-02 05:17:28,576 INFO SenderThread:38198 [sender.py:_save_file():939] saving file wandb-summary.json with policy end
15901
+ 2022-02-02 05:17:28,576 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: defer
15902
+ 2022-02-02 05:17:28,576 INFO SenderThread:38198 [sender.py:send_request_defer():383] handle sender defer: 3
15903
+ 2022-02-02 05:17:28,576 INFO SenderThread:38198 [sender.py:transition_state():387] send defer: 4
15904
+ 2022-02-02 05:17:28,577 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: defer
15905
+ 2022-02-02 05:17:28,578 INFO HandlerThread:38198 [handler.py:handle_request_defer():147] handle defer: 4
15906
+ 2022-02-02 05:17:28,578 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: defer
15907
+ 2022-02-02 05:17:28,578 INFO SenderThread:38198 [sender.py:send_request_defer():383] handle sender defer: 4
15908
+ 2022-02-02 05:17:28,590 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15909
+ 2022-02-02 05:17:29,255 INFO Thread-8 :38198 [dir_watcher.py:_on_file_modified():230] file/dir modified: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/output.log
15910
+ 2022-02-02 05:17:29,255 INFO SenderThread:38198 [sender.py:transition_state():387] send defer: 5
15911
+ 2022-02-02 05:17:29,255 INFO Thread-8 :38198 [dir_watcher.py:_on_file_modified():230] file/dir modified: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/wandb-summary.json
15912
+ 2022-02-02 05:17:29,256 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15913
+ 2022-02-02 05:17:29,257 INFO Thread-8 :38198 [dir_watcher.py:_on_file_modified():230] file/dir modified: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/config.yaml
15914
+ 2022-02-02 05:17:29,258 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: defer
15915
+ 2022-02-02 05:17:29,258 INFO HandlerThread:38198 [handler.py:handle_request_defer():147] handle defer: 5
15916
+ 2022-02-02 05:17:29,259 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: defer
15917
+ 2022-02-02 05:17:29,259 INFO SenderThread:38198 [sender.py:send_request_defer():383] handle sender defer: 5
15918
+ 2022-02-02 05:17:29,259 INFO SenderThread:38198 [dir_watcher.py:finish():283] shutting down directory watcher
15919
+ 2022-02-02 05:17:29,359 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15920
+ 2022-02-02 05:17:30,258 INFO SenderThread:38198 [dir_watcher.py:finish():313] scan: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files
15921
+ 2022-02-02 05:17:30,258 INFO SenderThread:38198 [dir_watcher.py:finish():327] scan save: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/output.log output.log
15922
+ 2022-02-02 05:17:30,259 INFO SenderThread:38198 [dir_watcher.py:finish():327] scan save: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/wandb-metadata.json wandb-metadata.json
15923
+ 2022-02-02 05:17:30,259 INFO SenderThread:38198 [dir_watcher.py:finish():327] scan save: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/config.yaml config.yaml
15924
+ 2022-02-02 05:17:30,260 INFO SenderThread:38198 [dir_watcher.py:finish():327] scan save: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/conda-environment.yaml conda-environment.yaml
15925
+ 2022-02-02 05:17:30,267 INFO SenderThread:38198 [dir_watcher.py:finish():327] scan save: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/wandb-summary.json wandb-summary.json
15926
+ 2022-02-02 05:17:30,273 INFO SenderThread:38198 [dir_watcher.py:finish():327] scan save: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/requirements.txt requirements.txt
15927
+ 2022-02-02 05:17:30,281 INFO SenderThread:38198 [sender.py:transition_state():387] send defer: 6
15928
+ 2022-02-02 05:17:30,282 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15929
