commit with good tokenizer

README.md

@@ -0,0 +1,103 @@
+---
+language:
+- fr
+license: apache-2.0
+tags:
+- automatic-speech-recognition
+- mozilla-foundation/common_voice_8_0
+- generated_from_trainer
+- robust-speech-event
+datasets:
+- common_voice
+model-index:
+- name: xls-r-300m-fr
+  results:
+  - task: 
+      name: Speech Recognition
+      type: automatic-speech-recognition
+    dataset:
+      name: Common Voice 8.0 fr
+      type: mozilla-foundation/common_voice_8_0
+      args: fr
+    metrics:
+       - name: Test WER
+         type: wer
+         value: 36.81
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# 
+
+This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - FR dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.2388
+- Wer: 0.3681
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 0.0001
+- train_batch_size: 64
+- eval_batch_size: 64
+- seed: 42
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- lr_scheduler_warmup_steps: 1500
+- num_epochs: 2.0
+- mixed_precision_training: Native AMP
+
+### Training results
+
+| Training Loss | Epoch | Step  | Validation Loss | Wer    |
+|:-------------:|:-----:|:-----:|:---------------:|:------:|
+| 4.3748        | 0.07  | 500   | 3.8784          | 1.0    |
+| 2.8068        | 0.14  | 1000  | 2.8289          | 0.9826 |
+| 1.6698        | 0.22  | 1500  | 0.8811          | 0.7127 |
+| 1.3488        | 0.29  | 2000  | 0.5166          | 0.5369 |
+| 1.2239        | 0.36  | 2500  | 0.4105          | 0.4741 |
+| 1.1537        | 0.43  | 3000  | 0.3585          | 0.4448 |
+| 1.1184        | 0.51  | 3500  | 0.3336          | 0.4292 |
+| 1.0968        | 0.58  | 4000  | 0.3195          | 0.4180 |
+| 1.0737        | 0.65  | 4500  | 0.3075          | 0.4141 |
+| 1.0677        | 0.72  | 5000  | 0.3015          | 0.4089 |
+| 1.0462        | 0.8   | 5500  | 0.2971          | 0.4077 |
+| 1.0392        | 0.87  | 6000  | 0.2870          | 0.3997 |
+| 1.0178        | 0.94  | 6500  | 0.2805          | 0.3963 |
+| 0.992         | 1.01  | 7000  | 0.2748          | 0.3935 |
+| 1.0197        | 1.09  | 7500  | 0.2691          | 0.3884 |
+| 1.0056        | 1.16  | 8000  | 0.2682          | 0.3889 |
+| 0.9826        | 1.23  | 8500  | 0.2647          | 0.3868 |
+| 0.9815        | 1.3   | 9000  | 0.2603          | 0.3832 |
+| 0.9717        | 1.37  | 9500  | 0.2561          | 0.3807 |
+| 0.9605        | 1.45  | 10000 | 0.2523          | 0.3783 |
+| 0.96          | 1.52  | 10500 | 0.2494          | 0.3788 |
+| 0.9442        | 1.59  | 11000 | 0.2478          | 0.3760 |
+| 0.9564        | 1.66  | 11500 | 0.2454          | 0.3733 |
+| 0.9436        | 1.74  | 12000 | 0.2439          | 0.3747 |
+| 0.938         | 1.81  | 12500 | 0.2411          | 0.3716 |
+| 0.9353        | 1.88  | 13000 | 0.2397          | 0.3698 |
+| 0.9271        | 1.95  | 13500 | 0.2388          | 0.3681 |
+
+
+### Framework versions
+
+- Transformers 4.17.0.dev0
+- Pytorch 1.10.2+cu102
+- Datasets 1.18.2.dev0
+- Tokenizers 0.11.0

added_tokens.json

@@ -0,0 +1 @@
+{}

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.23875188827514648,
+    "eval_runtime": 294.1776,
+    "eval_samples": 5792,
+    "eval_samples_per_second": 19.689,
+    "eval_steps_per_second": 0.309,
+    "eval_wer": 0.3680797679950471,
+    "train_loss": 1.442369053426242,
+    "train_runtime": 53680.5392,
+    "train_samples": 442265,
+    "train_samples_per_second": 16.478,
+    "train_steps_per_second": 0.257
+}

config.json

@@ -0,0 +1,107 @@
+{
+  "_name_or_path": "facebook/wav2vec2-xls-r-300m",
+  "activation_dropout": 0.05,
+  "adapter_kernel_size": 3,
+  "adapter_stride": 2,
+  "add_adapter": false,
+  "apply_spec_augment": true,
+  "architectures": [
+    "Wav2Vec2ForCTC"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "classifier_proj_size": 256,
+  "codevector_dim": 768,
+  "contrastive_logits_temperature": 0.1,
+  "conv_bias": true,
+  "conv_dim": [
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512
+  ],
+  "conv_kernel": [
+    10,
+    3,
+    3,
+    3,
+    3,
+    2,
+    2
+  ],
+  "conv_stride": [
+    5,
+    2,
+    2,
+    2,
+    2,
+    2,
+    2
+  ],
+  "ctc_loss_reduction": "mean",
+  "ctc_zero_infinity": false,
+  "diversity_loss_weight": 0.1,
+  "do_stable_layer_norm": true,
+  "eos_token_id": 2,
+  "feat_extract_activation": "gelu",
+  "feat_extract_dropout": 0.0,
+  "feat_extract_norm": "layer",
+  "feat_proj_dropout": 0.0,
+  "feat_quantizer_dropout": 0.0,
+  "final_dropout": 0.0,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.0,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "layerdrop": 0.0,
+  "mask_feature_length": 10,
+  "mask_feature_min_masks": 0,
+  "mask_feature_prob": 0.4,
+  "mask_time_length": 10,
+  "mask_time_min_masks": 2,
+  "mask_time_prob": 0.75,
+  "model_type": "wav2vec2",
+  "num_adapter_layers": 3,
+  "num_attention_heads": 16,
+  "num_codevector_groups": 2,
+  "num_codevectors_per_group": 320,
+  "num_conv_pos_embedding_groups": 16,
+  "num_conv_pos_embeddings": 128,
+  "num_feat_extract_layers": 7,
+  "num_hidden_layers": 24,
+  "num_negatives": 100,
+  "output_hidden_size": 1024,
+  "pad_token_id": 216,
+  "proj_codevector_dim": 768,
+  "tdnn_dilation": [
+    1,
+    2,
+    3,
+    1,
+    1
+  ],
+  "tdnn_dim": [
+    512,
+    512,
+    512,
+    512,
+    1500
+  ],
+  "tdnn_kernel": [
+    5,
+    3,
+    3,
+    1,
+    1
+  ],
+  "torch_dtype": "float32",
+  "transformers_version": "4.17.0.dev0",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 218,
+  "xvector_output_dim": 512
+}

eval.py

@@ -0,0 +1,153 @@
+#!/usr/bin/env python3
+import argparse
+import re
+import unicodedata
+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)
+
+
+chars_to_remove_regex = r'[\,\?\.\!\-\_\;\:\"\“\%\‘\”\�\^]'
+
+def remove_accents(text):
+    nfkd_form = unicodedata.normalize('NFKD', text)
+    return u"".join([c for c in nfkd_form if not unicodedata.combining(c)])
+
+def remove_special_characters(text):
+    text = re.sub(chars_to_remove_regex, '', text).lower()
+    text = re.sub("ç", r'[cedille]', text)
+    text = re.sub("&", r'et', text)    
+    text = re.sub("%", r' pourcents', text)   
+    text = re.sub("([0-9]+)(,|.)([0-9+])", r'\1 virgule \3', text) 
+    text = re.sub("\$", r'dollar', text)
+    text = re.sub("\£", r'livre', text)
+    text = re.sub("\€", r'euro', text)
+    text = remove_accents(text)
+    text = re.sub(r"\[cedille\]", 'ç', text) + " "
+    return text
+    
+def normalize_text(text: str) -> str:
+    text = remove_special_characters(text)
+
+    # 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(20))
+
+    # 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"]# "".join(prediction["text"].split("<s>"))
+        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": 2.0,
+    "eval_loss": 0.23875188827514648,
+    "eval_runtime": 294.1776,
+    "eval_samples": 5792,
+    "eval_samples_per_second": 19.689,
+    "eval_steps_per_second": 0.309,
+    "eval_wer": 0.3680797679950471
+}

preprocessor_config.json

@@ -0,0 +1,9 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0,
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7be356c0416d66c909300c8a24b255d6bc972bb2572b661bdcc3e0167f8aaba0
+size 1262817457

run.sh

@@ -0,0 +1,41 @@
+WANDB_PROJECT=auto-speech-recognition-french
+python run_speech_recognition_ctc.py \
+--dataset_name="mozilla-foundation/common_voice_8_0" \
+--model_name_or_path="facebook/wav2vec2-xls-r-300m" \
+--dataset_config_name="fr" \
+--tokenizer_name_or_path="./" \
+--output_dir="./" \
+--overwrite_output_dir \
+--num_train_epochs="2" \
+--per_device_train_batch_size="64" \
+--per_device_eval_batch_size="64" \
+--gradient_accumulation_steps="1" \
+--learning_rate="1e-4" \
+--warmup_steps="1500" \
+--length_column_name="input_length" \
+--evaluation_strategy="steps" \
+--text_column_name="sentence" \
+--save_steps="500" \
+--eval_steps="500" \
+--logging_steps="100" \
+--layerdrop="0.0" \
+--activation_dropout="0.05" \
+--save_total_limit="2" \
+--freeze_feature_encoder \
+--feat_proj_dropout="0.0" \
+--mask_time_prob="0.75" \
+--mask_time_length="10" \
+--mask_feature_prob="0.4" \
+--mask_feature_length="10" \
+--gradient_checkpointing \
+--report_to="wandb" \
+--run_name="xls-r-300m-fr" \
+--max_eval_samples="6000" \
+--max_duration_in_seconds="9" \
+--use_auth_token \
+--fp16 \
+--group_by_length \
+--preprocessing_num_workers="64" \
+--do_train --do_eval \
+--load_best_model_at_end \
+--push_to_hub

run_speech_recognition_ctc.py

@@ -0,0 +1,748 @@
+#!/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 unicodedata
+
+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,
+    Wav2Vec2CTCTokenizer,
+    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 '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: 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"].shuffle(seed=42).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
+    text_column_name = data_args.text_column_name
+
+    chars_to_remove_regex = r'[\,\?\.\!\-\_\;\:\"\“\%\‘\”\�\^]'
+    
+    def remove_accents(input_str):
+        nfkd_form = unicodedata.normalize('NFKD', input_str)
+        return u"".join([c for c in nfkd_form if not unicodedata.combining(c)])
+
+    def remove_special_characters(batch):
+        batch["target_text"] = re.sub(chars_to_remove_regex, '', batch[text_column_name]).lower()
+        batch["target_text"] = re.sub("ç", r'[cedille]', batch["target_text"])
+        batch["target_text"] = re.sub("&", r'et', batch["target_text"])    
+        batch["target_text"] = re.sub("%", r' pourcents', batch["target_text"])   
+        batch["target_text"] = re.sub("([0-9]+)(,|.)([0-9+])", r'\1 virgule \3', batch["target_text"]) 
+        batch["target_text"] = re.sub("\$", r'dollar', batch["target_text"])
+        batch["target_text"] = re.sub("\£", r'livre', batch["target_text"])
+        batch["target_text"] = re.sub("\€", r'euro', batch["target_text"])
+        batch["target_text"] = remove_accents(batch["target_text"])
+        batch["target_text"] = re.sub(r"\[cedille\]", 'ç', batch["target_text"]) + " "
+        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 = Wav2Vec2CTCTokenizer.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, 
+            batch_size=-1,
+            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, skip_special_tokens=True)#being sure to remove <s> from the output
+        # 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()

