End of training

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 100.0,
+    "eval_loss": 0.2915233075618744,
+    "eval_runtime": 109.4935,
+    "eval_samples": 2173,
+    "eval_samples_per_second": 19.846,
+    "eval_steps_per_second": 19.846,
+    "eval_wer": 0.24811411329035798,
+    "train_loss": 1.3542817559995148,
+    "train_runtime": 20867.045,
+    "train_samples": 4848,
+    "train_samples_per_second": 23.233,
+    "train_steps_per_second": 0.728
+}

eval.py

@@ -0,0 +1,137 @@
+#!/usr/bin/env python3
+import argparse
+import re
+from typing import Dict
+
+import torch
+from datasets import Audio, Dataset, load_dataset, load_metric
+
+from transformers import AutoFeatureExtractor, pipeline
+
+
+def log_results(result: Dataset, args: Dict[str, str]):
+    """DO NOT CHANGE. This function computes and logs the result metrics."""
+
+    log_outputs = args.log_outputs
+    dataset_id = "_".join(args.dataset.split("/") + [args.config, args.split])
+
+    # load metric
+    wer = load_metric("wer")
+    cer = load_metric("cer")
+
+    # compute metrics
+    wer_result = wer.compute(references=result["target"], predictions=result["prediction"])
+    cer_result = cer.compute(references=result["target"], predictions=result["prediction"])
+
+    # print & log results
+    result_str = f"WER: {wer_result}\n" f"CER: {cer_result}"
+    print(result_str)
+
+    with open(f"{dataset_id}_eval_results.txt", "w") as f:
+        f.write(result_str)
+
+    # log all results in text file. Possibly interesting for analysis
+    if log_outputs is not None:
+        pred_file = f"log_{dataset_id}_predictions.txt"
+        target_file = f"log_{dataset_id}_targets.txt"
+
+        with open(pred_file, "w") as p, open(target_file, "w") as t:
+
+            # mapping function to write output
+            def write_to_file(batch, i):
+                p.write(f"{i}" + "\n")
+                p.write(batch["prediction"] + "\n")
+                t.write(f"{i}" + "\n")
+                t.write(batch["target"] + "\n")
+
+            result.map(write_to_file, with_indices=True)
+
+
+def normalize_text(text: str) -> str:
+    """DO ADAPT FOR YOUR USE CASE. this function normalizes the target text."""
+
+    chars_to_ignore_regex = '[,?.!\-\;\:"“%‘”�—’…–]'  # noqa: W605 IMPORTANT: this should correspond to the chars that were ignored during training
+
+    text = re.sub(chars_to_ignore_regex, "", text.lower())
+
+    # In addition, we can normalize the target text, e.g. removing new lines characters etc...
+    # note that order is important here!
+    token_sequences_to_ignore = ["\n\n", "\n", "   ", "  "]
+
+    for t in token_sequences_to_ignore:
+        text = " ".join(text.split(t))
+
+    return text
+
+
+def main(args):
+    # load dataset
+    dataset = load_dataset(args.dataset, args.config, split=args.split, use_auth_token=True)
+
+    # for testing: only process the first two examples as a test
+    # dataset = dataset.select(range(10))
+
+    # load processor
+    feature_extractor = AutoFeatureExtractor.from_pretrained(args.model_id)
+    sampling_rate = feature_extractor.sampling_rate
+
+    # resample audio
+    dataset = dataset.cast_column("audio", Audio(sampling_rate=sampling_rate))
+
+    # load eval pipeline
+    if args.device is None:
+        args.device = 0 if torch.cuda.is_available() else -1
+    asr = pipeline("automatic-speech-recognition", model=args.model_id, device=args.device)
+
+    # map function to decode audio
+    def map_to_pred(batch):
+        prediction = asr(
+            batch["audio"]["array"], chunk_length_s=args.chunk_length_s, stride_length_s=args.stride_length_s
+        )
+
+        batch["prediction"] = prediction["text"]
+        batch["target"] = normalize_text(batch["sentence"])
+        return batch
+
+    # run inference on all examples
+    result = dataset.map(map_to_pred, remove_columns=dataset.column_names)
+
+    # compute and log_results
+    # do not change function below
+    log_results(result, args)
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+
+    parser.add_argument(
+        "--model_id", type=str, required=True, help="Model identifier. Should be loadable with 🤗 Transformers"
+    )
+    parser.add_argument(
+        "--dataset",
+        type=str,
+        required=True,
+        help="Dataset name to evaluate the `model_id`. Should be loadable with 🤗 Datasets",
+    )
+    parser.add_argument(
+        "--config", type=str, required=True, help="Config of the dataset. *E.g.* `'en'`  for Common Voice"
+    )
+    parser.add_argument("--split", type=str, required=True, help="Split of the dataset. *E.g.* `'test'`")
+    parser.add_argument(
+        "--chunk_length_s", type=float, default=None, help="Chunk length in seconds. Defaults to 5 seconds."