+ 2022-02-02 05:17:30,286 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: defer
15930
+ 2022-02-02 05:17:30,287 INFO HandlerThread:38198 [handler.py:handle_request_defer():147] handle defer: 6
15931
+ 2022-02-02 05:17:30,287 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: defer
15932
+ 2022-02-02 05:17:30,288 INFO SenderThread:38198 [sender.py:send_request_defer():383] handle sender defer: 6
15933
+ 2022-02-02 05:17:30,288 INFO SenderThread:38198 [file_pusher.py:finish():177] shutting down file pusher
15934
+ 2022-02-02 05:17:30,384 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15935
+ 2022-02-02 05:17:30,385 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15936
+ 2022-02-02 05:17:30,488 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15937
+ 2022-02-02 05:17:30,489 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15938
+ 2022-02-02 05:17:30,593 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15939
+ 2022-02-02 05:17:30,593 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15940
+ 2022-02-02 05:17:30,696 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15941
+ 2022-02-02 05:17:30,696 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15942
+ 2022-02-02 05:17:30,799 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15943
+ 2022-02-02 05:17:30,799 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15944
+ 2022-02-02 05:17:30,902 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15945
+ 2022-02-02 05:17:30,902 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15946
+ 2022-02-02 05:17:31,005 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15947
+ 2022-02-02 05:17:31,005 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15948
+ 2022-02-02 05:17:31,043 INFO Thread-12 :38198 [upload_job.py:push():137] Uploaded file /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/output.log
15949
+ 2022-02-02 05:17:31,090 INFO Thread-14 :38198 [upload_job.py:push():137] Uploaded file /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/wandb-summary.json
15950
+ 2022-02-02 05:17:31,098 INFO Thread-15 :38198 [upload_job.py:push():137] Uploaded file /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/requirements.txt
15951
+ 2022-02-02 05:17:31,108 INFO Thread-13 :38198 [upload_job.py:push():137] Uploaded file /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/files/config.yaml
15952
+ 2022-02-02 05:17:31,110 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15953
+ 2022-02-02 05:17:31,110 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15954
+ 2022-02-02 05:17:31,213 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15955
+ 2022-02-02 05:17:31,214 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15956
+ 2022-02-02 05:17:31,310 INFO Thread-7 :38198 [sender.py:transition_state():387] send defer: 7
15957
+ 2022-02-02 05:17:31,311 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: defer
15958
+ 2022-02-02 05:17:31,311 INFO HandlerThread:38198 [handler.py:handle_request_defer():147] handle defer: 7
15959
+ 2022-02-02 05:17:31,311 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: defer
15960
+ 2022-02-02 05:17:31,311 INFO SenderThread:38198 [sender.py:send_request_defer():383] handle sender defer: 7
15961
+ 2022-02-02 05:17:31,322 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15962
+ 2022-02-02 05:17:32,128 INFO SenderThread:38198 [sender.py:transition_state():387] send defer: 8
15963
+ 2022-02-02 05:17:32,128 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15964
+ 2022-02-02 05:17:32,129 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: defer
15965
+ 2022-02-02 05:17:32,130 INFO HandlerThread:38198 [handler.py:handle_request_defer():147] handle defer: 8
15966
+ 2022-02-02 05:17:32,130 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: defer