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": null, "eos_token": null, "unk_token": "[UNK]", "pad_token": "[PAD]", "additional_special_tokens": [{"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}]}

tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "[UNK]", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "[PAD]", "do_lower_case": false, "word_delimiter_token": "|", "config": null, "tokenizer_type": "wav2vec2", "special_tokens_map_file": null, "tokenizer_file": null, "name_or_path": "./", "tokenizer_class": "Wav2Vec2CTCTokenizer"}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.0,
+    "train_loss": 1.442369053426242,
+    "train_runtime": 53680.5392,
+    "train_samples": 442265,
+    "train_samples_per_second": 16.478,
+    "train_steps_per_second": 0.257
+}

trainer_state.json

@@ -0,0 +1,1096 @@
+{
+  "best_metric": 0.23875188827514648,
+  "best_model_checkpoint": "./checkpoint-13500",
+  "epoch": 2.0,
+  "global_step": 13822,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.533333333333333e-06,
+      "loss": 17.2403,
+      "step": 100
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.32e-05,
+      "loss": 10.2311,
+      "step": 200
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9800000000000004e-05,
+      "loss": 7.834,
+      "step": 300
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.646666666666667e-05,
+      "loss": 6.0656,
+      "step": 400
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 3.313333333333333e-05,
+      "loss": 4.3748,
+      "step": 500
+    },
+    {
+      "epoch": 0.07,
+      "eval_loss": 3.878422975540161,
+      "eval_runtime": 285.8223,
+      "eval_samples_per_second": 20.264,
+      "eval_steps_per_second": 0.318,
+      "eval_wer": 1.0,
+      "step": 500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 3.9800000000000005e-05,
+      "loss": 3.2923,
+      "step": 600
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.646666666666667e-05,
+      "loss": 2.9475,
+      "step": 700
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.3133333333333335e-05,
+      "loss": 2.8639,
+      "step": 800
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.9800000000000003e-05,
+      "loss": 2.8265,
+      "step": 900
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 6.646666666666667e-05,
+      "loss": 2.8068,
+      "step": 1000
+    },
+    {
+      "epoch": 0.14,
+      "eval_loss": 2.828850746154785,
+      "eval_runtime": 292.3877,
+      "eval_samples_per_second": 19.809,
+      "eval_steps_per_second": 0.311,
+      "eval_wer": 0.9826485059793412,
+      "step": 1000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 7.306666666666668e-05,
+      "loss": 2.779,
+      "step": 1100
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 7.973333333333334e-05,
+      "loss": 2.6402,
+      "step": 1200
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 8.64e-05,
+      "loss": 2.1119,
+      "step": 1300
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.306666666666667e-05,
+      "loss": 1.7965,
+      "step": 1400
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.973333333333334e-05,
+      "loss": 1.6698,
+      "step": 1500
+    },
+    {
+      "epoch": 0.22,
+      "eval_loss": 0.881136417388916,
+      "eval_runtime": 297.1806,
+      "eval_samples_per_second": 19.49,
+      "eval_steps_per_second": 0.306,
+      "eval_wer": 0.7127472384241911,
+      "step": 1500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.92209056971271e-05,
+      "loss": 1.5882,
+      "step": 1600
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.840934913163448e-05,
+      "loss": 1.5172,
+      "step": 1700
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.759779256614186e-05,
+      "loss": 1.4579,
+      "step": 1800
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.678623600064926e-05,
+      "loss": 1.3829,
+      "step": 1900
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.597467943515663e-05,
+      "loss": 1.3488,
+      "step": 2000
+    },
+    {
+      "epoch": 0.29,
+      "eval_loss": 0.516592800617218,
+      "eval_runtime": 301.2842,
+      "eval_samples_per_second": 19.224,
+      "eval_steps_per_second": 0.302,
+      "eval_wer": 0.5369024731988661,
+      "step": 2000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.516312286966402e-05,
+      "loss": 1.2981,
+      "step": 2100
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.43515663041714e-05,
+      "loss": 1.2845,
+      "step": 2200
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.354000973867879e-05,
+      "loss": 1.2459,
+      "step": 2300
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.272845317318618e-05,
+      "loss": 1.2255,
+      "step": 2400
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.191689660769356e-05,
+      "loss": 1.2239,
+      "step": 2500
+    },
+    {
+      "epoch": 0.36,
+      "eval_loss": 0.4104757010936737,
+      "eval_runtime": 299.1395,
+      "eval_samples_per_second": 19.362,
+      "eval_steps_per_second": 0.304,
+      "eval_wer": 0.474111245071524,
+      "step": 2500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.110534004220094e-05,
+      "loss": 1.2024,
+      "step": 2600
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.030189904236326e-05,
+      "loss": 1.1851,
+      "step": 2700
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 8.949034247687063e-05,
+      "loss": 1.1768,
+      "step": 2800
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 8.867878591137803e-05,
+      "loss": 1.1641,
+      "step": 2900
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 8.786722934588541e-05,
+      "loss": 1.1537,
+      "step": 3000
+    },
+    {
+      "epoch": 0.43,
+      "eval_loss": 0.35850802063941956,
+      "eval_runtime": 299.662,
+      "eval_samples_per_second": 19.328,
+      "eval_steps_per_second": 0.304,
+      "eval_wer": 0.4448499462348073,
+      "step": 3000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 8.70556727803928e-05,
+      "loss": 1.1449,
+      "step": 3100
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 8.624411621490018e-05,
+      "loss": 1.1379,
+      "step": 3200
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 8.543255964940758e-05,
+      "loss": 1.1331,
+      "step": 3300
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 8.462100308391495e-05,
+      "loss": 1.1205,
+      "step": 3400
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 8.380944651842234e-05,
+      "loss": 1.1184,
+      "step": 3500
+    },
+    {
+      "epoch": 0.51,
+      "eval_loss": 0.333638072013855,
+      "eval_runtime": 297.0402,
+      "eval_samples_per_second": 19.499,
+      "eval_steps_per_second": 0.306,
+      "eval_wer": 0.42922545537489004,
+      "step": 3500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 8.299788995292971e-05,
+      "loss": 1.1014,
+      "step": 3600
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 8.218633338743711e-05,
+      "loss": 1.1114,
+      "step": 3700
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.13747768219445e-05,
+      "loss": 1.117,
+      "step": 3800
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.056322025645188e-05,
+      "loss": 1.102,
+      "step": 3900
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.975166369095926e-05,
+      "loss": 1.0968,
+      "step": 4000
+    },
+    {
+      "epoch": 0.58,
+      "eval_loss": 0.31949570775032043,
+      "eval_runtime": 296.0172,
+      "eval_samples_per_second": 19.566,
+      "eval_steps_per_second": 0.307,
+      "eval_wer": 0.4180162273127179,
+      "step": 4000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.894822269112158e-05,
+      "loss": 1.0942,
+      "step": 4100
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.813666612562897e-05,
+      "loss": 1.0859,
+      "step": 4200
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 7.732510956013635e-05,
+      "loss": 1.0767,
+      "step": 4300
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 7.652166856029866e-05,
+      "loss": 1.0766,
+      "step": 4400
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 7.571011199480604e-05,
+      "loss": 1.0737,
+      "step": 4500
+    },
+    {
+      "epoch": 0.65,
+      "eval_loss": 0.30754634737968445,
+      "eval_runtime": 296.2378,
+      "eval_samples_per_second": 19.552,
+      "eval_steps_per_second": 0.307,
+      "eval_wer": 0.41408973899442797,
+      "step": 4500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.489855542931342e-05,
+      "loss": 1.0807,
+      "step": 4600
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.40869988638208e-05,
+      "loss": 1.071,
+      "step": 4700
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 7.32754422983282e-05,
+      "loss": 1.0613,
+      "step": 4800
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 7.246388573283557e-05,
+      "loss": 1.0635,
+      "step": 4900
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 7.165232916734297e-05,
+      "loss": 1.0677,
+      "step": 5000
+    },
+    {
+      "epoch": 0.72,
+      "eval_loss": 0.30150118470191956,
+      "eval_runtime": 297.4742,
+      "eval_samples_per_second": 19.471,
+      "eval_steps_per_second": 0.306,
+      "eval_wer": 0.4089250219948516,
+      "step": 5000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 7.084077260185034e-05,
+      "loss": 1.0707,
+      "step": 5100
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 7.002921603635774e-05,
+      "loss": 1.0617,
+      "step": 5200
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 6.921765947086512e-05,
+      "loss": 1.0566,
+      "step": 5300
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 6.84061029053725e-05,
+      "loss": 1.0518,
+      "step": 5400
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.760266190553481e-05,
+      "loss": 1.0462,
+      "step": 5500
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 0.297052800655365,
+      "eval_runtime": 296.1592,
+      "eval_samples_per_second": 19.557,
+      "eval_steps_per_second": 0.307,
+      "eval_wer": 0.4077193782788621,
+      "step": 5500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 6.679110534004221e-05,
+      "loss": 1.0514,
+      "step": 5600
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 6.597954877454959e-05,
+      "loss": 1.0446,
+      "step": 5700
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 6.516799220905698e-05,
+      "loss": 1.0358,
+      "step": 5800
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 6.435643564356436e-05,
+      "loss": 1.0364,
+      "step": 5900
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 6.354487907807174e-05,
+      "loss": 1.0392,
+      "step": 6000
+    },
+    {
+      "epoch": 0.87,
+      "eval_loss": 0.2870033383369446,
+      "eval_runtime": 295.9814,
+      "eval_samples_per_second": 19.569,
+      "eval_steps_per_second": 0.307,
+      "eval_wer": 0.3997034768157972,
+      "step": 6000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 6.273332251257913e-05,
+      "loss": 1.0375,
+      "step": 6100
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 6.192176594708652e-05,
+      "loss": 1.0408,
+      "step": 6200
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 6.11102093815939e-05,
+      "loss": 1.0382,
+      "step": 6300
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 6.0298652816101284e-05,
+      "loss": 1.0335,
+      "step": 6400
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 5.948709625060867e-05,
+      "loss": 1.0178,
+      "step": 6500
+    },
+    {
+      "epoch": 0.94,
+      "eval_loss": 0.28046590089797974,
+      "eval_runtime": 297.8045,
+      "eval_samples_per_second": 19.449,
+      "eval_steps_per_second": 0.306,
+      "eval_wer": 0.39629834794225943,
+      "step": 6500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.867553968511605e-05,
+      "loss": 1.0312,
+      "step": 6600
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.786398311962344e-05,
+      "loss": 1.033,
+      "step": 6700
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 5.7052426554130825e-05,
+      "loss": 1.0289,
+      "step": 6800
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 5.624086998863821e-05,
