File size: 32,549 Bytes
dbec483
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a80f928
dbec483
 
a80f928
dbec483
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
#!/usr/bin/env python
# coding=utf-8
# Copyright 2021 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and

""" Fine-tuning a 🤗 Transformers CTC model for automatic speech recognition"""

import functools
import json
import logging
import os
import re
import sys
import warnings
from dataclasses import dataclass, field
from typing import Dict, List, Optional, Union

import datasets
import numpy as np
import torch
from datasets import DatasetDict, load_dataset, load_metric

import transformers
from transformers import (
    AutoConfig,
    AutoFeatureExtractor,
    AutoModelForCTC,
    AutoProcessor,
    AutoTokenizer,
    HfArgumentParser,
    Trainer,
    TrainingArguments,
    Wav2Vec2Processor,
    set_seed,
)
from transformers.trainer_utils import get_last_checkpoint, is_main_process
from transformers.utils import check_min_version
from transformers.utils.versions import require_version

import string

# Will error if the minimal version of Transformers is not installed. Remove at your own risks.
check_min_version("4.17.0.dev0")

require_version("datasets>=1.13.3", "To fix: pip install -r examples/pytorch/text-classification/requirements.txt")


logger = logging.getLogger(__name__)


def list_field(default=None, metadata=None):
    return field(default_factory=lambda: default, metadata=metadata)


@dataclass
class ModelArguments:
    """
    Arguments pertaining to which model/config/tokenizer we are going to fine-tune from.
    """

    model_name_or_path: str = field(
        metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models"}
    )
    tokenizer_name_or_path: Optional[str] = field(
        default=None,
        metadata={"help": "Path to pretrained tokenizer or tokenizer identifier from huggingface.co/models"},
    )
    cache_dir: Optional[str] = field(
        default=None,
        metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"},
    )
    freeze_feature_encoder: bool = field(
        default=True, metadata={"help": "Whether to freeze the feature encoder layers of the model."}
    )
    attention_dropout: float = field(
        default=0.0, metadata={"help": "The dropout ratio for the attention probabilities."}
    )
    activation_dropout: float = field(
        default=0.0, metadata={"help": "The dropout ratio for activations inside the fully connected layer."}
    )
    feat_proj_dropout: float = field(default=0.0, metadata={"help": "The dropout ratio for the projected features."})
    hidden_dropout: float = field(
        default=0.0,
        metadata={
            "help": "The dropout probability for all fully connected layers in the embeddings, encoder, and pooler."
        },
    )
    final_dropout: float = field(
        default=0.0,
        metadata={"help": "The dropout probability for the final projection layer."},
    )
    mask_time_prob: float = field(
        default=0.05,
        metadata={
            "help": "Probability of each feature vector along the time axis to be chosen as the start of the vector"
            "span to be masked. Approximately ``mask_time_prob * sequence_length // mask_time_length`` feature"
            "vectors will be masked along the time axis."
        },
    )
    mask_time_length: int = field(
        default=10,
        metadata={"help": "Length of vector span to mask along the time axis."},
    )
    mask_feature_prob: float = field(
        default=0.0,
        metadata={
            "help": "Probability of each feature vector along the feature axis to be chosen as the start of the vector"
            "span to be masked. Approximately ``mask_feature_prob * sequence_length // mask_feature_length`` feature bins will be masked along the time axis."
        },
    )
    mask_feature_length: int = field(
        default=10,
        metadata={"help": "Length of vector span to mask along the feature axis."},
    )
    layerdrop: float = field(default=0.0, metadata={"help": "The LayerDrop probability."})
    ctc_loss_reduction: Optional[str] = field(
        default="mean", metadata={"help": "The way the ctc loss should be reduced. Should be one of 'mean' or 'sum'."}
    )


@dataclass
class DataTrainingArguments:
    """
    Arguments pertaining to what data we are going to input our model for training and eval.