15967
+ 2022-02-02 05:17:32,130 INFO SenderThread:38198 [sender.py:send_request_defer():383] handle sender defer: 8
15968
+ 2022-02-02 05:17:32,130 INFO SenderThread:38198 [sender.py:transition_state():387] send defer: 9
15969
+ 2022-02-02 05:17:32,132 DEBUG SenderThread:38198 [sender.py:send():234] send: final
15970
+ 2022-02-02 05:17:32,133 DEBUG SenderThread:38198 [sender.py:send():234] send: footer
15971
+ 2022-02-02 05:17:32,133 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: defer
15972
+ 2022-02-02 05:17:32,134 INFO HandlerThread:38198 [handler.py:handle_request_defer():147] handle defer: 9
15973
+ 2022-02-02 05:17:32,134 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: defer
15974
+ 2022-02-02 05:17:32,134 INFO SenderThread:38198 [sender.py:send_request_defer():383] handle sender defer: 9
15975
+ 2022-02-02 05:17:32,231 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: poll_exit
15976
+ 2022-02-02 05:17:32,232 DEBUG SenderThread:38198 [sender.py:send_request():248] send_request: poll_exit
15977
+ 2022-02-02 05:17:32,232 INFO SenderThread:38198 [file_pusher.py:join():182] waiting for file pusher
15978
+ 2022-02-02 05:17:32,548 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: get_summary
15979
+ 2022-02-02 05:17:32,620 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: sampled_history
15980
+ 2022-02-02 05:17:32,624 DEBUG HandlerThread:38198 [handler.py:handle_request():130] handle_request: shutdown
15981
+ 2022-02-02 05:17:32,624 INFO HandlerThread:38198 [handler.py:finish():731] shutting down handler
15982
+ 2022-02-02 05:17:33,133 INFO WriterThread:38198 [datastore.py:close():281] close: /workspace/xls-r-300m-pt/wandb/run-20220201_223624-2b1hcyq3/run-2b1hcyq3.wandb
15983
+ 2022-02-02 05:17:33,546 INFO SenderThread:38198 [sender.py:finish():1070] shutting down sender
15984
+ 2022-02-02 05:17:33,546 INFO SenderThread:38198 [file_pusher.py:finish():177] shutting down file pusher
15985
+ 2022-02-02 05:17:33,547 INFO SenderThread:38198 [file_pusher.py:join():182] waiting for file pusher
15986
+ 2022-02-02 05:17:33,554 INFO MainThread:38198 [internal.py:handle_exit():77] Internal process exited
wandb/run-20220201_223624-2b1hcyq3/logs/debug.log CHANGED
@@ -23,3 +23,133 @@ config: {}
23
  2022-02-01 22:36:26,500 INFO MainThread:37310 [wandb_init.py:init():633] run started, returning control to user process
24
  2022-02-01 22:36:26,504 INFO MainThread:37310 [wandb_run.py:_config_callback():956] config_cb None None {'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': 'float32', 'use_bfloat16': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'chunk_size_feed_forward': 0, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'architectures': ['Wav2Vec2ForPreTraining'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': 1, 'pad_token_id': 50, 'eos_token_id': 2, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'facebook/wav2vec2-xls-r-300m', 'transformers_version': '4.17.0.dev0', 'feat_extract_dropout': 0.0, 'model_type': 'wav2vec2', 'num_feat_extract_layers': 7, 'hidden_size': 1024, 'feat_extract_norm': 'layer', 'feat_extract_activation': 'gelu', 'conv_dim': [512, 512, 512, 512, 512, 512, 512], 'conv_stride': [5, 2, 2, 2, 2, 2, 2], 'conv_kernel': [10, 3, 3, 3, 3, 2, 2], 'conv_bias': True, 'num_conv_pos_embeddings': 128, 'num_conv_pos_embedding_groups': 16, 'num_hidden_layers': 24, 'intermediate_size': 4096, 'hidden_act': 'gelu', 'num_attention_heads': 16, 'hidden_dropout': 0.0, 'attention_dropout': 0.0, 'activation_dropout': 0.05, 'feat_proj_dropout': 0.0, 'final_dropout': 0.0, 'layerdrop': 0.0, 'layer_norm_eps': 1e-05, 'initializer_range': 0.02, 'vocab_size': 53, 'do_stable_layer_norm': True, 'use_weighted_layer_sum': False, 'apply_spec_augment': True, 