+    )
+    parser.add_argument(
+        "--stride_length_s", type=float, default=None, help="Stride of the audio chunks. Defaults to 1 second."
+    )
+    parser.add_argument(
+        "--log_outputs", action="store_true", help="If defined, write outputs to log file for analysis."
+    )
+    parser.add_argument(
+        "--device",
+        type=int,
+        default=None,
+        help="The device to run the pipeline on. -1 for CPU (default), 0 for the first GPU and so on.",
+    )
+    args = parser.parse_args()
+
+    main(args)

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 100.0,
+    "eval_loss": 0.2915233075618744,
+    "eval_runtime": 109.4935,
+    "eval_samples": 2173,
+    "eval_samples_per_second": 19.846,
+    "eval_steps_per_second": 19.846,
+    "eval_wer": 0.24811411329035798
+}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4fe94f5923d068a79150fed2c12e8ebfd18a585adf3c11016ad4130ef3adecdc
+oid sha256:1a0e6f8ae13f422d3032f643d7b16e4bd10b03aa147373b144ac7c8e8860b495
 size 1262136881

run_speech_recognition_ctc.py

@@ -0,0 +1,737 @@
+#!/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
+
+
+# Will error if the minimal version of Transformers is not installed. Remove at your own risks.
+check_min_version("4.16.0.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'"
+        },
+    )
+    eval_split_name: str = field(
+        default="test",
+        metadata={
+            "help": "The name of the training data set split to use (via the datasets library). Defaults to 'train'"
+        },
+    )
+    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: Optional[List[str]] = list_field(
+        default=None,
+        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:
+            batch["target_text"] = re.sub(chars_to_ignore_regex, "", batch[text_column_name]).lower() + " "
+        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 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
+    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()

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 100.0,
+    "train_loss": 1.3542817559995148,
+    "train_runtime": 20867.045,
+    "train_samples": 4848,
+    "train_samples_per_second": 23.233,
+    "train_steps_per_second": 0.728
+}

trainer_state.json

@@ -0,0 +1,982 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 100.0,
+  "global_step": 15200,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.2399999999999997e-06,
+      "loss": 29.0164,
+      "step": 100
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.5733333333333325e-06,
+      "loss": 19.5282,
+      "step": 200
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 6.906666666666666e-06,
+      "loss": 10.0976,
+      "step": 300
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 9.24e-06,
+      "loss": 7.1314,
+      "step": 400
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 1.1573333333333332e-05,
+      "loss": 6.253,
+      "step": 500
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 1.3906666666666665e-05,
+      "loss": 5.0478,
+      "step": 600
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 1.624e-05,
+      "loss": 4.2473,
+      "step": 700
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 1.857333333333333e-05,
+      "loss": 3.8186,
+      "step": 800