+      "loss": 1.027,
+      "step": 6900
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 5.54293134231456e-05,
+      "loss": 0.992,
+      "step": 7000
+    },
+    {
+      "epoch": 1.01,
+      "eval_loss": 0.2747785747051239,
+      "eval_runtime": 298.4841,
+      "eval_samples_per_second": 19.405,
+      "eval_steps_per_second": 0.305,
+      "eval_wer": 0.39352862589201343,
+      "step": 7000
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 5.461775685765298e-05,
+      "loss": 1.0025,
+      "step": 7100
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 5.3806200292160366e-05,
+      "loss": 1.0122,
+      "step": 7200
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 5.299464372666775e-05,
+      "loss": 1.018,
+      "step": 7300
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 5.218308716117514e-05,
+      "loss": 0.9936,
+      "step": 7400
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 5.137153059568252e-05,
+      "loss": 1.0197,
+      "step": 7500
+    },
+    {
+      "epoch": 1.09,
+      "eval_loss": 0.26907604932785034,
+      "eval_runtime": 298.796,
+      "eval_samples_per_second": 19.384,
+      "eval_steps_per_second": 0.305,
+      "eval_wer": 0.3884453713056796,
+      "step": 7500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 5.055997403018991e-05,
+      "loss": 1.008,
+      "step": 7600
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.97484174646973e-05,
+      "loss": 1.0028,
+      "step": 7700
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.893686089920468e-05,
+      "loss": 0.9929,
+      "step": 7800
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.8125304333712064e-05,
+      "loss": 0.995,
+      "step": 7900
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.731374776821945e-05,
+      "loss": 1.0056,
+      "step": 8000
+    },
+    {
+      "epoch": 1.16,
+      "eval_loss": 0.26817116141319275,
+      "eval_runtime": 298.1504,
+      "eval_samples_per_second": 19.426,
+      "eval_steps_per_second": 0.305,
+      "eval_wer": 0.3888526833718922,
+      "step": 8000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.6510306768381754e-05,
+      "loss": 0.9971,
+      "step": 8100
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.5698750202889145e-05,
+      "loss": 0.9976,
+      "step": 8200
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.488719363739653e-05,
+      "loss": 1.0014,
+      "step": 8300
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.407563707190391e-05,
+      "loss": 0.9835,
+      "step": 8400
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.3272196072066225e-05,
+      "loss": 0.9826,
+      "step": 8500
+    },
+    {
+      "epoch": 1.23,
+      "eval_loss": 0.26473307609558105,
+      "eval_runtime": 299.0161,
+      "eval_samples_per_second": 19.37,
+      "eval_steps_per_second": 0.304,
+      "eval_wer": 0.38675095311023494,
+      "step": 8500
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.2460639506573615e-05,
+      "loss": 0.9838,
+      "step": 8600
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.1649082941081e-05,
+      "loss": 0.9836,
+      "step": 8700
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.0845641941243305e-05,
+      "loss": 0.9824,
+      "step": 8800
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.003408537575069e-05,
+      "loss": 0.9715,
+      "step": 8900
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.922252881025807e-05,
+      "loss": 0.9815,
+      "step": 9000
+    },
+    {
+      "epoch": 1.3,
+      "eval_loss": 0.26034271717071533,
+      "eval_runtime": 299.6782,
+      "eval_samples_per_second": 19.327,
+      "eval_steps_per_second": 0.304,
+      "eval_wer": 0.3831828994102121,
+      "step": 9000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.841097224476546e-05,
+      "loss": 0.9757,
+      "step": 9100
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.7599415679272846e-05,
+      "loss": 0.9689,
+      "step": 9200
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.678785911378023e-05,
+      "loss": 0.9778,
+      "step": 9300
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.5976302548287614e-05,
+      "loss": 0.9794,
+      "step": 9400
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.5164745982795e-05,
+      "loss": 0.9717,
+      "step": 9500
+    },
+    {
+      "epoch": 1.37,
+      "eval_loss": 0.25609487295150757,
+      "eval_runtime": 299.6976,
+      "eval_samples_per_second": 19.326,
+      "eval_steps_per_second": 0.304,
+      "eval_wer": 0.3807064420476392,
+      "step": 9500
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.435318941730239e-05,
+      "loss": 0.9752,
+      "step": 9600
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 3.354163285180977e-05,
+      "loss": 0.965,
+      "step": 9700
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 3.2730076286317155e-05,
+      "loss": 0.9522,
+      "step": 9800
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.191851972082454e-05,
+      "loss": 0.9718,
+      "step": 9900
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.110696315533193e-05,
+      "loss": 0.9605,
+      "step": 10000
+    },
+    {
+      "epoch": 1.45,
+      "eval_loss": 0.25231894850730896,
+      "eval_runtime": 297.5796,
+      "eval_samples_per_second": 19.464,
+      "eval_steps_per_second": 0.306,
+      "eval_wer": 0.3782951546156603,
+      "step": 10000
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.0295406589839315e-05,
+      "loss": 0.9635,
+      "step": 10100
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.94838500243467e-05,
+      "loss": 0.9632,
+      "step": 10200
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.8672293458854082e-05,
+      "loss": 0.9548,
+      "step": 10300
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.7868852459016392e-05,
+      "loss": 0.9554,
+      "step": 10400
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.7057295893523783e-05,
+      "loss": 0.96,
+      "step": 10500
+    },
+    {
+      "epoch": 1.52,
+      "eval_loss": 0.24939315021038055,
+      "eval_runtime": 300.036,
+      "eval_samples_per_second": 19.304,
+      "eval_steps_per_second": 0.303,
+      "eval_wer": 0.3787513441298185,
+      "step": 10500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.6245739328031166e-05,
+      "loss": 0.9611,
+      "step": 10600
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.543418276253855e-05,
+      "loss": 0.9594,
+      "step": 10700
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4622626197045937e-05,
+      "loss": 0.9589,
+      "step": 10800
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.381106963155332e-05,
+      "loss": 0.9441,
+      "step": 10900
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.2999513066060704e-05,
+      "loss": 0.9442,
+      "step": 11000
+    },
+    {
+      "epoch": 1.59,
+      "eval_loss": 0.24783751368522644,
+      "eval_runtime": 297.7741,
+      "eval_samples_per_second": 19.451,
+      "eval_steps_per_second": 0.306,
+      "eval_wer": 0.3760142070448695,
+      "step": 11000
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.218795650056809e-05,
+      "loss": 0.9496,
+      "step": 11100
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.1376399935075474e-05,
+      "loss": 0.9486,
+      "step": 11200
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.056484336958286e-05,
+      "loss": 0.9558,
+      "step": 11300
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.9753286804090245e-05,
+      "loss": 0.9486,
+      "step": 11400
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.8941730238597632e-05,
+      "loss": 0.9564,
+      "step": 11500
+    },
+    {
+      "epoch": 1.66,
+      "eval_loss": 0.2454409897327423,
+      "eval_runtime": 296.0265,
+      "eval_samples_per_second": 19.566,
+      "eval_steps_per_second": 0.307,
+      "eval_wer": 0.3733096549252175,
+      "step": 11500
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.8130173673105015e-05,
+      "loss": 0.9427,
+      "step": 11600
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.73186171076124e-05,
+      "loss": 0.9423,
+      "step": 11700
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.6507060542119786e-05,
+      "loss": 0.9503,
+      "step": 11800
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.5695503976627173e-05,
+      "loss": 0.9383,
+      "step": 11900
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.4883947411134558e-05,
+      "loss": 0.9436,
+      "step": 12000
+    },
+    {
+      "epoch": 1.74,
+      "eval_loss": 0.24390804767608643,
+      "eval_runtime": 295.4584,
+      "eval_samples_per_second": 19.603,
+      "eval_steps_per_second": 0.308,
+      "eval_wer": 0.37467822346769203,
+      "step": 12000
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.4072390845641942e-05,
+      "loss": 0.9491,
+      "step": 12100
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.3260834280149325e-05,
+      "loss": 0.9419,
+      "step": 12200
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.2449277714656712e-05,
+      "loss": 0.9517,
+      "step": 12300
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.1637721149164097e-05,
+      "loss": 0.9367,
+      "step": 12400
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.0826164583671483e-05,
+      "loss": 0.938,
+      "step": 12500
+    },
+    {
+      "epoch": 1.81,
+      "eval_loss": 0.24111612141132355,
+      "eval_runtime": 296.9314,
+      "eval_samples_per_second": 19.506,
+      "eval_steps_per_second": 0.306,
+      "eval_wer": 0.37159894424712436,
+      "step": 12500
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.0014608018178868e-05,
+      "loss": 0.9337,
+      "step": 12600
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 9.203051452686253e-06,
+      "loss": 0.9284,
+      "step": 12700
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 8.391494887193638e-06,
+      "loss": 0.938,
+      "step": 12800
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 7.579938321701023e-06,
+      "loss": 0.9365,
+      "step": 12900
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.768381756208409e-06,
+      "loss": 0.9353,
+      "step": 13000
+    },
+    {
+      "epoch": 1.88,
+      "eval_loss": 0.23965783417224884,
+      "eval_runtime": 296.6078,
+      "eval_samples_per_second": 19.527,
+      "eval_steps_per_second": 0.307,
+      "eval_wer": 0.3697904786731402,
+      "step": 13000
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 5.956825190715793e-06,
+      "loss": 0.9413,
+      "step": 13100
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 5.153384190878104e-06,
+      "loss": 0.9356,
+      "step": 13200
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 4.34182762538549e-06,
+      "loss": 0.9209,
+      "step": 13300
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.530271059892875e-06,
+      "loss": 0.9362,
+      "step": 13400
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.71871449440026e-06,
+      "loss": 0.9271,
+      "step": 13500
+    },
+    {
+      "epoch": 1.95,
+      "eval_loss": 0.23875188827514648,
+      "eval_runtime": 296.9414,
+      "eval_samples_per_second": 19.506,
+      "eval_steps_per_second": 0.306,
+      "eval_wer": 0.3680797679950471,
+      "step": 13500
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.907157928907645e-06,
+      "loss": 0.9288,
+      "step": 13600
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.0956013634150302e-06,
+      "loss": 0.9345,
+      "step": 13700
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 2.840447979224152e-07,
+      "loss": 0.9326,
+      "step": 13800
+    },
+    {
+      "epoch": 2.0,
+      "step": 13822,
+      "total_flos": 1.2600843645735263e+20,
+      "train_loss": 1.442369053426242,
+      "train_runtime": 53680.5392,
+      "train_samples_per_second": 16.478,
+      "train_steps_per_second": 0.257
+    }
+  ],
+  "max_steps": 13822,
+  "num_train_epochs": 2,
+  "total_flos": 1.2600843645735263e+20,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3515fcb3146dcedc28056bf8ff0c0b0a7c0ebdb32789529ea25826588fbd9491
+size 3055