'mask_time_prob': 0.65, 'mask_time_length': 10, 'mask_time_min_masks': 2, 'mask_feature_prob': 0.3, 'mask_feature_length': 10, 'mask_feature_min_masks': 0, 'num_codevectors_per_group': 320, 'num_codevector_groups': 2, 'contrastive_logits_temperature': 0.1, 'feat_quantizer_dropout': 0.0, 'num_negatives': 100, 'codevector_dim': 768, 'proj_codevector_dim': 768, 'diversity_loss_weight': 0.1, 'ctc_loss_reduction': 'mean', 'ctc_zero_infinity': False, 'add_adapter': False, 'adapter_kernel_size': 3, 'adapter_stride': 2, 'num_adapter_layers': 3, 'output_hidden_size': 1024, 'classifier_proj_size': 256, 'tdnn_dim': [512, 512, 512, 512, 1500], 'tdnn_kernel': [5, 3, 3, 1, 1], 'tdnn_dilation': [1, 2, 3, 1, 1], 'xvector_output_dim': 512, 'output_dir': './', 'overwrite_output_dir': True, 'do_train': True, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'steps', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 32, 'per_gpu_train_batch_size': 'None', 'per_gpu_eval_batch_size': 'None', 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': 'None', 'learning_rate': 0.0002, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 15.0, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'warmup_ratio': 0.0, 'warmup_steps': 1500, 'log_level': -1, 'log_level_replica': -1, 'log_on_each_node': True, 'logging_dir': './runs/Feb01_22-34-39_job-1abccd0a-3293-4ffe-8274-9e8f841f653f', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 100, 'logging_nan_inf_filter': True, 'save_strategy': 'steps', 'save_steps': 500, 'save_total_limit': 3, 'save_on_each_node': False, 'no_cuda': False, 'seed': 42, 'bf16': False, 'fp16': True, 'fp16_opt_level': 'O1', 'half_precision_backend': 'amp', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': 'None', 'local_rank': -1, 'xpu_backend': 'None', 'tpu_num_cores': 'None', 'tpu_metrics_debug': False, 'debug': '[]', 'dataloader_drop_last': False, 'eval_steps': 500, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': 'xls-r-300m-fr', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': 'None', 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'sharded_ddp': '[]', 'deepspeed': 'None', 'label_smoothing_factor': 0.0, 'optim': 'adamw_hf', 'adafactor': False, 'group_by_length': True, 'length_column_name': 'input_length', 'report_to': "['wandb']", 'ddp_find_unused_parameters': 'None', 'ddp_bucket_cap_mb': 'None', 'dataloader_pin_memory': True, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': 'None', 'hub_model_id': 'None', 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'gradient_checkpointing': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': 'None', 'push_to_hub_organization': 'None', 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', '_n_gpu': 1, 'mp_parameters': '', 'train_batch_size': 32, 'eval_batch_size': 32}
25
  2022-02-01 22:36:26,509 INFO MainThread:37310 [wandb_watch.py:watch():43] Watching
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
23
  2022-02-01 22:36:26,500 INFO MainThread:37310 [wandb_init.py:init():633] run started, returning control to user process
24
  2022-02-01 22:36:26,504 INFO MainThread:37310 [wandb_run.py:_config_callback():956] config_cb None None {'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': 'float32', 'use_bfloat16': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'chunk_size_feed_forward': 0, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'architectures': ['Wav2Vec2ForPreTraining'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': 1, 'pad_token_id': 50, 'eos_token_id': 2, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'facebook/wav2vec2-xls-r-300m', 'transformers_version': '4.17.0.dev0', 'feat_extract_dropout': 0.0, 'model_type': 'wav2vec2', 'num_feat_extract_layers': 7, 'hidden_size': 