    Using `HfArgumentParser` we can turn this class
    into argparse arguments to be able to specify them on
    the command line.
    """

    dataset_name: str = field(
        metadata={"help": "The configuration name of the dataset to use (via the datasets library)."}
    )
    dataset_config_name: str = field(
        default=None, metadata={"help": "The configuration name of the dataset to use (via the datasets library)."}
    )
    train_split_name: str = field(
        default="train+validation",
        metadata={
            "help": "The name of the training data set split to use (via the datasets library). Defaults to 'train+validation'"
        },
    )
    eval_split_name: str = field(
        default="test",
        metadata={
            "help": "The name of the evaluation data set split to use (via the datasets library). Defaults to 'test'"
        },
    )
    audio_column_name: str = field(
        default="audio",
        metadata={"help": "The name of the dataset column containing the audio data. Defaults to 'audio'"},
    )
    text_column_name: str = field(
        default="text",
        metadata={"help": "The name of the dataset column containing the text data. Defaults to 'text'"},
    )
    overwrite_cache: bool = field(
        default=False, metadata={"help": "Overwrite the cached preprocessed datasets or not."}
    )
    preprocessing_num_workers: Optional[int] = field(
        default=None,
        metadata={"help": "The number of processes to use for the preprocessing."},
    )
    max_train_samples: Optional[int] = field(
        default=None,
        metadata={
            "help": "For debugging purposes or quicker training, truncate the number of training examples to this "
            "value if set."
        },
    )
    max_eval_samples: Optional[int] = field(
        default=None,
        metadata={
            "help": "For debugging purposes or quicker training, truncate the number of validation examples to this "
            "value if set."
        },
    )
    chars_to_ignore: List[str] = list_field(
        default=[",", "?", "¿", ".", "!", "¡", ";", ";", ":", '""', "%", '"', "�", "ʿ", "·", "჻", "~", "՞",
                 "؟", "،", "।", "॥", "«", "»", "„", "“", "”", "「", "」", "‘", "’", "《", "》", "(", ")", "[", "]",
                 "{", "}", "=", "`", "_", "+", "<", ">", "…", "–", "°", "´", "ʾ", "‹", "›", "©", "®", "—", "→", "。",
                 "、", "﹂", "﹁", "‧", "~", "﹏", ",", "{", "}", "(", ")", "[", "]", "【", "】", "‥", "〽",
                 "『", "』", "〝", "〟", "⟨", "⟩", "〜", ":", "!", "?", "♪", "؛", "/", "\\", "º", "−", "^", "ʻ", "ˆ"],
        metadata={"help": "A list of characters to remove from the transcripts."},
    )
    eval_metrics: List[str] = list_field(
        default=["wer"],
        metadata={"help": "A list of metrics the model should be evaluated on. E.g. `'wer cer'`"},
    )
    max_duration_in_seconds: float = field(
        default=20.0,
        metadata={
            "help": "Filter audio files that are longer than `max_duration_in_seconds` seconds to 'max_duration_in_seconds`"
        },
    )
    min_duration_in_seconds: float = field(
        default=0.0, metadata={"help": "Filter audio files that are shorter than `min_duration_in_seconds` seconds"}
    )
    preprocessing_only: bool = field(
        default=False,
        metadata={
            "help": "Whether to only do data preprocessing and skip training. "
            "This is especially useful when data preprocessing errors out in distributed training due to timeout. "
            "In this case, one should run the preprocessing in a non-distributed setup with `preprocessing_only=True` "
            "so that the cached datasets can consequently be loaded in distributed training"
        },
    )
    use_auth_token: bool = field(
        default=False,
        metadata={
            "help": "If :obj:`True`, will use the token generated when running"
            ":obj:`transformers-cli login` as HTTP bearer authorization for remote files."
        },
    )
    unk_token: str = field(
        default="[UNK]",
        metadata={"help": "The unk token for the tokenizer"},
    )
    pad_token: str = field(
        default="[PAD]",
        metadata={"help": "The padding token for the tokenizer"},
    )
    word_delimiter_token: str = field(
        default="|",
        metadata={"help": "The word delimiter token for the tokenizer"},
    )
    phoneme_language: Optional[str] = field(
        default=None,
        metadata={
            "help": "The target language that should be used be"
            " passed to the tokenizer for tokenization. Note that"
            " this is only relevant if the model classifies the"
            " input audio to a sequence of phoneme sequences."
        },
    )