vocab.json

@@ -0,0 +1 @@
+{"'": 1, "(": 2, ")": 3, "*": 4, ".": 5, "/": 6, "1": 7, "2": 8, "=": 9, "C": 10, "E": 11, "N": 12, "Q": 13, "R": 14, "Z": 15, "`": 16, "a": 17, "b": 18, "c": 19, "d": 20, "e": 21, "f": 22, "g": 23, "h": 24, "i": 25, "j": 26, "k": 27, "l": 28, "m": 29, "n": 30, "o": 31, "p": 32, "q": 33, "r": 34, "s": 35, "t": 36, "u": 37, "v": 38, "w": 39, "x": 40, "y": 41, "z": 42, "{": 43, "|": 0, "}": 45, "~": 46, "§": 47, "«": 48, "®": 49, "°": 50, "±": 51, "·": 52, "»": 53, "×": 54, "ß": 55, "æ": 56, "ç": 57, "ð": 58, "ø": 59, "þ": 60, "đ": 61, "ħ": 62, "ı": 63, "ł": 64, "œ": 65, "ǀ": 66, "ǃ": 67, "ɑ": 68, "ə": 69, "ɨ": 70, "ʉ": 71, "ʔ": 72, "ʻ": 73, "ʼ": 74, "ʽ": 75, "ʾ": 76, "ʿ": 77, "ː": 78, "α": 79, "β": 80, "γ": 81, "δ": 82, "ε": 83, "ζ": 84, "η": 85, "θ": 86, "ι": 87, "κ": 88, "λ": 89, "μ": 90, "ν": 91, "ο": 92, "π": 93, "ρ": 94, "ς": 95, "σ": 96, "τ": 97, "υ": 98, "φ": 99, "χ": 100, "ψ": 101, "ω": 102, "а": 103, "г": 104, "е": 105, "з": 106, "и": 107, "к": 108, "м": 109, "н": 110, "о": 111, "п": 112, "р": 113, "ц": 114, "ч": 115, "э": 116, "я": 117, "є": 118, "і": 119, "ј": 120, "џ": 121, "ҫ": 122, "ӌ": 123, "գ": 124, "զ": 125, "ا": 126, "ب": 127, "ة": 128, "د": 129, "ر": 130, "ل": 131, "م": 132, "ن": 133, "و": 134, "ي": 135, "ᄀ": 136, "ᄆ": 137, "ᄉ": 138, "ᄌ": 139, "ᅡ": 140, "ᅢ": 141, "ᅥ": 142, "ᅩ": 143, "ᅵ": 144, "ᆨ": 145, "ᆷ": 146, "ᆸ": 147, "ᆼ": 148, "ቀ": 149, "ከ": 150, "ወ": 151, "ደ": 152, "ጀ": 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, "ⱅ": 180, "ⱎ": 181, "い": 182, "う": 183, "た": 184, "つ": 185, "の": 186, "ひ": 187, "へ": 188, "ま": 189, "む": 190, "め": 191, "も": 192, "や": 193, "三": 194, "丹": 195, "乃": 196, "京": 197, "保": 198, "北": 199, "厳": 200, "宇": 201, "扬": 202, "文": 203, "星": 204, "术": 205, "杜": 206, "津": 207, "牡": 208, "甌": 209, "美": 210, "西": 211, "貴": 212, "青": 213, "馆": 214, "ꝑ": 215, "[UNK]": 215, "[PAD]": 216}