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 2.0906666666666668e-05,
+      "loss": 3.561,
+      "step": 900
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 2.3239999999999998e-05,
+      "loss": 3.4988,
+      "step": 1000
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 2.557333333333333e-05,
+      "loss": 3.4272,
+      "step": 1100
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 2.7906666666666665e-05,
+      "loss": 3.3321,
+      "step": 1200
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 3.024e-05,
+      "loss": 3.2363,
+      "step": 1300
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 3.257333333333333e-05,
+      "loss": 3.1793,
+      "step": 1400
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 3.490666666666666e-05,
+      "loss": 3.1351,
+      "step": 1500
+    },
+    {
+      "epoch": 10.53,
+      "learning_rate": 3.7239999999999996e-05,
+      "loss": 3.0,
+      "step": 1600
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 3.957333333333333e-05,
+      "loss": 2.7348,
+      "step": 1700
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 4.190666666666666e-05,
+      "loss": 2.2213,
+      "step": 1800
+    },
+    {
+      "epoch": 12.5,
+      "learning_rate": 4.4239999999999997e-05,
+      "loss": 1.8715,
+      "step": 1900
+    },
+    {
+      "epoch": 13.16,
+      "learning_rate": 4.657333333333333e-05,
+      "loss": 1.6957,
+      "step": 2000
+    },
+    {
+      "epoch": 13.82,
+      "learning_rate": 4.8906666666666664e-05,
+      "loss": 1.5455,
+      "step": 2100
+    },
+    {
+      "epoch": 14.47,
+      "learning_rate": 5.124e-05,
+      "loss": 1.4168,
+      "step": 2200
+    },
+    {
+      "epoch": 15.13,
+      "learning_rate": 5.3573333333333324e-05,
+      "loss": 1.2693,
+      "step": 2300
+    },
+    {
+      "epoch": 15.79,
+      "learning_rate": 5.590666666666666e-05,
+      "loss": 1.2215,
+      "step": 2400
+    },
+    {
+      "epoch": 16.45,
+      "learning_rate": 5.823999999999999e-05,
+      "loss": 1.172,
+      "step": 2500
+    },
+    {
+      "epoch": 17.11,
+      "learning_rate": 6.0573333333333325e-05,
+      "loss": 1.0927,
+      "step": 2600
+    },
+    {
+      "epoch": 17.76,
+      "learning_rate": 6.290666666666665e-05,
+      "loss": 1.0716,
+      "step": 2700
+    },
+    {
+      "epoch": 18.42,
+      "learning_rate": 6.523999999999999e-05,
+      "loss": 1.0267,
+      "step": 2800
+    },
+    {
+      "epoch": 19.08,
+      "learning_rate": 6.757333333333333e-05,
+      "loss": 1.0305,
+      "step": 2900
+    },
+    {
+      "epoch": 19.74,
+      "learning_rate": 6.990666666666666e-05,
+      "loss": 1.0076,
+      "step": 3000
+    },
+    {
+      "epoch": 19.74,
+      "eval_loss": 0.3274237811565399,
+      "eval_runtime": 108.7735,
+      "eval_samples_per_second": 19.977,
+      "eval_steps_per_second": 19.977,
+      "eval_wer": 0.38060622685502676,
+      "step": 3000
+    },
+    {
+      "epoch": 20.39,
+      "learning_rate": 6.944918032786885e-05,
+      "loss": 0.9555,
+      "step": 3100
+    },
+    {
+      "epoch": 21.05,
+      "learning_rate": 6.887540983606556e-05,
+      "loss": 0.9422,
+      "step": 3200
+    },
+    {
+      "epoch": 21.71,
+      "learning_rate": 6.830163934426229e-05,
+      "loss": 0.9229,
+      "step": 3300
+    },
+    {