@dataclass
class DataCollatorCTCWithPadding:
    """
    Data collator that will dynamically pad the inputs received.
    Args:
        processor (:class:`~transformers.AutoProcessor`)
            The processor used for proccessing the data.
        padding (:obj:`bool`, :obj:`str` or :class:`~transformers.tokenization_utils_base.PaddingStrategy`, `optional`, defaults to :obj:`True`):
            Select a strategy to pad the returned sequences (according to the model's padding side and padding index)
            among:
            * :obj:`True` or :obj:`'longest'`: Pad to the longest sequence in the batch (or no padding if only a single
              sequence if provided).
            * :obj:`'max_length'`: Pad to a maximum length specified with the argument :obj:`max_length` or to the
              maximum acceptable input length for the model if that argument is not provided.
            * :obj:`False` or :obj:`'do_not_pad'` (default): No padding (i.e., can output a batch with sequences of
              different lengths).
        max_length (:obj:`int`, `optional`):
            Maximum length of the ``input_values`` of the returned list and optionally padding length (see above).
        max_length_labels (:obj:`int`, `optional`):
            Maximum length of the ``labels`` returned list and optionally padding length (see above).
        pad_to_multiple_of (:obj:`int`, `optional`):
            If set will pad the sequence to a multiple of the provided value.
            This is especially useful to enable the use of Tensor Cores on NVIDIA hardware with compute capability >=
            7.5 (Volta).
    """

    processor: AutoProcessor
    padding: Union[bool, str] = "longest"
    pad_to_multiple_of: Optional[int] = None
    pad_to_multiple_of_labels: Optional[int] = None

    def __call__(self, features: List[Dict[str, Union[List[int], torch.Tensor]]]) -> Dict[str, torch.Tensor]:
        # split inputs and labels since they have to be of different lenghts and need
        # different padding methods
        input_features = [{"input_values": feature["input_values"]} for feature in features]
        label_features = [{"input_ids": feature["labels"]} for feature in features]

        batch = self.processor.pad(
            input_features,
            padding=self.padding,
            pad_to_multiple_of=self.pad_to_multiple_of,
            return_tensors="pt",
        )

        with self.processor.as_target_processor():
            labels_batch = self.processor.pad(
                label_features,
                padding=self.padding,
                pad_to_multiple_of=self.pad_to_multiple_of_labels,
                return_tensors="pt",
            )

        # replace padding with -100 to ignore loss correctly
        labels = labels_batch["input_ids"].masked_fill(labels_batch.attention_mask.ne(1), -100)

        batch["labels"] = labels

        return batch


def create_vocabulary_from_data(
    datasets: DatasetDict,
    word_delimiter_token: Optional[str] = None,
    unk_token: Optional[str] = None,
    pad_token: Optional[str] = None,
):
    # Given training and test labels create vocabulary
    def extract_all_chars(batch):
        all_text = " ".join(batch["target_text"])
        vocab = list(set(all_text))
        return {"vocab": [vocab], "all_text": [all_text]}

    vocabs = datasets.map(
        extract_all_chars,
        batched=True,
        batch_size=-1,
        keep_in_memory=True,
        remove_columns=datasets["train"].column_names,
    )

    # take union of all unique characters in each dataset
    vocab_set = functools.reduce(
        lambda vocab_1, vocab_2: set(vocab_1["vocab"][0]) | set(vocab_2["vocab"][0]), vocabs.values()
    )

    vocab_dict = {v: k for k, v in enumerate(sorted(list(vocab_set)))}

    # replace white space with delimiter token
    if word_delimiter_token is not None:
        vocab_dict[word_delimiter_token] = vocab_dict[" "]
        del vocab_dict[" "]

    # add unk and pad token
    if unk_token is not None:
        vocab_dict[unk_token] = len(vocab_dict)

    if pad_token is not None:
        vocab_dict[pad_token] = len(vocab_dict)

    return vocab_dict


def main():
    # See all possible arguments in src/transformers/training_args.py
    # or by passing the --help flag to this script.
    # We now keep distinct sets of args, for a cleaner separation of concerns.