wandb/debug-internal.log

@@ -0,0 +1 @@
+run-20220129_215451-1vipdbow/logs/debug-internal.log

wandb/debug.log

@@ -0,0 +1 @@
+run-20220129_215451-1vipdbow/logs/debug.log

wandb/latest-run

@@ -0,0 +1 @@
+run-20220129_215451-1vipdbow

wandb/run-20220129_131141-h6nhqm30/files/output.log

@@ -0,0 +1,9003 @@
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  0%|▋                                                                                                                                                              | 100/21520 [04:25<9:53:42,  1.66s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  1%|█▍                                                                                                                                                            | 199/21520 [08:51<10:08:31,  1.71s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  1%|██▏                                                                                                                                                            | 300/21520 [13:20<9:42:10,  1.65s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  2%|██▉                                                                                                                                                            | 399/21520 [17:44<9:49:47,  1.68s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  2%|███▋                                                                                                                                                           | 500/21520 [22:14<9:36:18,  1.65s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+{'loss': 5.4101, 'learning_rate': 2.3239999999999998e-05, 'epoch': 0.12}
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Configuration saved in ./checkpoint-500/config.json
+{'eval_loss': 5.053680896759033, 'eval_wer': 1.0, 'eval_runtime': 62.0508, 'eval_samples_per_second': 29.637, 'eval_steps_per_second': 0.467, 'epoch': 0.12}
+Model weights saved in ./checkpoint-500/pytorch_model.bin
+Configuration saved in ./checkpoint-500/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  3%|████▍                                                                                                                                                          | 600/21520 [28:28<9:27:51,  1.63s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  3%|█████▏                                                                                                                                                         | 700/21520 [32:54<9:54:50,  1.71s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  4%|█████▉                                                                                                                                                         | 799/21520 [37:16<9:45:40,  1.70s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  4%|██████▋                                                                                                                                                        | 899/21520 [41:40<9:40:54,  1.69s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  5%|███████▍                                                                                                                                                       | 999/21520 [46:03<9:58:48,  1.75s/it]
+  5%|███████▎                                                                                                                                                      | 1000/21520 [46:05<9:34:03,  1.68s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:55<00:00,  1.82s/it]
+
+Configuration saved in ./checkpoint-1000/config.json
+Model weights saved in ./checkpoint-1000/pytorch_model.bin
+Configuration saved in ./checkpoint-1000/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  5%|████████                                                                                                                                                      | 1099/21520 [53:19<9:33:07,  1.68s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  6%|████████▊                                                                                                                                                     | 1199/21520 [57:45<9:59:52,  1.77s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  6%|█████████▍                                                                                                                                                  | 1299/21520 [1:02:09<9:15:12,  1.65s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  7%|██████████▏                                                                                                                                                 | 1399/21520 [1:06:35<9:28:45,  1.70s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  7%|██████████▊                                                                                                                                                 | 1499/21520 [1:10:59<9:13:59,  1.66s/it]
+  7%|██████████▊                                                                                                                                                 | 1500/21520 [1:11:00<8:58:38,  1.61s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Configuration saved in ./checkpoint-1500/config.json
+{'eval_loss': 1.2048969268798828, 'eval_wer': 0.8062855432505238, 'eval_runtime': 64.3059, 'eval_samples_per_second': 28.598, 'eval_steps_per_second': 0.451, 'epoch': 0.35}
+Model weights saved in ./checkpoint-1500/pytorch_model.bin
+Configuration saved in ./checkpoint-1500/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Adding files tracked by Git LFS: ['wandb/run-20220129_131141-h6nhqm30/run-h6nhqm30.wandb']. This may take a bit of time if the files are large.
+01/29/2022 14:25:13 - WARNING - huggingface_hub.repository - Adding files tracked by Git LFS: ['wandb/run-20220129_131141-h6nhqm30/run-h6nhqm30.wandb']. This may take a bit of time if the files are large.
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  7%|███████████▌                                                                                                                                                | 1599/21520 [1:18:18<9:12:21,  1.66s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  8%|████████████▎                                                                                                                                               | 1700/21520 [1:22:44<8:47:54,  1.60s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  8%|█████████████                                                                                                                                               | 1799/21520 [1:27:06<9:14:58,  1.69s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  9%|█████████████▊                                                                                                                                              | 1900/21520 [1:31:34<8:50:40,  1.62s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  9%|██████████████▍                                                                                                                                             | 2000/21520 [1:35:55<8:46:26,  1.62s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+{'loss': 1.5498, 'learning_rate': 6.826923076923076e-05, 'epoch': 0.46}
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:55<00:00,  1.81s/it]
+
+Configuration saved in ./checkpoint-2000/config.json
+Model weights saved in ./checkpoint-2000/pytorch_model.bin
+Configuration saved in ./checkpoint-2000/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Deleting older checkpoint [checkpoint-500] due to args.save_total_limit
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 10%|███████████████▏                                                                                                                                            | 2100/21520 [1:43:15<8:45:13,  1.62s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 10%|███████████████▉                                                                                                                                            | 2200/21520 [1:47:39<8:48:39,  1.64s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 11%|████████████████▋                                                                                                                                           | 2299/21520 [1:52:02<8:59:39,  1.68s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 11%|█████████████████▍                                                                                                                                          | 2400/21520 [1:56:27<8:53:22,  1.67s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 12%|██████████████████                                                                                                                                          | 2499/21520 [2:00:52<9:23:31,  1.78s/it]
+ 12%|██████████████████                                                                                                                                          | 2500/21520 [2:00:53<8:59:59,  1.70s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Configuration saved in ./checkpoint-2500/config.json
+{'eval_loss': 0.4956221282482147, 'eval_wer': 0.5256510026938043, 'eval_runtime': 65.2452, 'eval_samples_per_second': 28.186, 'eval_steps_per_second': 0.444, 'epoch': 0.58}
+Model weights saved in ./checkpoint-2500/pytorch_model.bin
+Configuration saved in ./checkpoint-2500/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Deleting older checkpoint [checkpoint-1000] due to args.save_total_limit
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 12%|██████████████████▊                                                                                                                                         | 2600/21520 [2:08:12<8:38:33,  1.64s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 13%|███████████████████▌                                                                                                                                        | 2699/21520 [2:12:34<9:20:50,  1.79s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 13%|████████████████████▎                                                                                                                                       | 2799/21520 [2:16:58<8:44:37,  1.68s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 13%|█████████████████████                                                                                                                                       | 2900/21520 [2:21:21<8:26:00,  1.63s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 14%|█████████████████████▋                                                                                                                                      | 2998/21520 [2:25:39<9:03:14,  1.76s/it]
+ 14%|█████████████████████▋                                                                                                                                      | 3000/21520 [2:25:42<8:12:20,  1.60s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:55<00:00,  1.82s/it]
+
+Configuration saved in ./checkpoint-3000/config.json
+Model weights saved in ./checkpoint-3000/pytorch_model.bin
+Configuration saved in ./checkpoint-3000/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Deleting older checkpoint [checkpoint-1500] due to args.save_total_limit
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 14%|██████████████████████▍                                                                                                                                     | 3100/21520 [2:33:03<8:22:01,  1.64s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 15%|███████████████████████▏                                                                                                                                    | 3199/21520 [2:37:24<8:25:51,  1.66s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 15%|███████████████████████▉                                                                                                                                    | 3300/21520 [2:41:52<8:21:22,  1.65s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 16%|████████████████████████▋                                                                                                                                   | 3398/21520 [2:46:11<8:41:39,  1.73s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 16%|█████████████████████████▎                                                                                                                                  | 3499/21520 [2:50:38<8:32:27,  1.71s/it]
+ 16%|█████████████████████████▎                                                                                                                                  | 3500/21520 [2:50:39<8:11:10,  1.64s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Configuration saved in ./checkpoint-3500/config.json
+{'eval_loss': 0.38118353486061096, 'eval_wer': 0.4635139179886262, 'eval_runtime': 65.5254, 'eval_samples_per_second': 28.065, 'eval_steps_per_second': 0.443, 'epoch': 0.81}
+Model weights saved in ./checkpoint-3500/pytorch_model.bin
+Configuration saved in ./checkpoint-3500/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Deleting older checkpoint [checkpoint-2000] due to args.save_total_limit
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 17%|██████████████████████████                                                                                                                                  | 3600/21520 [2:58:00<8:11:02,  1.64s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 17%|██████████████████████████▊                                                                                                                                 | 3699/21520 [3:02:22<8:11:14,  1.65s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 18%|███████████████████████████▌                                                                                                                                | 3800/21520 [3:06:46<7:57:07,  1.62s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 18%|████████████████████████████▎                                                                                                                               | 3898/21520 [3:11:08<8:29:00,  1.73s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 19%|████████████████████████████▉                                                                                                                               | 4000/21520 [3:15:36<7:50:45,  1.61s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+{'loss': 1.2001, 'learning_rate': 6.128321678321677e-05, 'epoch': 0.93}
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:55<00:00,  1.79s/it]
+
+Configuration saved in ./checkpoint-4000/config.json
+Model weights saved in ./checkpoint-4000/pytorch_model.bin
+Configuration saved in ./checkpoint-4000/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Deleting older checkpoint [checkpoint-2500] due to args.save_total_limit
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 19%|█████████████████████████████▋                                                                                                                              | 4100/21520 [3:22:57<7:45:28,  1.60s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 20%|██████████████████████████████▍                                                                                                                             | 4199/21520 [3:27:18<7:58:53,  1.66s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 20%|███████████████████████████████▏                                                                                                                            | 4299/21520 [3:31:42<8:00:50,  1.68s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 20%|███████████████████████████████▉                                                                                                                            | 4400/21520 [3:36:10<9:34:20,  2.01s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 21%|████████████████████████████████▌                                                                                                                           | 4500/21520 [3:40:32<9:18:48,  1.97s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+{'loss': 1.1671, 'learning_rate': 5.953496503496503e-05, 'epoch': 1.05}
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:56<00:00,  1.81s/it]
+
+Configuration saved in ./checkpoint-4500/config.json
+Model weights saved in ./checkpoint-4500/pytorch_model.bin
+Configuration saved in ./checkpoint-4500/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Deleting older checkpoint [checkpoint-3000] due to args.save_total_limit
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 21%|█████████████████████████████████▎                                                                                                                          | 4599/21520 [3:47:51<9:07:25,  1.94s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 22%|██████████████████████████████████                                                                                                                          | 4700/21520 [3:52:17<9:25:30,  2.02s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 22%|██████████████████████████████████▊                                                                                                                         | 4800/21520 [3:56:42<9:16:50,  2.00s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 23%|███████████████████████████████████▌                                                                                                                        | 4900/21520 [4:01:06<9:15:05,  2.00s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 23%|████████████████████████████████████▏                                                                                                                       | 5000/21520 [4:05:30<8:56:08,  1.95s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+{'loss': 1.1599, 'learning_rate': 5.779020979020979e-05, 'epoch': 1.16}
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Configuration saved in ./checkpoint-5000/config.json
+{'eval_loss': 0.3355213701725006, 'eval_wer': 0.4330439988027537, 'eval_runtime': 63.7849, 'eval_samples_per_second': 28.831, 'eval_steps_per_second': 0.455, 'epoch': 1.16}
+Model weights saved in ./checkpoint-5000/pytorch_model.bin
+Configuration saved in ./checkpoint-5000/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Deleting older checkpoint [checkpoint-3500] due to args.save_total_limit
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 24%|████████████████████████████████████▉                                                                                                                       | 5100/21520 [4:12:48<8:56:31,  1.96s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 24%|█████████████████████████████████████▋                                                                                                                      | 5200/21520 [4:17:12<9:28:46,  2.09s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 25%|██████████████████████████████████████▍                                                                                                                     | 5299/21520 [4:21:32<8:51:36,  1.97s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 25%|███████████████████████████████████████▏                                                                                                                    | 5399/21520 [4:25:55<9:07:16,  2.04s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 26%|███████████████████████████████████████▊                                                                                                                    | 5500/21520 [4:30:20<8:58:27,  2.02s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+{'loss': 1.1568, 'learning_rate': 5.6041958041958036e-05, 'epoch': 1.28}
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Configuration saved in ./checkpoint-5500/config.json
+{'eval_loss': 0.327897846698761, 'eval_wer': 0.42490272373540855, 'eval_runtime': 64.3828, 'eval_samples_per_second': 28.564, 'eval_steps_per_second': 0.45, 'epoch': 1.28}
+Model weights saved in ./checkpoint-5500/pytorch_model.bin
+Configuration saved in ./checkpoint-5500/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Deleting older checkpoint [checkpoint-4000] due to args.save_total_limit
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 26%|████████████████████████████████████████▌                                                                                                                   | 5599/21520 [4:37:38<8:44:34,  1.98s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 26%|█████████████████████████████████████████▎                                                                                                                  | 5699/21520 [4:42:01<8:33:59,  1.95s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 27%|██████████████████████████████████████████                                                                                                                  | 5799/21520 [4:46:25<8:30:31,  1.95s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 27%|██████████████████████████████████████████▊                                                                                                                 | 5899/21520 [4:50:48<8:50:56,  2.04s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 28%|███████████████████████████████████████████▍                                                                                                                | 6000/21520 [4:55:17<8:35:28,  1.99s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+{'loss': 1.0994, 'learning_rate': 5.4297202797202796e-05, 'epoch': 1.39}
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Configuration saved in ./checkpoint-6000/config.json
+{'eval_loss': 0.318962037563324, 'eval_wer': 0.4240646513020054, 'eval_runtime': 65.8584, 'eval_samples_per_second': 27.924, 'eval_steps_per_second': 0.44, 'epoch': 1.39}
+Model weights saved in ./checkpoint-6000/pytorch_model.bin
+Configuration saved in ./checkpoint-6000/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Deleting older checkpoint [checkpoint-4500] due to args.save_total_limit
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 28%|████████████████████████████████████████████▏                                                                                                               | 6099/21520 [5:02:30<8:27:43,  1.98s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 29%|████████████████████████████████████████████▉                                                                                                               | 6199/21520 [5:06:54<8:34:39,  2.02s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 29%|█████████████████████████████████████████████▋                                                                                                              | 6299/21520 [5:11:18<8:31:19,  2.02s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 30%|██████████████████████████████████████████████▍                                                                                                             | 6400/21520 [5:15:44<8:25:03,  2.00s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 30%|███████████████████████████████████████████████                                                                                                             | 6500/21520 [5:20:08<8:19:08,  1.99s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+{'loss': 1.1201, 'learning_rate': 5.2548951048951044e-05, 'epoch': 1.51}
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:57<00:00,  1.83s/it]
+
+Configuration saved in ./checkpoint-6500/config.json
+Model weights saved in ./checkpoint-6500/pytorch_model.bin
+Configuration saved in ./checkpoint-6500/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Deleting older checkpoint [checkpoint-5000] due to args.save_total_limit
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 31%|███████████████████████████████████████████████▊                                                                                                            | 6600/21520 [5:27:24<8:15:18,  1.99s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 31%|████████████████████████████████████████████████▌                                                                                                           | 6700/21520 [5:31:47<8:03:10,  1.96s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 32%|█████████████████████████████████████████████████▎                                                                                                          | 6799/21520 [5:36:08<7:56:35,  1.94s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 32%|██████████████████████████████████████████████████                                                                                                          | 6900/21520 [5:40:33<8:02:43,  1.98s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 33%|██████████████████████████████████████████████████▋                                                                                                         | 6999/21520 [5:44:52<7:54:39,  1.96s/it]
+ 33%|██████████████████████████████████████████████████▋                                                                                                         | 7000/21520 [5:44:54<7:55:25,  1.96s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+100%|████████████████��████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:54<00:00,  1.80s/it]
+
+Configuration saved in ./checkpoint-7000/config.json
+Model weights saved in ./checkpoint-7000/pytorch_model.bin
+Configuration saved in ./checkpoint-7000/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Deleting older checkpoint [checkpoint-5500] due to args.save_total_limit
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 33%|███████████████████████████████████████████████████▍                                                                                                        | 7100/21520 [5:52:08<8:10:16,  2.04s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 33%|████████████████████████████████████████████████████▏                                                                                                       | 7199/21520 [5:56:30<8:01:29,  2.02s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 34%|████████████████████████████████████████████████████▉                                                                                                       | 7299/21520 [6:00:54<7:59:22,  2.02s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 34%|█████████████████████████████████████████████████████▋                                                                                                      | 7399/21520 [6:05:20<7:52:51,  2.01s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 35%|██████████████████████████████████████████████████████▎                                                                                                     | 7499/21520 [6:09:43<7:42:55,  1.98s/it]
+ 35%|██████████████████████████████████████████████████████▎                                                                                                     | 7500/21520 [6:09:45<7:45:02,  1.99s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Configuration saved in ./checkpoint-7500/config.json
+{'eval_loss': 0.3057877719402313, 'eval_wer': 0.4125112241843759, 'eval_runtime': 65.7311, 'eval_samples_per_second': 27.978, 'eval_steps_per_second': 0.441, 'epoch': 1.74}
+Model weights saved in ./checkpoint-7500/pytorch_model.bin
+Configuration saved in ./checkpoint-7500/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Deleting older checkpoint [checkpoint-6000] due to args.save_total_limit
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 35%|███████████████████████████████████████████████████████                                                                                                     | 7600/21520 [6:16:59<7:31:32,  1.95s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 36%|███████████████████████████████████████████████████████▊                                                                                                    | 7699/21520 [6:21:22<7:59:09,  2.08s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 36%|████████████████████████████████████████████████████████▌                                                                                                   | 7800/21520 [6:25:47<7:34:58,  1.99s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 37%|█████████████████████████████████████████████████████████▎                                                                                                  | 7899/21520 [6:30:08<7:34:20,  2.00s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 37%|█████████████████████████████████████████████████████████▉                                                                                                  | 8000/21520 [6:34:33<7:26:34,  1.98s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+{'loss': 1.1101, 'learning_rate': 4.7311188811188806e-05, 'epoch': 1.86}
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Configuration saved in ./checkpoint-8000/config.json
+{'eval_loss': 0.3026394248008728, 'eval_wer': 0.4099970068841664, 'eval_runtime': 64.4321, 'eval_samples_per_second': 28.542, 'eval_steps_per_second': 0.45, 'epoch': 1.86}
+Model weights saved in ./checkpoint-8000/pytorch_model.bin
+Configuration saved in ./checkpoint-8000/preprocessor_config.json
+Configuration saved in ./preprocessor_config.json
+Deleting older checkpoint [checkpoint-6500] due to args.save_total_limit
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 38%|██████████████████████████████████████████████████████████▋                                                                                                 | 8099/21520 [6:41:43<7:19:00,  1.96s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 38%|███████████████████████████████████████████████████████████▍                                                                                                | 8200/21520 [6:46:08<7:12:30,  1.95s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 39%|████████████████████████████████████████████████████████████▏                                                                                               | 8301/21520 [6:50:33<7:07:31,  1.94s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 39%|████████████████████████████████████████████████████████████▉                                                                                               | 8400/21520 [6:54:55<7:36:42,  2.09s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 39%|█████████████████████████████████████████████████████████████▌                                                                                              | 8499/21520 [6:59:15<7:12:41,  1.99s/it]
+ 39%|█████████████████████████████████████████████████████████████▌                                                                                              | 8500/21520 [6:59:17<7:12:02,  1.99s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.
+***** Running Evaluation *****
+  Num examples = 1839
+  Batch size = 64
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:56<00:00,  1.81s/it]
+
+Configuration saved in ./checkpoint-8500/config.json
+Model weights saved in ./checkpoint-8500/pytorch_model.bin
+Configuration saved in ./checkpoint-8500/preprocessor_config.json