1024, 'feat_extract_norm': 'layer', 'feat_extract_activation': 'gelu', 'conv_dim': [512, 512, 512, 512, 512, 512, 512], 'conv_stride': [5, 2, 2, 2, 2, 2, 2], 'conv_kernel': [10, 3, 3, 3, 3, 2, 2], 'conv_bias': True, 'num_conv_pos_embeddings': 128, 'num_conv_pos_embedding_groups': 16, 'num_hidden_layers': 24, 'intermediate_size': 4096, 'hidden_act': 'gelu', 'num_attention_heads': 16, 'hidden_dropout': 0.0, 'attention_dropout': 0.0, 'activation_dropout': 0.05, 'feat_proj_dropout': 0.0, 'final_dropout': 0.0, 'layerdrop': 0.0, 'layer_norm_eps': 1e-05, 'initializer_range': 0.02, 'vocab_size': 53, 'do_stable_layer_norm': True, 'use_weighted_layer_sum': False, 'apply_spec_augment': True, 'mask_time_prob': 0.65, 'mask_time_length': 10, 'mask_time_min_masks': 2, 'mask_feature_prob': 0.3, 'mask_feature_length': 10, 'mask_feature_min_masks': 0, 'num_codevectors_per_group': 320, 'num_codevector_groups': 2, 'contrastive_logits_temperature': 0.1, 'feat_quantizer_dropout': 0.0, 'num_negatives': 100, 'codevector_dim': 768, 'proj_codevector_dim': 768, 'diversity_loss_weight': 0.1, 'ctc_loss_reduction': 'mean', 'ctc_zero_infinity': False, 'add_adapter': False, 'adapter_kernel_size': 3, 'adapter_stride': 2, 'num_adapter_layers': 3, 'output_hidden_size': 1024, 'classifier_proj_size': 256, 'tdnn_dim': [512, 512, 512, 512, 1500], 'tdnn_kernel': [5, 3, 3, 1, 1], 'tdnn_dilation': [1, 2, 3, 1, 1], 'xvector_output_dim': 512, 'output_dir': './', 'overwrite_output_dir': True, 'do_train': True, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'steps', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 32, 'per_gpu_train_batch_size': 'None', 'per_gpu_eval_batch_size': 'None', 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': 'None', 'learning_rate': 0.0002, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 15.0, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'warmup_ratio': 0.0, 'warmup_steps': 1500, 'log_level': -1, 'log_level_replica': -1, 'log_on_each_node': True, 'logging_dir': './runs/Feb01_22-34-39_job-1abccd0a-3293-4ffe-8274-9e8f841f653f', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 100, 'logging_nan_inf_filter': True, 'save_strategy': 'steps', 'save_steps': 500, 'save_total_limit': 3, 'save_on_each_node': False, 'no_cuda': False, 'seed': 42, 'bf16': False, 'fp16': True, 'fp16_opt_level': 'O1', 'half_precision_backend': 'amp', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': 'None', 'local_rank': -1, 'xpu_backend': 'None', 'tpu_num_cores': 'None', 'tpu_metrics_debug': False, 'debug': '[]', 'dataloader_drop_last': False, 'eval_steps': 500, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': 'xls-r-300m-fr', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': 'None', 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'sharded_ddp': '[]', 'deepspeed': 'None', 'label_smoothing_factor': 0.0, 'optim': 'adamw_hf', 'adafactor': False, 'group_by_length': True, 'length_column_name': 'input_length', 'report_to': "['wandb']", 'ddp_find_unused_parameters': 'None', 'ddp_bucket_cap_mb': 'None', 'dataloader_pin_memory': True, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': 'None', 'hub_model_id': 'None', 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'gradient_checkpointing': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': 'None', 'push_to_hub_organization': 'None', 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', '_n_gpu': 1, 'mp_parameters': '', 'train_batch_size': 32, 'eval_batch_size': 32}
25
  2022-02-01 22:36:26,509 INFO MainThread:37310 [wandb_watch.py:watch():43] Watching
26
+ 2022-02-02 05:17:25,827 INFO MainThread:37310 [wandb_run.py:_atexit_cleanup():1780] got exitcode: 0
27
+ 2022-02-02 05:17:25,831 INFO MainThread:37310 [wandb_run.py:_restore():1752] restore