    parser = HfArgumentParser((ModelArguments, DataTrainingArguments, TrainingArguments))
    if len(sys.argv) == 2 and sys.argv[1].endswith(".json"):
        # If we pass only one argument to the script and it's the path to a json file,
        # let's parse it to get our arguments.
        model_args, data_args, training_args = parser.parse_json_file(json_file=os.path.abspath(sys.argv[1]))
    else:
        model_args, data_args, training_args = parser.parse_args_into_dataclasses()

    # Detecting last checkpoint.
    last_checkpoint = None
    if os.path.isdir(training_args.output_dir) and training_args.do_train and not training_args.overwrite_output_dir:
        last_checkpoint = get_last_checkpoint(training_args.output_dir)
        if last_checkpoint is None and len(os.listdir(training_args.output_dir)) > 0:
            raise ValueError(
                f"Output directory ({training_args.output_dir}) already exists and is not empty. "
                "Use --overwrite_output_dir to overcome."
            )
        elif last_checkpoint is not None:
            logger.info(
                f"Checkpoint detected, resuming training at {last_checkpoint}. To avoid this behavior, change "
                "the `--output_dir` or add `--overwrite_output_dir` to train from scratch."
            )

    # Setup logging
    logging.basicConfig(
        format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
        datefmt="%m/%d/%Y %H:%M:%S",
        handlers=[logging.StreamHandler(sys.stdout)],
    )
    logger.setLevel(logging.INFO if is_main_process(training_args.local_rank) else logging.WARN)

    # Log on each process the small summary:
    logger.warning(
        f"Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu}"
        f"distributed training: {bool(training_args.local_rank != -1)}, 16-bits training: {training_args.fp16}"
    )
    # Set the verbosity to info of the Transformers logger (on main process only):
    if is_main_process(training_args.local_rank):
        transformers.utils.logging.set_verbosity_info()
    logger.info("Training/evaluation parameters %s", training_args)

    # Set seed before initializing model.
    set_seed(training_args.seed)

    # 1. First, let's load the dataset
    raw_datasets = DatasetDict()

    if training_args.do_train:
        raw_datasets["train"]  = load_dataset(
            data_args.dataset_name,
            data_args.dataset_config_name,
            split=data_args.train_split_name,
            use_auth_token=data_args.use_auth_token,
        )

        if data_args.audio_column_name not in raw_datasets["train"].column_names:
            raise ValueError(
                f"--audio_column_name '{data_args.audio_column_name}' not found in dataset '{data_args.dataset_name}'. "
                "Make sure to set `--audio_column_name` to the correct audio column - one of "
                f"{', '.join(raw_datasets['train'].column_names)}."
            )

        if data_args.text_column_name not in raw_datasets["train"].column_names:
            raise ValueError(
                f"--text_column_name {data_args.text_column_name} not found in dataset '{data_args.dataset_name}'. "
                "Make sure to set `--text_column_name` to the correct text column - one of "
                f"{', '.join(raw_datasets['train'].column_names)}."
            )

        if data_args.max_train_samples is not None:
            raw_datasets["train"] = raw_datasets["train"].select(range(data_args.max_train_samples))

    if training_args.do_eval:
        raw_datasets["eval"] = load_dataset(
            data_args.dataset_name,
            data_args.dataset_config_name,
            split=data_args.eval_split_name,
            use_auth_token=data_args.use_auth_token,
        )

        if data_args.max_eval_samples is not None:
            raw_datasets["eval"] = raw_datasets["eval"].select(range(data_args.max_eval_samples))

    # 2. We remove some special characters from the datasets
    # that make training complicated and do not help in transcribing the speech
    # E.g. characters, such as `,` and `.` do not really have an acoustic characteristic
    # that could be easily picked up by the model
    chars_to_ignore_regex = (
        f'[{"".join(data_args.chars_to_ignore)}]' if data_args.chars_to_ignore is not None else None
    )
    text_column_name = data_args.text_column_name

    def remove_special_characters(batch):
        if chars_to_ignore_regex is not None:
            sen = re.sub(chars_to_ignore_regex, "", batch[text_column_name]).lower() + " "
            