wandb/run-20220129_131141-h6nhqm30/files/requirements.txt

@@ -0,0 +1,180 @@
+aiohttp==3.8.1
+aiosignal==1.2.0
+analytics-python==1.4.0
+anyio==3.5.0
+appdirs==1.4.4
+argon2-cffi-bindings==21.2.0
+argon2-cffi==21.3.0
+asgiref==3.5.0
+asttokens==2.0.5
+async-timeout==4.0.2
+attrs==21.4.0
+audioread==2.1.9
+backcall==0.2.0
+backoff==1.10.0
+bcrypt==3.2.0
+beautifulsoup4==4.9.3
+black==21.12b0
+bleach==4.1.0
+brotlipy==0.7.0
+certifi==2020.12.5
+cffi==1.14.3
+chardet==3.0.4
+charset-normalizer==2.0.10
+click==8.0.3
+conda-build==3.21.4
+conda-package-handling==1.7.2
+conda==4.9.2
+configparser==5.2.0
+cryptography==3.2.1
+cycler==0.11.0
+datasets==1.18.2.dev0
+debugpy==1.5.1
+decorator==4.4.2
+defusedxml==0.7.1
+dill==0.3.4
+dnspython==2.1.0
+docker-pycreds==0.4.0
+entrypoints==0.3
+executing==0.8.2
+fastapi==0.73.0
+ffmpy==0.3.0
+filelock==3.0.12
+fonttools==4.29.0
+frozenlist==1.3.0
+fsspec==2022.1.0
+gitdb==4.0.9
+gitpython==3.1.26
+glob2==0.7
+gradio==2.7.5.2
+h11==0.13.0
+huggingface-hub==0.4.0
+idna==2.10
+importlib-resources==5.4.0
+ipykernel==6.7.0
+ipython-genutils==0.2.0
+ipython==8.0.1
+ipywidgets==7.6.3
+jedi==0.17.0
+jinja2==2.11.3
+jiwer==2.3.0
+joblib==1.1.0
+json5==0.9.6
+jsonschema==4.4.0
+jupyter-client==7.1.2
+jupyter-core==4.9.1
+jupyterlab-pygments==0.1.2
+jupyterlab-server==1.2.0
+jupyterlab-widgets==1.0.2
+jupyterlab==2.2.9
+kiwisolver==1.3.2
+libarchive-c==2.9
+librosa==0.8.1
+llvmlite==0.38.0
+markdown2==2.4.2
+markupsafe==1.1.1
+matplotlib-inline==0.1.3
+matplotlib==3.5.1
+mistune==0.8.4
+mkl-fft==1.3.0
+mkl-random==1.1.1
+mkl-service==2.3.0
+monotonic==1.6
+multidict==6.0.2
+multiprocess==0.70.12.2
+mypy-extensions==0.4.3
+nano==0.10.0
+nbclient==0.5.10
+nbconvert==6.4.1
+nbformat==5.1.3
+nest-asyncio==1.5.4
+notebook==6.4.8
+numba==0.55.1
+numpy==1.19.2
+olefile==0.46
+packaging==21.3
+pandas==1.4.0
+pandocfilters==1.5.0
+paramiko==2.9.2
+parso==0.8.1
+pathspec==0.9.0
+pathtools==0.1.2
+pexpect==4.8.0
+pickleshare==0.7.5
+pillow==8.1.2
+pip==21.3.1
+pkginfo==1.7.0
+platformdirs==2.4.1
+pooch==1.6.0
+prometheus-client==0.13.0
+promise==2.3
+prompt-toolkit==3.0.8
+protobuf==3.19.4
+psutil==5.8.0
+ptyprocess==0.7.0
+pure-eval==0.2.2
+pyarrow==6.0.1
+pycosat==0.6.3
+pycparser==2.20
+pycryptodome==3.13.0
+pydantic==1.9.0
+pydub==0.25.1
+pygments==2.8.0
+pynacl==1.5.0
+pyopenssl==19.1.0
+pyparsing==3.0.7
+pyrsistent==0.18.1
+pysocks==1.7.1
+python-dateutil==2.8.2
+python-etcd==0.4.5
+python-levenshtein==0.12.2
+python-multipart==0.0.5
+pytz==2021.1
+pyyaml==5.4.1
+pyzmq==22.3.0
+regex==2022.1.18
+requests==2.24.0
+resampy==0.2.2
+ruamel-yaml==0.15.87
+sacremoses==0.0.47
+scikit-learn==1.0.2
+scipy==1.7.3
+send2trash==1.8.0
+sentry-sdk==1.5.4
+setuptools==50.3.1.post20201107
+shortuuid==1.0.8
+six==1.15.0
+smmap==5.0.0
+sniffio==1.2.0
+soundfile==0.10.3.post1
+soupsieve==2.2
+stack-data==0.1.4
+starlette==0.17.1
+subprocess32==3.5.4
+termcolor==1.1.0
+terminado==0.13.1
+testpath==0.5.0
+threadpoolctl==3.0.0
+tokenizers==0.11.4
+tomli==1.2.3
+torch==1.10.2
+torchaudio==0.10.2
+torchelastic==0.2.2
+torchtext==0.9.1
+torchvision==0.9.1
+tornado==6.1
+tqdm==4.62.3
+traitlets==5.1.1
+transformers==4.17.0.dev0
+typing-extensions==4.0.1
+urllib3==1.25.11
+uvicorn==0.17.1
+wandb==0.12.9
+wcwidth==0.2.5
+webencodings==0.5.1
+wheel==0.35.1
+widgetsnbextension==3.5.2
+xxhash==2.0.2
+yarl==1.7.2
+yaspin==2.1.0
+zipp==3.7.0

wandb/run-20220129_131141-h6nhqm30/files/wandb-metadata.json

@@ -0,0 +1,64 @@
+{
+    "os": "Linux-4.15.0-151-generic-x86_64-with-glibc2.10",
+    "python": "3.8.8",
+    "heartbeatAt": "2022-01-29T13:11:43.886807",
+    "startedAt": "2022-01-29T13:11:41.626227",
+    "docker": null,
+    "gpu": "Tesla V100S-PCIE-32GB",
+    "gpu_count": 1,
+    "cpu_count": 60,
+    "cuda": null,
+    "args": [
+        "--dataset_name=mozilla-foundation/common_voice_8_0",
+        "--model_name_or_path=facebook/wav2vec2-xls-r-300m",
+        "--dataset_config_name=fr",
+        "--output_dir=./",
+        "--overwrite_output_dir",
+        "--num_train_epochs=5",
+        "--per_device_train_batch_size=64",
+        "--per_device_eval_batch_size=64",
+        "--gradient_accumulation_steps=1",
+        "--learning_rate=7e-5",
+        "--warmup_steps=1500",
+        "--length_column_name=input_length",
+        "--evaluation_strategy=steps",
+        "--text_column_name=sentence",
+        "--save_steps=500",
+        "--eval_steps=500",
+        "--logging_steps=100",
+        "--layerdrop=0.0",
+        "--activation_dropout=0.1",
+        "--save_total_limit=3",
+        "--freeze_feature_encoder",
+        "--feat_proj_dropout=0.0",
+        "--mask_time_prob=0.75",
+        "--mask_time_length=10",
+        "--mask_feature_prob=0.33",
+        "--mask_feature_length=10",
+        "--gradient_checkpointing",
+        "--report_to=wandb",
+        "--run_name=xls-r-300m-fr",
+        "--max_eval_samples=4000",
+        "--max_duration_in_seconds=5",
+        "--use_auth_token",
+        "--fp16",
+        "--group_by_length",
+        "--preprocessing_num_workers=64",
+        "--do_train",
+        "--do_eval",
+        "--load_best_model_at_end",
+        "--push_to_hub"
+    ],
+    "state": "running",
+    "program": "run_speech_recognition_ctc.py",
+    "codePath": "run_speech_recognition_ctc.py",
+    "git": {
+        "remote": "https://huggingface.co/AlexN/xls-r-300m-fr",
+        "commit": "44c47390840b2e07543aa79514aa51086a0ff3cf"
+    },
+    "email": "[email protected]",
+    "root": "/workspace/xls-r-300m-fr",
+    "host": "job-1abccd0a-3293-4ffe-8274-9e8f841f653f",
+    "username": "ovh",
+    "executable": "/opt/conda/bin/python"
+}