+      "epoch": 22.37,
+      "learning_rate": 6.772786885245901e-05,
+      "loss": 0.9312,
+      "step": 3400
+    },
+    {
+      "epoch": 23.03,
+      "learning_rate": 6.715409836065574e-05,
+      "loss": 0.9204,
+      "step": 3500
+    },
+    {
+      "epoch": 23.68,
+      "learning_rate": 6.658032786885245e-05,
+      "loss": 0.8966,
+      "step": 3600
+    },
+    {
+      "epoch": 24.34,
+      "learning_rate": 6.600655737704917e-05,
+      "loss": 0.8867,
+      "step": 3700
+    },
+    {
+      "epoch": 25.0,
+      "learning_rate": 6.543278688524589e-05,
+      "loss": 0.895,
+      "step": 3800
+    },
+    {
+      "epoch": 25.66,
+      "learning_rate": 6.485901639344262e-05,
+      "loss": 0.8397,
+      "step": 3900
+    },
+    {
+      "epoch": 26.32,
+      "learning_rate": 6.428524590163934e-05,
+      "loss": 0.8523,
+      "step": 4000
+    },
+    {
+      "epoch": 26.97,
+      "learning_rate": 6.371147540983606e-05,
+      "loss": 0.831,
+      "step": 4100
+    },
+    {
+      "epoch": 27.63,
+      "learning_rate": 6.313770491803279e-05,
+      "loss": 0.8191,
+      "step": 4200
+    },
+    {
+      "epoch": 28.29,
+      "learning_rate": 6.25639344262295e-05,
+      "loss": 0.8222,
+      "step": 4300
+    },
+    {
+      "epoch": 28.95,
+      "learning_rate": 6.199016393442622e-05,
+      "loss": 0.8067,
+      "step": 4400
+    },
+    {
+      "epoch": 29.61,
+      "learning_rate": 6.141639344262295e-05,
+      "loss": 0.7992,
+      "step": 4500
+    },
+    {
+      "epoch": 30.26,
+      "learning_rate": 6.084262295081967e-05,
+      "loss": 0.8202,
+      "step": 4600
+    },
+    {
+      "epoch": 30.92,
+      "learning_rate": 6.026885245901639e-05,
+      "loss": 0.7756,
+      "step": 4700
+    },
+    {
+      "epoch": 31.58,
+      "learning_rate": 5.969508196721311e-05,
+      "loss": 0.7759,
+      "step": 4800
+    },
+    {
+      "epoch": 32.24,
+      "learning_rate": 5.912131147540983e-05,
+      "loss": 0.7778,
+      "step": 4900
+    },
+    {
+      "epoch": 32.89,
+      "learning_rate": 5.854754098360655e-05,
+      "loss": 0.7528,
+      "step": 5000
+    },
+    {
+      "epoch": 33.55,
+      "learning_rate": 5.797377049180327e-05,
+      "loss": 0.7656,
+      "step": 5100
+    },
+    {
+      "epoch": 34.21,
+      "learning_rate": 5.739999999999999e-05,
+      "loss": 0.7529,
+      "step": 5200
+    },
+    {
+      "epoch": 34.87,
+      "learning_rate": 5.682622950819672e-05,
+      "loss": 0.7371,
+      "step": 5300
+    },
+    {
+      "epoch": 35.53,
+      "learning_rate": 5.625245901639344e-05,
+      "loss": 0.7432,
+      "step": 5400
+    },
+    {
+      "epoch": 36.18,
+      "learning_rate": 5.5678688524590164e-05,
+      "loss": 0.7285,
+      "step": 5500
+    },
+    {
+      "epoch": 36.84,
+      "learning_rate": 5.510491803278688e-05,
+      "loss": 0.7116,
+      "step": 5600
+    },
+    {
+      "epoch": 37.5,
+      "learning_rate": 5.4531147540983605e-05,
+      "loss": 0.7109,
+      "step": 5700
+    },
+    {
+      "epoch": 38.16,
+      "learning_rate": 5.395737704918032e-05,
+      "loss": 0.7084,
+      "step": 5800
+    },
+    {
+      "epoch": 38.82,
+      "learning_rate": 5.3383606557377046e-05,
+      "loss": 0.6951,
+      "step": 5900
+    },
+    {
+      "epoch": 39.47,