            # convert 'D' and 'd' to '啲' if there a 'D' in sentence
            # hacky stuff, wont work on 'D', 'd' co-occure with normal english words
            # wont work on multiple 'D'
            if "d" in sen:
                if len([c for c in sen if c in string.ascii_lowercase]) == 1:
                    sen = sen.replace("d", "啲")
            batch["target_text"] = sen    
        else:
            batch["target_text"] = batch[text_column_name].lower() + " "
        return batch

    with training_args.main_process_first(desc="dataset map special characters removal"):
        raw_datasets = raw_datasets.map(
            remove_special_characters,
            remove_columns=[text_column_name],
            desc="remove special characters from datasets",
        )

    # save special tokens for tokenizer
    word_delimiter_token = data_args.word_delimiter_token
    unk_token = data_args.unk_token
    pad_token = data_args.pad_token

    # 3. Next, let's load the config as we might need it to create
    # the tokenizer
    # load config
    config = AutoConfig.from_pretrained(
        model_args.model_name_or_path, cache_dir=model_args.cache_dir, use_auth_token=data_args.use_auth_token
    )

    # 4. Next, if no tokenizer file is defined,
    # we create the vocabulary of the model by extracting all unique characters from
    # the training and evaluation datasets
    # We need to make sure that only first rank saves vocabulary
    # make sure all processes wait until vocab is created
    tokenizer_name_or_path = model_args.tokenizer_name_or_path
    tokenizer_kwargs = {}
    if tokenizer_name_or_path is None:
        # save vocab in training output dir
        tokenizer_name_or_path = training_args.output_dir

        vocab_file = os.path.join(tokenizer_name_or_path, "vocab.json")

        with training_args.main_process_first():
            if training_args.overwrite_output_dir and os.path.isfile(vocab_file):
                os.remove(vocab_file)

        with training_args.main_process_first(desc="dataset map vocabulary creation"):
            if not os.path.isfile(vocab_file):
                os.makedirs(tokenizer_name_or_path, exist_ok=True)
                vocab_dict = create_vocabulary_from_data(
                    raw_datasets,
                    word_delimiter_token=word_delimiter_token,
                    unk_token=unk_token,
                    pad_token=pad_token,
                )

                # save vocab dict to be loaded into tokenizer
                with open(vocab_file, "w") as file:
                    json.dump(vocab_dict, file)

        # if tokenizer has just been created
        # it is defined by `tokenizer_class` if present in config else by `model_type`
        tokenizer_kwargs = {
            "config": config if config.tokenizer_class is not None else None,
            "tokenizer_type": config.model_type if config.tokenizer_class is None else None,
            "unk_token": unk_token,
            "pad_token": pad_token,
            "word_delimiter_token": word_delimiter_token,
        }

    # 5. Now we can instantiate the feature extractor, tokenizer and model
    # Note for distributed training, the .from_pretrained methods guarantee that only
    # one local process can concurrently download model & vocab.

    # load feature_extractor and tokenizer
    tokenizer = AutoTokenizer.from_pretrained(
        tokenizer_name_or_path,
        use_auth_token=data_args.use_auth_token,
        **tokenizer_kwargs,
    )
    feature_extractor = AutoFeatureExtractor.from_pretrained(
        model_args.model_name_or_path, cache_dir=model_args.cache_dir, use_auth_token=data_args.use_auth_token
    )

    # adapt config
    config.update(
        {
            "feat_proj_dropout": model_args.feat_proj_dropout,
            "attention_dropout": model_args.attention_dropout,
            "hidden_dropout": model_args.hidden_dropout,
            "final_dropout": model_args.final_dropout,
            "mask_time_prob": model_args.mask_time_prob,
            "mask_time_length": model_args.mask_time_length,
            "mask_feature_prob": model_args.mask_feature_prob,
            "mask_feature_length": model_args.mask_feature_length,
            "gradient_checkpointing": training_args.gradient_checkpointing,
            "layerdrop": model_args.layerdrop,
            "ctc_loss_reduction": model_args.ctc_loss_reduction,
            "pad_token_id": tokenizer.pad_token_id,
            "vocab_size": len(tokenizer),
            "activation_dropout": model_args.activation_dropout,
        }
    )

    # create model
    model = AutoModelForCTC.from_pretrained(
        model_args.model_name_or_path,
        cache_dir=model_args.cache_dir,
        config=config,
        use_auth_token=data_args.use_auth_token,
    )

