jonatasgrosman commited on
Commit
69edc61
1 Parent(s): 5fb6f8f
README.md ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - fr
4
+ license: apache-2.0
5
+ tags:
6
+ - automatic-speech-recognition
7
+ - mozilla-foundation/common_voice_8_0
8
+ - fr
9
+ - robust-speech-event
10
+ datasets:
11
+ - mozilla-foundation/common_voice_8_0
12
+ model-index:
13
+ - name: XLS-R-1B - French
14
+ results:
15
+ - task:
16
+ name: Automatic Speech Recognition
17
+ type: automatic-speech-recognition
18
+ dataset:
19
+ name: Common Voice 8
20
+ type: mozilla-foundation/common_voice_8_0
21
+ args: fr
22
+ metrics:
23
+ - name: Test WER
24
+ type: wer
25
+ value: 16.87
26
+ - name: Test CER
27
+ type: cer
28
+ value: 4.67
29
+ - name: Test WER (+LM)
30
+ type: wer
31
+ value: 15.18
32
+ - name: Test CER (+LM)
33
+ type: cer
34
+ value: 4.28
35
+ ---
36
+
37
+ # XLS-R-1B-FRENCH
38
+
39
+ This model is a fine-tuned version of [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - FR dataset.
40
+
41
+
42
+ ## Evaluation Commands
43
+
44
+ 1. To evaluate on `mozilla-foundation/common_voice_8_0` with split `test`
45
+
46
+ ```bash
47
+ python eval.py --model_id jonatasgrosman/wav2vec2-xls-r-1b-french --dataset mozilla-foundation/common_voice_8_0 --config fr --split test
48
+ ```
49
+
50
+ 2. To evaluate on `speech-recognition-community-v2/dev_data`
51
+
52
+ ```bash
53
+ python eval.py --model_id jonatasgrosman/wav2vec2-xls-r-1b-french --dataset speech-recognition-community-v2/dev_data --config fr --split validation --chunk_length_s 5.0 --stride_length_s 1.0
54
+ ```
alphabet.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"labels": ["", "<s>", "</s>", "\u2047", " ", "'", "-", "a", "b", "c", "d", "e", "f", "g", "h", "i", "j", "k", "l", "m", "n", "o", "p", "q", "r", "s", "t", "u", "v", "w", "x", "y", "z", "\u00e0", "\u00e2", "\u00e3", "\u00e7", "\u00e8", "\u00e9", "\u00ea", "\u00eb", "\u00ee", "\u00ef", "\u00f4", "\u00f9", "\u00fb", "\u0153"], "is_bpe": false}
config.json ADDED
@@ -0,0 +1,107 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-xls-r-1b",
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": 0.05,
12
+ "bos_token_id": 1,
13
+ "classifier_proj_size": 256,
14
+ "codevector_dim": 1024,
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": 0.05,
53
+ "feat_quantizer_dropout": 0.0,
54
+ "final_dropout": 0.05,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.05,
57
+ "hidden_size": 1280,
58
+ "initializer_range": 0.02,
59
+ "intermediate_size": 5120,
60
+ "layer_norm_eps": 1e-05,
61
+ "layerdrop": 0.05,
62
+ "mask_feature_length": 10,
63
+ "mask_feature_min_masks": 0,
64
+ "mask_feature_prob": 0.0,
65
+ "mask_time_length": 10,
66
+ "mask_time_min_masks": 2,
67
+ "mask_time_prob": 0.05,
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": 48,
77
+ "num_negatives": 100,
78
+ "output_hidden_size": 1280,
79
+ "pad_token_id": 0,
80
+ "proj_codevector_dim": 1024,
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": "4.16.0.dev0",
104
+ "use_weighted_layer_sum": false,
105
+ "vocab_size": 47,
106
+ "xvector_output_dim": 512
107
+ }
eval.py ADDED
@@ -0,0 +1,146 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ from datasets import load_dataset, load_metric, Audio, Dataset
3
+ from transformers import pipeline, AutoFeatureExtractor, AutoTokenizer
4
+ import re
5
+ import torch
6
+ import argparse
7
+ from typing import Dict
8
+
9
+
10
+ def log_results(result: Dataset, args: Dict[str, str]):
11
+ """ DO NOT CHANGE. This function computes and logs the result metrics. """
12
+
13
+ log_outputs = args.log_outputs
14
+ dataset_id = "_".join(args.dataset.split("/") + [args.config, args.split])
15
+
16
+ # load metric
17
+ wer = load_metric("wer")