+      "learning_rate": 5.280983606557376e-05,
+      "loss": 0.6889,
+      "step": 6000
+    },
+    {
+      "epoch": 39.47,
+      "eval_loss": 0.2824408710002899,
+      "eval_runtime": 108.4203,
+      "eval_samples_per_second": 20.042,
+      "eval_steps_per_second": 20.042,
+      "eval_wer": 0.2941983267041558,
+      "step": 6000
+    },
+    {
+      "epoch": 40.13,
+      "learning_rate": 5.2236065573770486e-05,
+      "loss": 0.6805,
+      "step": 6100
+    },
+    {
+      "epoch": 40.79,
+      "learning_rate": 5.1662295081967203e-05,
+      "loss": 0.7018,
+      "step": 6200
+    },
+    {
+      "epoch": 41.45,
+      "learning_rate": 5.1094262295081965e-05,
+      "loss": 0.6773,
+      "step": 6300
+    },
+    {
+      "epoch": 42.11,
+      "learning_rate": 5.052049180327869e-05,
+      "loss": 0.6828,
+      "step": 6400
+    },
+    {
+      "epoch": 42.76,
+      "learning_rate": 4.9952459016393436e-05,
+      "loss": 0.6875,
+      "step": 6500
+    },
+    {
+      "epoch": 43.42,
+      "learning_rate": 4.937868852459015e-05,
+      "loss": 0.6846,
+      "step": 6600
+    },
+    {
+      "epoch": 44.08,
+      "learning_rate": 4.880491803278688e-05,
+      "loss": 0.6826,
+      "step": 6700
+    },
+    {
+      "epoch": 44.74,
+      "learning_rate": 4.823114754098361e-05,
+      "loss": 0.645,
+      "step": 6800
+    },
+    {
+      "epoch": 45.39,
+      "learning_rate": 4.7657377049180325e-05,
+      "loss": 0.6468,
+      "step": 6900
+    },
+    {
+      "epoch": 46.05,
+      "learning_rate": 4.708360655737705e-05,
+      "loss": 0.6638,
+      "step": 7000
+    },
+    {
+      "epoch": 46.71,
+      "learning_rate": 4.6509836065573765e-05,
+      "loss": 0.6451,
+      "step": 7100
+    },
+    {
+      "epoch": 47.37,
+      "learning_rate": 4.593606557377049e-05,
+      "loss": 0.6396,
+      "step": 7200
+    },
+    {
+      "epoch": 48.03,
+      "learning_rate": 4.5362295081967206e-05,
+      "loss": 0.6643,
+      "step": 7300
+    },
+    {
+      "epoch": 48.68,
+      "learning_rate": 4.478852459016393e-05,
+      "loss": 0.6453,
+      "step": 7400
+    },
+    {
+      "epoch": 49.34,
+      "learning_rate": 4.422049180327869e-05,
+      "loss": 0.6579,
+      "step": 7500
+    },
+    {
+      "epoch": 50.0,
+      "learning_rate": 4.364672131147541e-05,
+      "loss": 0.6433,
+      "step": 7600
+    },
+    {
+      "epoch": 50.66,
+      "learning_rate": 4.307295081967213e-05,
+      "loss": 0.615,
+      "step": 7700
+    },
+    {
+      "epoch": 51.32,
+      "learning_rate": 4.249918032786885e-05,
+      "loss": 0.6422,
+      "step": 7800
+    },
+    {
+      "epoch": 51.97,
+      "learning_rate": 4.192540983606557e-05,
+      "loss": 0.6204,
+      "step": 7900
+    },
+    {
+      "epoch": 52.63,
+      "learning_rate": 4.135163934426229e-05,
+      "loss": 0.6173,
+      "step": 8000
+    },
+    {
+      "epoch": 53.29,
+      "learning_rate": 4.0777868852459014e-05,
+      "loss": 0.597,
+      "step": 8100
+    },
+    {
+      "epoch": 53.95,
+      "learning_rate": 4.020409836065573e-05,
+      "loss": 0.6106,
+      "step": 8200
+    },
+    {
+      "epoch": 54.61,
+      "learning_rate": 3.9630327868852455e-05,
+      "loss": 0.5915,
+      "step": 8300
+    },
+    {
+      "epoch": 55.26,