    # freeze encoder
    if model_args.freeze_feature_encoder:
        model.freeze_feature_encoder()

    # 6. Now we preprocess the datasets including remove long audio sample, loading the audio, resampling and normalization
    # Thankfully, `datasets` takes care of automatically loading and resampling the audio,
    # so that we just need to set the correct target sampling rate and normalize the input
    # via the `feature_extractor`

    # make sure that dataset decodes audio with correct sampling rate
    dataset_sampling_rate = next(iter(raw_datasets.values())).features[data_args.audio_column_name].sampling_rate
    
    # print("data sample rate:", dataset_sampling_rate) # 48_000
    # print("feature sample rate:", feature_extractor.sampling_rate) # 16_000
    
    # # remove long common voice
    # def remove_long_common_voicedata(dataset, max_seconds=6):
    #     #convert pyarrow table to pandas
    #     dftest= dataset.to_pandas()

    #     #find out length of input_values
    #     dftest['len']= dftest['target_text'].apply(len)

    #     #for wav2vec training we already resampled to 16khz
    #     #remove data that is longer than max_seconds (6 seconds ideal)
    #     maxLength = max_seconds * 16000 
    #     dftest= dftest[dftest['len']<maxLength]
    #     dftest = dftest.drop('len', 1)

    #     #convert back to pyarrow table to use in trainer
    #     dataset= dataset.from_pandas(dftest)

    #     #directly remove do not wait for gc
    #     del dftest
    #     return dataset

    # raw_datasets['train'] = remove_long_common_voicedata(raw_datasets['train'], max_seconds=3)
    # raw_datasets['eval'] = remove_long_common_voicedata(raw_datasets['eval'], max_seconds=3)


    # casting column
    if dataset_sampling_rate != feature_extractor.sampling_rate:
        raw_datasets = raw_datasets.cast_column(
            data_args.audio_column_name, datasets.features.Audio(sampling_rate=feature_extractor.sampling_rate)
        )


    # derive max & min input length for sample rate & max duration
    max_input_length = data_args.max_duration_in_seconds * feature_extractor.sampling_rate
    min_input_length = data_args.min_duration_in_seconds * feature_extractor.sampling_rate
    audio_column_name = data_args.audio_column_name
    num_workers = data_args.preprocessing_num_workers

    # `phoneme_language` is only relevant if the model is fine-tuned on phoneme classification
    phoneme_language = data_args.phoneme_language

    # Preprocessing the datasets.
    # We need to read the audio files as arrays and tokenize the targets.
    def prepare_dataset(batch):
        # load audio
        sample = batch[audio_column_name]

        inputs = feature_extractor(sample["array"], sampling_rate=sample["sampling_rate"])
        batch["input_values"] = inputs.input_values[0]
        batch["input_length"] = len(batch["input_values"])

        # encode targets
        additional_kwargs = {}
        if phoneme_language is not None:
            additional_kwargs["phonemizer_lang"] = phoneme_language

        batch["labels"] = tokenizer(batch["target_text"], **additional_kwargs).input_ids
        return batch

    with training_args.main_process_first(desc="dataset map preprocessing"):
        vectorized_datasets = raw_datasets.map(
            prepare_dataset,
            remove_columns=next(iter(raw_datasets.values())).column_names,
            num_proc=num_workers,
            desc="preprocess datasets",
        )

        def is_audio_in_length_range(length):
            return length > min_input_length and length < max_input_length

        # filter data that is shorter than min_input_length
        vectorized_datasets = vectorized_datasets.filter(
            is_audio_in_length_range,
            num_proc=num_workers,
            input_columns=["input_length"],
        )

    # 7. Next, we can prepare the training.
    # Let's use word error rate (WER) as our evaluation metric,
    # instantiate a data collator and the trainer

    # Define evaluation metrics during training, *i.e.* word error rate, character error rate
    eval_metrics = {metric: load_metric(metric) for metric in data_args.eval_metrics}