28
+ 2022-02-02 05:17:28,379 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
29
+ wandb_count: 1
30
+ }
31
+ pusher_stats {
32
+ uploaded_bytes: 2149
33
+ total_bytes: 2149
34
+ }
35
+
36
+ 2022-02-02 05:17:28,489 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
37
+ wandb_count: 1
38
+ }
39
+ pusher_stats {
40
+ uploaded_bytes: 2149
41
+ total_bytes: 2149
42
+ }
43
+
44
+ 2022-02-02 05:17:29,258 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
45
+ wandb_count: 1
46
+ }
47
+ pusher_stats {
48
+ uploaded_bytes: 2149
49
+ total_bytes: 2149
50
+ }
51
+
52
+ 2022-02-02 05:17:30,283 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
53
+ wandb_count: 4
54
+ }
55
+ pusher_stats {
56
+ uploaded_bytes: 2149
57
+ total_bytes: 2176207
58
+ }
59
+
60
+ 2022-02-02 05:17:30,386 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
61
+ wandb_count: 5
62
+ }
63
+ pusher_stats {
64
+ uploaded_bytes: 2149
65
+ total_bytes: 2179340
66
+ }
67
+
68
+ 2022-02-02 05:17:30,491 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
69
+ wandb_count: 5
70
+ }
71
+ pusher_stats {
72
+ uploaded_bytes: 2149
73
+ total_bytes: 2179340
74
+ }
75
+
76
+ 2022-02-02 05:17:30,594 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
77
+ wandb_count: 5
78
+ }
79
+ pusher_stats {
80
+ uploaded_bytes: 2176207
81
+ total_bytes: 2179340
82
+ }
83
+
84
+ 2022-02-02 05:17:30,697 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
85
+ wandb_count: 5
86
+ }
87
+ pusher_stats {
88
+ uploaded_bytes: 2176207
89
+ total_bytes: 2179340
90
+ }
91
+
92
+ 2022-02-02 05:17:30,800 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
93
+ wandb_count: 5
94
+ }
95
+ pusher_stats {
96
+ uploaded_bytes: 2179340
97
+ total_bytes: 2179340
98
+ }
99
+
100
+ 2022-02-02 05:17:30,903 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
101
+ wandb_count: 5
102
+ }
103
+ pusher_stats {
104
+ uploaded_bytes: 2179340
105
+ total_bytes: 2179340
106
+ }
107
+
108
+ 2022-02-02 05:17:31,006 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
109
+ wandb_count: 5
110
+ }
111
+ pusher_stats {
112
+ uploaded_bytes: 2179340
113
+ total_bytes: 2179340
114
+ }
115
+
116
+ 2022-02-02 05:17:31,111 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
117
+ wandb_count: 5
118
+ }
119
+ pusher_stats {
120
+ uploaded_bytes: 2179340
121
+ total_bytes: 2179340
122
+ }
123
+
124
+ 2022-02-02 05:17:31,215 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
125
+ wandb_count: 5
126
+ }
127
+ pusher_stats {
128
+ uploaded_bytes: 2179340
129
+ total_bytes: 2179340
130
+ }
131
+
132
+ 2022-02-02 05:17:32,130 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: file_counts {
133
+ wandb_count: 5
134
+ }
135
+ pusher_stats {
136
+ uploaded_bytes: 2179340
137
+ total_bytes: 2179340
138
+ }
139
+
140
+ 2022-02-02 05:17:32,547 INFO MainThread:37310 [wandb_run.py:_wait_for_finish():1912] got exit ret: done: true
141
+ exit_result {
142
+ }
143
+ file_counts {
144
+ wandb_count: 5
145
+ }
146
+ pusher_stats {
147
+ uploaded_bytes: 2179340
148
+ total_bytes: 2179340
149
+ }
150
+ local_info {
151
+ }
152
+
153
+ 2022-02-02 05:17:33,722 INFO MainThread:37310 [wandb_run.py:_append_history():2130] rendering history
154
+ 2022-02-02 05:17:33,724 INFO MainThread:37310 [wandb_run.py:_append_summary():2085] rendering summary
155
+ 2022-02-02 05:17:33,726 INFO MainThread:37310 [wandb_run.py:_append_files():2180] logging synced files
wandb/run-20220201_223624-2b1hcyq3/run-2b1hcyq3.wandb CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1f386743d47e668fa7dece2d08651b0b861610e6786ea118035ce7e2d03c2e86
3
- size 77278807
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1aafe6fd34e4c360ead35aa95b2e51a9f57c14909dbdc54d0f630f197ff3eda8
3
+ size 77927361