wandb/run-20220129_131141-h6nhqm30/logs/debug.log

@@ -0,0 +1,24 @@
+2022-01-29 13:11:41,630 INFO    MainThread:44227 [wandb_setup.py:_flush():71] setting env: {}
+2022-01-29 13:11:41,630 INFO    MainThread:44227 [wandb_setup.py:_flush():71] setting login settings: {}
+2022-01-29 13:11:41,631 INFO    MainThread:44227 [wandb_init.py:_log_setup():371] Logging user logs to /workspace/xls-r-300m-fr/wandb/run-20220129_131141-h6nhqm30/logs/debug.log
+2022-01-29 13:11:41,631 INFO    MainThread:44227 [wandb_init.py:_log_setup():372] Logging internal logs to /workspace/xls-r-300m-fr/wandb/run-20220129_131141-h6nhqm30/logs/debug-internal.log
+2022-01-29 13:11:41,631 INFO    MainThread:44227 [wandb_init.py:init():404] calling init triggers
+2022-01-29 13:11:41,631 INFO    MainThread:44227 [wandb_init.py:init():409] wandb.init called with sweep_config: {}
+config: {}
+2022-01-29 13:11:41,631 INFO    MainThread:44227 [wandb_init.py:init():460] starting backend
+2022-01-29 13:11:41,631 INFO    MainThread:44227 [backend.py:_multiprocessing_setup():99] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
+2022-01-29 13:11:42,043 INFO    MainThread:44227 [backend.py:ensure_launched():216] starting backend process...
+2022-01-29 13:11:42,437 INFO    MainThread:44227 [backend.py:ensure_launched():221] started backend process with pid: 13445
+2022-01-29 13:11:42,439 INFO    MainThread:44227 [wandb_init.py:init():469] backend started and connected
+2022-01-29 13:11:42,449 INFO    MainThread:44227 [wandb_init.py:init():533] updated telemetry
+2022-01-29 13:11:43,278 INFO    MainThread:44227 [wandb_init.py:init():563] communicating current version
+2022-01-29 13:11:43,680 INFO    MainThread:44227 [wandb_init.py:init():568] got version response 
+2022-01-29 13:11:43,681 INFO    MainThread:44227 [wandb_init.py:init():578] communicating run to backend with 30 second timeout
+2022-01-29 13:11:43,876 INFO    MainThread:44227 [wandb_init.py:init():606] starting run threads in backend
+2022-01-29 13:11:44,463 INFO    MainThread:44227 [wandb_run.py:_console_start():1810] atexit reg
+2022-01-29 13:11:44,464 INFO    MainThread:44227 [wandb_run.py:_redirect():1684] redirect: SettingsConsole.REDIRECT
+2022-01-29 13:11:44,464 INFO    MainThread:44227 [wandb_run.py:_redirect():1689] Redirecting console.
+2022-01-29 13:11:44,471 INFO    MainThread:44227 [wandb_run.py:_redirect():1745] Redirects installed.
+2022-01-29 13:11:44,471 INFO    MainThread:44227 [wandb_init.py:init():633] run started, returning control to user process
+2022-01-29 13:11:44,474 INFO    MainThread:44227 [wandb_run.py:_config_callback():956] config_cb None None {'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': 'float32', 'use_bfloat16': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'chunk_size_feed_forward': 0, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'architectures': ['Wav2Vec2ForPreTraining'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': 1, 'pad_token_id': 216, 'eos_token_id': 2, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'facebook/wav2vec2-xls-r-300m', 'transformers_version': '4.17.0.dev0', 'feat_extract_dropout': 0.0, 'model_type': 'wav2vec2', 'num_feat_extract_layers': 7, 'hidden_size': 1024, 'feat_extract_norm': 'layer', 'feat_extract_activation': 'gelu', 'conv_dim': [512, 512, 512, 512, 512, 512, 512], 'conv_stride': [5, 2, 2, 2, 2, 2, 2], 'conv_kernel': [10, 3, 3, 3, 3, 2, 2], 'conv_bias': True, 'num_conv_pos_embeddings': 128, 'num_conv_pos_embedding_groups': 16, 'num_hidden_layers': 24, 'intermediate_size': 4096, 'hidden_act': 'gelu', 'num_attention_heads': 16, 'hidden_dropout': 0.0, 'attention_dropout': 0.0, 'activation_dropout': 0.1, 'feat_proj_dropout': 0.0, 'final_dropout': 0.0, 'layerdrop': 0.0, 'layer_norm_eps': 1e-05, 'initializer_range': 0.02, 'vocab_size': 218, 'do_stable_layer_norm': True, 'use_weighted_layer_sum': False, 'apply_spec_augment': True, 'mask_time_prob': 0.75, 'mask_time_length': 10, 'mask_time_min_masks': 2, 'mask_feature_prob': 0.33, 'mask_feature_length': 10, 'mask_feature_min_masks': 0, 'num_codevectors_per_group': 320, 'num_codevector_groups': 2, 'contrastive_logits_temperature': 0.1, 'feat_quantizer_dropout': 0.0, 'num_negatives': 100, 'codevector_dim': 768, 'proj_codevector_dim': 768, 'diversity_loss_weight': 0.1, 'ctc_loss_reduction': 'mean', 'ctc_zero_infinity': False, 'add_adapter': False, 'adapter_kernel_size': 3, 'adapter_stride': 2, 'num_adapter_layers': 3, 'output_hidden_size': 1024, 'classifier_proj_size': 256, 'tdnn_dim': [512, 512, 512, 512, 1500], 'tdnn_kernel': [5, 3, 3, 1, 1], 'tdnn_dilation': [1, 2, 3, 1, 1], 'xvector_output_dim': 512, 'output_dir': './', 'overwrite_output_dir': True, 'do_train': True, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'steps', 'prediction_loss_only': False, 'per_device_train_batch_size': 64, 'per_device_eval_batch_size': 64, 'per_gpu_train_batch_size': 'None', 'per_gpu_eval_batch_size': 'None', 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': 'None', 'learning_rate': 7e-05, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 5.0, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'warmup_ratio': 0.0, 'warmup_steps': 1500, 'log_level': -1, 'log_level_replica': -1, 'log_on_each_node': True, 'logging_dir': './runs/Jan29_11-53-08_job-1abccd0a-3293-4ffe-8274-9e8f841f653f', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 100, 'logging_nan_inf_filter': True, 'save_strategy': 'steps', 'save_steps': 500, 'save_total_limit': 3, 'save_on_each_node': False, 'no_cuda': False, 'seed': 42, 'bf16': False, 'fp16': True, 'fp16_opt_level': 'O1', 'half_precision_backend': 'amp', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': 'None', 'local_rank': -1, 'xpu_backend': 'None', 'tpu_num_cores': 'None', 'tpu_metrics_debug': False, 'debug': '[]', 'dataloader_drop_last': False, 'eval_steps': 500, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': 'xls-r-300m-fr', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': 'None', 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'sharded_ddp': '[]', 'deepspeed': 'None', 'label_smoothing_factor': 0.0, 'optim': 'adamw_hf', 'adafactor': False, 'group_by_length': True, 'length_column_name': 'input_length', 'report_to': "['wandb']", 'ddp_find_unused_parameters': 'None', 'ddp_bucket_cap_mb': 'None', 'dataloader_pin_memory': True, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': 'None', 'hub_model_id': 'None', 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'gradient_checkpointing': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': 'None', 'push_to_hub_organization': 'None', 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', '_n_gpu': 1, 'mp_parameters': '', 'train_batch_size': 64, 'eval_batch_size': 64}
+2022-01-29 13:11:44,479 INFO    MainThread:44227 [wandb_watch.py:watch():43] Watching

wandb/run-20220129_131141-h6nhqm30/run-h6nhqm30.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:624404f6778fa63cf42e5d65fc81506fe3d0200ae05be92d55229cbe4a878ab4
+size 62254538