    # for large datasets it is advised to run the preprocessing on a
    # single machine first with ``args.preprocessing_only`` since there will mostly likely
    # be a timeout when running the script in distributed mode.
    # In a second step ``args.preprocessing_only`` can then be set to `False` to load the
    # cached dataset
    if data_args.preprocessing_only:
        logger.info(f"Data preprocessing finished. Files cached at {vectorized_datasets.cache_files}")
        return

    def compute_metrics(pred):
        pred_logits = pred.predictions
        pred_ids = np.argmax(pred_logits, axis=-1)

        pred.label_ids[pred.label_ids == -100] = tokenizer.pad_token_id

        pred_str = tokenizer.batch_decode(pred_ids)
        # we do not want to group tokens when computing the metrics
        label_str = tokenizer.batch_decode(pred.label_ids, group_tokens=False)

        metrics = {k: v.compute(predictions=pred_str, references=label_str) for k, v in eval_metrics.items()}

        return metrics

    # Now save everything to be able to create a single processor later
    if is_main_process(training_args.local_rank):
        # save feature extractor, tokenizer and config
        feature_extractor.save_pretrained(training_args.output_dir)
        tokenizer.save_pretrained(training_args.output_dir)
        config.save_pretrained(training_args.output_dir)

    try:
        processor = AutoProcessor.from_pretrained(training_args.output_dir)
    except (OSError, KeyError):
        warnings.warn(
            "Loading a processor from a feature extractor config that does not"
            " include a `processor_class` attribute is deprecated and will be removed in v5. Please add the following "
            " attribute to your `preprocessor_config.json` file to suppress this warning: "
            " `'processor_class': 'Wav2Vec2Processor'`",
            FutureWarning,
        )
        processor = Wav2Vec2Processor.from_pretrained(training_args.output_dir)

    # Instantiate custom data collator
    data_collator = DataCollatorCTCWithPadding(processor=processor)

    # Initialize Trainer
    trainer = Trainer(
        model=model,
        data_collator=data_collator,
        args=training_args,
        compute_metrics=compute_metrics,
        train_dataset=vectorized_datasets["train"] if training_args.do_train else None,
        eval_dataset=vectorized_datasets["eval"] if training_args.do_eval else None,
        tokenizer=feature_extractor,
    )

    # 8. Finally, we can start training

    # Training
    if training_args.do_train:

        # use last checkpoint if exist
        if last_checkpoint is not None:
            checkpoint = last_checkpoint
        elif os.path.isdir(model_args.model_name_or_path):
            checkpoint = model_args.model_name_or_path
        else:
            checkpoint = None

        train_result = trainer.train(resume_from_checkpoint=checkpoint)
        trainer.save_model()

        metrics = train_result.metrics
        max_train_samples = (
            data_args.max_train_samples
            if data_args.max_train_samples is not None
            else len(vectorized_datasets["train"])
        )
        metrics["train_samples"] = min(max_train_samples, len(vectorized_datasets["train"]))

        trainer.log_metrics("train", metrics)
        trainer.save_metrics("train", metrics)
        trainer.save_state()

    # Evaluation
    results = {}
    if training_args.do_eval:
        logger.info("*** Evaluate ***")
        metrics = trainer.evaluate()
        max_eval_samples = (
            data_args.max_eval_samples if data_args.max_eval_samples is not None else len(vectorized_datasets["eval"])
        )
        metrics["eval_samples"] = min(max_eval_samples, len(vectorized_datasets["eval"]))

        trainer.log_metrics("eval", metrics)
        trainer.save_metrics("eval", metrics)

    # Write model card and (optionally) push to hub
    config_name = data_args.dataset_config_name if data_args.dataset_config_name is not None else "na"
    kwargs = {
        "finetuned_from": model_args.model_name_or_path,
        "tasks": "speech-recognition",
        "tags": ["automatic-speech-recognition", data_args.dataset_name],
        "dataset_args": f"Config: {config_name}, Training split: {data_args.train_split_name}, Eval split: {data_args.eval_split_name}",
        "dataset": f"{data_args.dataset_name.upper()} - {config_name.upper()}",
    }
    if "common_voice" in data_args.dataset_name:
        kwargs["language"] = config_name

    if training_args.push_to_hub:
        trainer.push_to_hub(**kwargs)
    else:
        trainer.create_model_card(**kwargs)

    return results


if __name__ == "__main__":
    main()