+      "learning_rate": 3.905655737704917e-05,
+      "loss": 0.5934,
+      "step": 8400
+    },
+    {
+      "epoch": 55.92,
+      "learning_rate": 3.8482786885245895e-05,
+      "loss": 0.5833,
+      "step": 8500
+    },
+    {
+      "epoch": 56.58,
+      "learning_rate": 3.790901639344262e-05,
+      "loss": 0.591,
+      "step": 8600
+    },
+    {
+      "epoch": 57.24,
+      "learning_rate": 3.733524590163934e-05,
+      "loss": 0.5708,
+      "step": 8700
+    },
+    {
+      "epoch": 57.89,
+      "learning_rate": 3.676147540983607e-05,
+      "loss": 0.5774,
+      "step": 8800
+    },
+    {
+      "epoch": 58.55,
+      "learning_rate": 3.6187704918032784e-05,
+      "loss": 0.5838,
+      "step": 8900
+    },
+    {
+      "epoch": 59.21,
+      "learning_rate": 3.561393442622951e-05,
+      "loss": 0.5863,
+      "step": 9000
+    },
+    {
+      "epoch": 59.21,
+      "eval_loss": 0.2699650526046753,
+      "eval_runtime": 108.4786,
+      "eval_samples_per_second": 20.032,
+      "eval_steps_per_second": 20.032,
+      "eval_wer": 0.2734878617473598,
+      "step": 9000
+    },
+    {
+      "epoch": 59.87,
+      "learning_rate": 3.5040163934426225e-05,
+      "loss": 0.5803,
+      "step": 9100
+    },
+    {
+      "epoch": 60.53,
+      "learning_rate": 3.446639344262295e-05,
+      "loss": 0.577,
+      "step": 9200
+    },
+    {
+      "epoch": 61.18,
+      "learning_rate": 3.389262295081967e-05,
+      "loss": 0.5599,
+      "step": 9300
+    },
+    {
+      "epoch": 61.84,
+      "learning_rate": 3.331885245901639e-05,
+      "loss": 0.5712,
+      "step": 9400
+    },
+    {
+      "epoch": 62.5,
+      "learning_rate": 3.274508196721311e-05,
+      "loss": 0.5615,
+      "step": 9500
+    },
+    {
+      "epoch": 63.16,
+      "learning_rate": 3.217131147540984e-05,
+      "loss": 0.5427,
+      "step": 9600
+    },
+    {
+      "epoch": 63.82,
+      "learning_rate": 3.1597540983606554e-05,
+      "loss": 0.5617,
+      "step": 9700
+    },
+    {
+      "epoch": 64.47,
+      "learning_rate": 3.102377049180328e-05,
+      "loss": 0.5602,
+      "step": 9800
+    },
+    {
+      "epoch": 65.13,
+      "learning_rate": 3.0449999999999998e-05,
+      "loss": 0.5325,
+      "step": 9900
+    },
+    {
+      "epoch": 65.79,
+      "learning_rate": 2.987622950819672e-05,
+      "loss": 0.5438,
+      "step": 10000
+    },
+    {
+      "epoch": 66.45,
+      "learning_rate": 2.930245901639344e-05,
+      "loss": 0.5538,
+      "step": 10100
+    },
+    {
+      "epoch": 67.11,
+      "learning_rate": 2.8728688524590163e-05,
+      "loss": 0.5408,
+      "step": 10200
+    },
+    {
+      "epoch": 67.76,
+      "learning_rate": 2.8154918032786883e-05,
+      "loss": 0.5424,
+      "step": 10300
+    },
+    {
+      "epoch": 68.42,
+      "learning_rate": 2.7581147540983604e-05,
+      "loss": 0.5132,
+      "step": 10400
+    },
+    {
+      "epoch": 69.08,
+      "learning_rate": 2.7007377049180324e-05,
+      "loss": 0.5396,
+      "step": 10500
+    },
+    {
+      "epoch": 69.74,
+      "learning_rate": 2.6433606557377044e-05,
+      "loss": 0.5398,
+      "step": 10600
+    },
+    {
+      "epoch": 70.39,
+      "learning_rate": 2.5859836065573768e-05,
+      "loss": 0.5067,
+      "step": 10700
+    },
+    {
+      "epoch": 71.05,