18
+ cer = load_metric("cer")
19
+
20
+ # compute metrics
21
+ wer_result = wer.compute(references=result["target"], predictions=result["prediction"])
22
+ cer_result = cer.compute(references=result["target"], predictions=result["prediction"])
23
+
24
+ # print & log results
25
+ result_str = (
26
+ f"WER: {wer_result}\n"
27
+ f"CER: {cer_result}"
28
+ )
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
+ def normalize_text(text: str, invalid_chars_regex: str, to_lower: bool) -> str:
52
+ """ DO ADAPT FOR YOUR USE CASE. this function normalizes the target text. """
53
+
54
+ text = text.lower() if to_lower else text.upper()
55
+
56
+ text = re.sub(invalid_chars_regex, " ", text)
57
+
58
+ text = re.sub("\s+", " ", text).strip()
59
+
60
+ return text
61
+
62
+
63
+ def main(args):
64
+ # load dataset
65
+ dataset = load_dataset(args.dataset, args.config, split=args.split, use_auth_token=True)
66
+
67
+ # for testing: only process the first two examples as a test
68
+ # dataset = dataset.select(range(10))
69
+
70
+ # load processor
71
+ feature_extractor = AutoFeatureExtractor.from_pretrained(args.model_id)
72
+ sampling_rate = feature_extractor.sampling_rate
73
+
74
+ # resample audio
75
+ dataset = dataset.cast_column("audio", Audio(sampling_rate=sampling_rate))
76
+
77
+ # load eval pipeline
78
+ if args.device is None:
79
+ args.device = 0 if torch.cuda.is_available() else -1
80
+ asr = pipeline("automatic-speech-recognition", model=args.model_id, device=args.device)
81
+
82
+ # build normalizer config
83
+ tokenizer = AutoTokenizer.from_pretrained(args.model_id)
84
+ tokens = [x for x in tokenizer.convert_ids_to_tokens(range(0, tokenizer.vocab_size))]
85
+ special_tokens = [
86
+ tokenizer.pad_token, tokenizer.word_delimiter_token,
87
+ tokenizer.unk_token, tokenizer.bos_token,
88
+ tokenizer.eos_token,
89
+ ]
90
+ non_special_tokens = [x for x in tokens if x not in special_tokens]
91
+ invalid_chars_regex = f"[^\s{re.escape(''.join(set(non_special_tokens)))}]"
92
+ normalize_to_lower = False
93
+ for token in non_special_tokens:
94
+ if token.isalpha() and token.islower():
95
+ normalize_to_lower = True
96
+ break
97
+
98
+ # map function to decode audio
99
+ def map_to_pred(batch, args=args, asr=asr, invalid_chars_regex=invalid_chars_regex, normalize_to_lower=normalize_to_lower):
100
+ prediction = asr(batch["audio"]["array"], chunk_length_s=args.chunk_length_s, stride_length_s=args.stride_length_s)
101
+
102
+ batch["prediction"] = prediction["text"]
103
+ batch["target"] = normalize_text(batch["sentence"], invalid_chars_regex, normalize_to_lower)
104
+ return batch
105
+
106
+ # run inference on all examples
107
+ result = dataset.map(map_to_pred, remove_columns=dataset.column_names)
108
+
109
+ # compute and log_results
110
+ # do not change function below
111
+ log_results(result, args)
112
+
113
+
114
+ if __name__ == "__main__":
115
+ parser = argparse.ArgumentParser()
116
+
117
+ parser.add_argument(
118
+ "--model_id", type=str, required=True, help="Model identifier. Should be loadable with 🤗 Transformers"
119
+ )
120
+ parser.add_argument(
121
+ "--dataset", type=str, required=True, help="Dataset name to evaluate the `model_id`. Should be loadable with 🤗 Datasets"
122
+ )
123
+ parser.add_argument(
124
+ "--config", type=str, required=True, help="Config of the dataset. *E.g.* `'en'` for Common Voice"
125
+ )
126
+ parser.add_argument(
127
+ "--split", type=str, required=True, help="Split of the dataset. *E.g.* `'test'`"
128
+ )
129
+ parser.add_argument(
130
+ "--chunk_length_s", type=float, default=None, help="Chunk length in seconds. Defaults to None. For long audio files a good value would be 5.0 seconds."
131
+ )
132
+ parser.add_argument(
133
+ "--stride_length_s", type=float, default=None, help="Stride of the audio chunks. Defaults to None. For long audio files a good value would be 1.0 seconds."
134
+ )
135
+ parser.add_argument(
136
+ "--log_outputs", action='store_true', help="If defined, write outputs to log file for analysis."
137
+ )
138
+ parser.add_argument(
139
+ "--device",
140
+ type=int,
141
+ default=None,
142
+ help="The device to run the pipeline on. -1 for CPU (default), 0 for the first GPU and so on.",
143
+ )
144
+ args = parser.parse_args()
145
+
146
+ main(args)
language_model/attrs.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5ffd02e1ceef6517476e72ebe7997ddef7e92d27cb5a23d6695d64c4317d6ad
3
+ size 78
language_model/lm.binary ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:22cc77043d9e39a4fad2221f50114d5e6fdc74d4151586fa71568c331a8e6b36
3
+ size 222726674
language_model/unigrams.txt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:22c371fad05183e06466ccf1770c28c19194a98cced128d22d938ab2c72ca755
3
+ size 2345295
log_mozilla-foundation_common_voice_8_0_fr_test_predictions.txt ADDED
The diff for this file is too large to render. See raw diff
 
log_mozilla-foundation_common_voice_8_0_fr_test_predictions_greedy.txt ADDED
The diff for this file is too large to render. See raw diff
 
log_mozilla-foundation_common_voice_8_0_fr_test_targets.txt ADDED
The diff for this file is too large to render. See raw diff
 
log_mozilla-foundation_common_voice_8_0_fr_test_targets_greedy.txt ADDED
The diff for this file is too large to render. See raw diff
 
mozilla-foundation_common_voice_8_0_fr_test_eval_results.txt ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ WER: 0.15185178176037548
2
+ CER: 0.04283279709697798
mozilla-foundation_common_voice_8_0_fr_test_eval_results_greedy.txt ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ WER: 0.16870738000164015
2
+ CER: 0.0467260199509307
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": "Wav2Vec2ProcessorWithLM",
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:c526b57c0fa8c775ae10130211e8e6ea900c5c3b89e10ad13651f2c35326a482
3
+ size 3850553521
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>"}
vocab.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"<pad>": 0, "<s>": 1, "</s>": 2, "<unk>": 3, "|": 4, "'": 5, "-": 6, "a": 7, "b": 8, "c": 9, "d": 10, "e": 11, "f": 12, "g": 13, "h": 14, "i": 15, "j": 16, "k": 17, "l": 18, "m": 19, "n": 20, "o": 21, "p": 22, "q": 23, "r": 24, "s": 25, "t": 26, "u": 27, "v": 28, "w": 29, "x": 30, "y": 31, "z": 32, "à": 33, "â": 34, "ã": 35, "ç": 36, "è": 37, "é": 38, "ê": 39, "ë": 40, "î": 41, "ï": 42, "ô": 43, "ù": 44, "û": 45, "œ": 46}