+      "learning_rate": 2.5291803278688523e-05,
+      "loss": 0.5248,
+      "step": 10800
+    },
+    {
+      "epoch": 71.71,
+      "learning_rate": 2.4718032786885243e-05,
+      "loss": 0.5148,
+      "step": 10900
+    },
+    {
+      "epoch": 72.37,
+      "learning_rate": 2.4144262295081967e-05,
+      "loss": 0.5155,
+      "step": 11000
+    },
+    {
+      "epoch": 73.03,
+      "learning_rate": 2.3570491803278687e-05,
+      "loss": 0.5195,
+      "step": 11100
+    },
+    {
+      "epoch": 73.68,
+      "learning_rate": 2.2996721311475408e-05,
+      "loss": 0.5185,
+      "step": 11200
+    },
+    {
+      "epoch": 74.34,
+      "learning_rate": 2.2422950819672128e-05,
+      "loss": 0.5184,
+      "step": 11300
+    },
+    {
+      "epoch": 75.0,
+      "learning_rate": 2.184918032786885e-05,
+      "loss": 0.5122,
+      "step": 11400
+    },
+    {
+      "epoch": 75.66,
+      "learning_rate": 2.127540983606557e-05,
+      "loss": 0.498,
+      "step": 11500
+    },
+    {
+      "epoch": 76.32,
+      "learning_rate": 2.0701639344262296e-05,
+      "loss": 0.5075,
+      "step": 11600
+    },
+    {
+      "epoch": 76.97,
+      "learning_rate": 2.0127868852459017e-05,
+      "loss": 0.479,
+      "step": 11700
+    },
+    {
+      "epoch": 77.63,
+      "learning_rate": 1.9554098360655737e-05,
+      "loss": 0.4826,
+      "step": 11800
+    },
+    {
+      "epoch": 78.29,
+      "learning_rate": 1.8980327868852457e-05,
+      "loss": 0.5057,
+      "step": 11900
+    },
+    {
+      "epoch": 78.95,
+      "learning_rate": 1.8406557377049178e-05,
+      "loss": 0.4798,
+      "step": 12000
+    },
+    {
+      "epoch": 78.95,
+      "eval_loss": 0.2843836545944214,
+      "eval_runtime": 109.262,
+      "eval_samples_per_second": 19.888,
+      "eval_steps_per_second": 19.888,
+      "eval_wer": 0.26018378823206695,
+      "step": 12000
+    },
+    {
+      "epoch": 79.61,
+      "learning_rate": 1.7832786885245898e-05,
+      "loss": 0.495,
+      "step": 12100
+    },
+    {
+      "epoch": 80.26,
+      "learning_rate": 1.7264754098360656e-05,
+      "loss": 0.4784,
+      "step": 12200
+    },
+    {
+      "epoch": 80.92,
+      "learning_rate": 1.6690983606557376e-05,
+      "loss": 0.477,
+      "step": 12300
+    },
+    {
+      "epoch": 81.58,
+      "learning_rate": 1.6117213114754097e-05,
+      "loss": 0.4936,
+      "step": 12400
+    },
+    {
+      "epoch": 82.24,
+      "learning_rate": 1.5543442622950817e-05,
+      "loss": 0.4942,
+      "step": 12500
+    },
+    {
+      "epoch": 82.89,
+      "learning_rate": 1.496967213114754e-05,
+      "loss": 0.474,
+      "step": 12600
+    },
+    {
+      "epoch": 83.55,
+      "learning_rate": 1.4395901639344262e-05,
+      "loss": 0.4857,
+      "step": 12700
+    },
+    {
+      "epoch": 84.21,
+      "learning_rate": 1.3822131147540982e-05,
+      "loss": 0.4628,
+      "step": 12800
+    },
+    {
+      "epoch": 84.87,
+      "learning_rate": 1.3248360655737704e-05,
+      "loss": 0.4653,
+      "step": 12900
+    },
+    {
+      "epoch": 85.53,
+      "learning_rate": 1.2674590163934426e-05,
+      "loss": 0.4787,
+      "step": 13000
+    },
+    {
+      "epoch": 86.18,
+      "learning_rate": 1.2100819672131147e-05,
+      "loss": 0.4726,
+      "step": 13100
+    },
+    {
+      "epoch": 86.84,
+      "learning_rate": 1.1527049180327867e-05,
+      "loss": 0.4535,
+      "step": 13200
+    },
+    {
+      "epoch": 87.5,
+      "learning_rate": 1.095327868852459e-05,
+      "loss": 0.4648,
+      "step": 13300
+    },
+    {
+      "epoch": 88.16,
+      "learning_rate": 1.0379508196721311e-05,
+      "loss": 0.4681,
+      "step": 13400
+    },
+    {
+      "epoch": 88.82,
+      "learning_rate": 9.805737704918032e-06,
+      "loss": 0.4741,
+      "step": 13500
+    },
+    {
+      "epoch": 89.47,
+      "learning_rate": 9.231967213114752e-06,
+      "loss": 0.4682,
+      "step": 13600
+    },
+    {
+      "epoch": 90.13,
+      "learning_rate": 8.658196721311476e-06,
+      "loss": 0.4627,
+      "step": 13700
+    },
+    {
+      "epoch": 90.79,
+      "learning_rate": 8.084426229508196e-06,
+      "loss": 0.453,
+      "step": 13800
+    },
+    {
+      "epoch": 91.45,
+      "learning_rate": 7.5106557377049175e-06,
+      "loss": 0.451,
+      "step": 13900
+    },
+    {
+      "epoch": 92.11,
+      "learning_rate": 6.936885245901639e-06,
+      "loss": 0.4705,
+      "step": 14000
+    },
+    {
+      "epoch": 92.76,
+      "learning_rate": 6.36311475409836e-06,
+      "loss": 0.4615,
+      "step": 14100
+    },
+    {
+      "epoch": 93.42,
+      "learning_rate": 5.789344262295081e-06,
+      "loss": 0.4507,
+      "step": 14200
+    },
+    {
+      "epoch": 94.08,
+      "learning_rate": 5.2155737704918034e-06,
+      "loss": 0.4526,
+      "step": 14300
+    },
+    {
+      "epoch": 94.74,
+      "learning_rate": 4.641803278688524e-06,
+      "loss": 0.4299,
+      "step": 14400
+    },
+    {
+      "epoch": 95.39,
+      "learning_rate": 4.068032786885246e-06,
+      "loss": 0.4448,
+      "step": 14500
+    },
+    {
+      "epoch": 96.05,
+      "learning_rate": 3.494262295081967e-06,
+      "loss": 0.4414,
+      "step": 14600
+    },
+    {
+      "epoch": 96.71,
+      "learning_rate": 2.920491803278688e-06,
+      "loss": 0.43,
+      "step": 14700
+    },
+    {
+      "epoch": 97.37,
+      "learning_rate": 2.3467213114754098e-06,
+      "loss": 0.4374,
+      "step": 14800
+    },
+    {
+      "epoch": 98.03,
+      "learning_rate": 1.772950819672131e-06,
+      "loss": 0.4582,
+      "step": 14900
+    },
+    {
+      "epoch": 98.68,
+      "learning_rate": 1.1991803278688525e-06,
+      "loss": 0.4399,
+      "step": 15000
+    },
+    {
+      "epoch": 98.68,
+      "eval_loss": 0.2906634509563446,
+      "eval_runtime": 110.6588,
+      "eval_samples_per_second": 19.637,
+      "eval_steps_per_second": 19.637,
+      "eval_wer": 0.24893704567274721,
+      "step": 15000
+    },
+    {
+      "epoch": 99.34,
+      "learning_rate": 6.254098360655737e-07,
+      "loss": 0.4309,
+      "step": 15100
+    },
+    {
+      "epoch": 100.0,
+      "learning_rate": 5.1639344262295075e-08,
+      "loss": 0.4456,
+      "step": 15200
+    },
+    {
+      "epoch": 100.0,
+      "step": 15200,
+      "total_flos": 5.9242678559261245e+19,
+      "train_loss": 1.3542817559995148,
+      "train_runtime": 20867.045,
+      "train_samples_per_second": 23.233,
+      "train_steps_per_second": 0.728
+    }
+  ],
+  "max_steps": 15200,
+  "num_train_epochs": 100,
+  "total_flos": 5.9242678559261245e+19,
+  "trial_name": null,
+  "trial_params": null
+}