wandb/run-20220129_215451-1vipdbow/files/requirements.txt

@@ -0,0 +1,180 @@
+aiohttp==3.8.1
+aiosignal==1.2.0
+analytics-python==1.4.0
+anyio==3.5.0
+appdirs==1.4.4
+argon2-cffi-bindings==21.2.0
+argon2-cffi==21.3.0
+asgiref==3.5.0
+asttokens==2.0.5
+async-timeout==4.0.2
+attrs==21.4.0
+audioread==2.1.9
+backcall==0.2.0
+backoff==1.10.0
+bcrypt==3.2.0
+beautifulsoup4==4.9.3
+black==21.12b0
+bleach==4.1.0
+brotlipy==0.7.0
+certifi==2020.12.5
+cffi==1.14.3
+chardet==3.0.4
+charset-normalizer==2.0.10
+click==8.0.3
+conda-build==3.21.4
+conda-package-handling==1.7.2
+conda==4.9.2
+configparser==5.2.0
+cryptography==3.2.1
+cycler==0.11.0
+datasets==1.18.2.dev0
+debugpy==1.5.1
+decorator==4.4.2
+defusedxml==0.7.1
+dill==0.3.4
+dnspython==2.1.0
+docker-pycreds==0.4.0
+entrypoints==0.3
+executing==0.8.2
+fastapi==0.73.0
+ffmpy==0.3.0
+filelock==3.0.12
+fonttools==4.29.0
+frozenlist==1.3.0
+fsspec==2022.1.0
+gitdb==4.0.9
+gitpython==3.1.26
+glob2==0.7
+gradio==2.7.5.2
+h11==0.13.0
+huggingface-hub==0.4.0
+idna==2.10
+importlib-resources==5.4.0
+ipykernel==6.7.0
+ipython-genutils==0.2.0
+ipython==8.0.1
+ipywidgets==7.6.3
+jedi==0.17.0
+jinja2==2.11.3
+jiwer==2.3.0
+joblib==1.1.0
+json5==0.9.6
+jsonschema==4.4.0
+jupyter-client==7.1.2
+jupyter-core==4.9.1
+jupyterlab-pygments==0.1.2
+jupyterlab-server==1.2.0
+jupyterlab-widgets==1.0.2
+jupyterlab==2.2.9
+kiwisolver==1.3.2
+libarchive-c==2.9
+librosa==0.8.1
+llvmlite==0.38.0
+markdown2==2.4.2
+markupsafe==1.1.1
+matplotlib-inline==0.1.3
+matplotlib==3.5.1
+mistune==0.8.4
+mkl-fft==1.3.0
+mkl-random==1.1.1
+mkl-service==2.3.0
+monotonic==1.6
+multidict==6.0.2
+multiprocess==0.70.12.2
+mypy-extensions==0.4.3
+nano==0.10.0
+nbclient==0.5.10
+nbconvert==6.4.1
+nbformat==5.1.3
+nest-asyncio==1.5.4
+notebook==6.4.8
+numba==0.55.1
+numpy==1.19.2
+olefile==0.46
+packaging==21.3
+pandas==1.4.0
+pandocfilters==1.5.0
+paramiko==2.9.2
+parso==0.8.1
+pathspec==0.9.0
+pathtools==0.1.2
+pexpect==4.8.0
+pickleshare==0.7.5
+pillow==8.1.2
+pip==21.3.1
+pkginfo==1.7.0
+platformdirs==2.4.1
+pooch==1.6.0
+prometheus-client==0.13.0
+promise==2.3
+prompt-toolkit==3.0.8
+protobuf==3.19.4
+psutil==5.8.0
+ptyprocess==0.7.0
+pure-eval==0.2.2
+pyarrow==6.0.1
+pycosat==0.6.3
+pycparser==2.20
+pycryptodome==3.13.0
+pydantic==1.9.0
+pydub==0.25.1
+pygments==2.8.0
+pynacl==1.5.0
+pyopenssl==19.1.0
+pyparsing==3.0.7
+pyrsistent==0.18.1
+pysocks==1.7.1
+python-dateutil==2.8.2
+python-etcd==0.4.5
+python-levenshtein==0.12.2
+python-multipart==0.0.5
+pytz==2021.1
+pyyaml==5.4.1
+pyzmq==22.3.0
+regex==2022.1.18
+requests==2.24.0
+resampy==0.2.2
+ruamel-yaml==0.15.87
+sacremoses==0.0.47
+scikit-learn==1.0.2
+scipy==1.7.3
+send2trash==1.8.0
+sentry-sdk==1.5.4
+setuptools==50.3.1.post20201107
+shortuuid==1.0.8
+six==1.15.0
+smmap==5.0.0
+sniffio==1.2.0
+soundfile==0.10.3.post1
+soupsieve==2.2
+stack-data==0.1.4
+starlette==0.17.1
+subprocess32==3.5.4
+termcolor==1.1.0
+terminado==0.13.1
+testpath==0.5.0
+threadpoolctl==3.0.0
+tokenizers==0.11.4
+tomli==1.2.3
+torch==1.10.2
+torchaudio==0.10.2
+torchelastic==0.2.2
+torchtext==0.9.1
+torchvision==0.9.1
+tornado==6.1
+tqdm==4.62.3
+traitlets==5.1.1
+transformers==4.17.0.dev0
+typing-extensions==4.0.1
+urllib3==1.25.11
+uvicorn==0.17.1
+wandb==0.12.9
+wcwidth==0.2.5
+webencodings==0.5.1
+wheel==0.35.1
+widgetsnbextension==3.5.2
+xxhash==2.0.2
+yarl==1.7.2
+yaspin==2.1.0
+zipp==3.7.0

wandb/run-20220129_215451-1vipdbow/files/wandb-metadata.json

@@ -0,0 +1,65 @@
+{
+    "os": "Linux-4.15.0-151-generic-x86_64-with-glibc2.10",
+    "python": "3.8.8",
+    "heartbeatAt": "2022-01-29T21:54:53.025463",
+    "startedAt": "2022-01-29T21:54:51.704109",
+    "docker": null,
+    "gpu": "Tesla V100S-PCIE-32GB",
+    "gpu_count": 1,
+    "cpu_count": 60,
+    "cuda": null,
+    "args": [
+        "--dataset_name=mozilla-foundation/common_voice_8_0",
+        "--model_name_or_path=facebook/wav2vec2-xls-r-300m",
+        "--dataset_config_name=fr",
+        "--tokenizer_name_or_path=./",
+        "--output_dir=./",
+        "--overwrite_output_dir",
+        "--num_train_epochs=2",
+        "--per_device_train_batch_size=64",
+        "--per_device_eval_batch_size=64",
+        "--gradient_accumulation_steps=1",
+        "--learning_rate=1e-4",
+        "--warmup_steps=1500",
+        "--length_column_name=input_length",
+        "--evaluation_strategy=steps",
+        "--text_column_name=sentence",
+        "--save_steps=500",
+        "--eval_steps=500",
+        "--logging_steps=100",
+        "--layerdrop=0.0",
+        "--activation_dropout=0.05",
+        "--save_total_limit=2",
+        "--freeze_feature_encoder",
+        "--feat_proj_dropout=0.0",
+        "--mask_time_prob=0.75",
+        "--mask_time_length=10",
+        "--mask_feature_prob=0.4",
+        "--mask_feature_length=10",
+        "--gradient_checkpointing",
+        "--report_to=wandb",
+        "--run_name=xls-r-300m-fr",
+        "--max_eval_samples=6000",
+        "--max_duration_in_seconds=9",
+        "--use_auth_token",
+        "--fp16",
+        "--group_by_length",
+        "--preprocessing_num_workers=64",
+        "--do_train",
+        "--do_eval",
+        "--load_best_model_at_end",
+        "--push_to_hub"
+    ],
+    "state": "running",
+    "program": "run_speech_recognition_ctc.py",
+    "codePath": "run_speech_recognition_ctc.py",
+    "git": {
+        "remote": "https://huggingface.co/AlexN/xls-r-300m-fr",
+        "commit": "c58dbaf4476093ff9758e2d11bbf63f828a0ecc3"
+    },
+    "email": "[email protected]",
+    "root": "/workspace/xls-r-300m-fr",
+    "host": "job-1abccd0a-3293-4ffe-8274-9e8f841f653f",
+    "username": "ovh",
+    "executable": "/opt/conda/bin/python"
+}

wandb/run-20220129_215451-1vipdbow/logs/debug.log

@@ -0,0 +1,24 @@
+2022-01-29 21:54:51,708 INFO    MainThread:7453 [wandb_setup.py:_flush():71] setting env: {}
+2022-01-29 21:54:51,708 INFO    MainThread:7453 [wandb_setup.py:_flush():71] setting login settings: {}
+2022-01-29 21:54:51,708 INFO    MainThread:7453 [wandb_init.py:_log_setup():371] Logging user logs to /workspace/xls-r-300m-fr/wandb/run-20220129_215451-1vipdbow/logs/debug.log
+2022-01-29 21:54:51,709 INFO    MainThread:7453 [wandb_init.py:_log_setup():372] Logging internal logs to /workspace/xls-r-300m-fr/wandb/run-20220129_215451-1vipdbow/logs/debug-internal.log
+2022-01-29 21:54:51,709 INFO    MainThread:7453 [wandb_init.py:init():404] calling init triggers
+2022-01-29 21:54:51,709 INFO    MainThread:7453 [wandb_init.py:init():409] wandb.init called with sweep_config: {}
+config: {}
+2022-01-29 21:54:51,709 INFO    MainThread:7453 [wandb_init.py:init():460] starting backend
+2022-01-29 21:54:51,709 INFO    MainThread:7453 [backend.py:_multiprocessing_setup():99] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
+2022-01-29 21:54:51,806 INFO    MainThread:7453 [backend.py:ensure_launched():216] starting backend process...
+2022-01-29 21:54:51,893 INFO    MainThread:7453 [backend.py:ensure_launched():221] started backend process with pid: 37713
+2022-01-29 21:54:51,897 INFO    MainThread:7453 [wandb_init.py:init():469] backend started and connected
+2022-01-29 21:54:51,906 INFO    MainThread:7453 [wandb_init.py:init():533] updated telemetry
+2022-01-29 21:54:52,074 INFO    MainThread:7453 [wandb_init.py:init():563] communicating current version
+2022-01-29 21:54:52,818 INFO    MainThread:7453 [wandb_init.py:init():568] got version response 
+2022-01-29 21:54:52,818 INFO    MainThread:7453 [wandb_init.py:init():578] communicating run to backend with 30 second timeout
+2022-01-29 21:54:53,017 INFO    MainThread:7453 [wandb_init.py:init():606] starting run threads in backend
+2022-01-29 21:54:53,621 INFO    MainThread:7453 [wandb_run.py:_console_start():1810] atexit reg
+2022-01-29 21:54:53,622 INFO    MainThread:7453 [wandb_run.py:_redirect():1684] redirect: SettingsConsole.REDIRECT
+2022-01-29 21:54:53,623 INFO    MainThread:7453 [wandb_run.py:_redirect():1689] Redirecting console.
+2022-01-29 21:54:53,628 INFO    MainThread:7453 [wandb_run.py:_redirect():1745] Redirects installed.
+2022-01-29 21:54:53,628 INFO    MainThread:7453 [wandb_init.py:init():633] run started, returning control to user process
+2022-01-29 21:54:53,630 INFO    MainThread:7453 [wandb_run.py:_config_callback():956] config_cb None None {'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': 'float32', 'use_bfloat16': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'chunk_size_feed_forward': 0, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'architectures': ['Wav2Vec2ForPreTraining'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': 1, 'pad_token_id': 216, 'eos_token_id': 2, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'facebook/wav2vec2-xls-r-300m', 'transformers_version': '4.17.0.dev0', 'feat_extract_dropout': 0.0, 'model_type': 'wav2vec2', 'num_feat_extract_layers': 7, 'hidden_size': 1024, 'feat_extract_norm': 'layer', 'feat_extract_activation': 'gelu', 'conv_dim': [512, 512, 512, 512, 512, 512, 512], 'conv_stride': [5, 2, 2, 2, 2, 2, 2], 'conv_kernel': [10, 3, 3, 3, 3, 2, 2], 'conv_bias': True, 'num_conv_pos_embeddings': 128, 'num_conv_pos_embedding_groups': 16, 'num_hidden_layers': 24, 'intermediate_size': 4096, 'hidden_act': 'gelu', 'num_attention_heads': 16, 'hidden_dropout': 0.0, 'attention_dropout': 0.0, 'activation_dropout': 0.05, 'feat_proj_dropout': 0.0, 'final_dropout': 0.0, 'layerdrop': 0.0, 'layer_norm_eps': 1e-05, 'initializer_range': 0.02, 'vocab_size': 218, 'do_stable_layer_norm': True, 'use_weighted_layer_sum': False, 'apply_spec_augment': True, 'mask_time_prob': 0.75, 'mask_time_length': 10, 'mask_time_min_masks': 2, 'mask_feature_prob': 0.4, 'mask_feature_length': 10, 'mask_feature_min_masks': 0, 'num_codevectors_per_group': 320, 'num_codevector_groups': 2, 'contrastive_logits_temperature': 0.1, 'feat_quantizer_dropout': 0.0, 'num_negatives': 100, 'codevector_dim': 768, 'proj_codevector_dim': 768, 'diversity_loss_weight': 0.1, 'ctc_loss_reduction': 'mean', 'ctc_zero_infinity': False, 'add_adapter': False, 'adapter_kernel_size': 3, 'adapter_stride': 2, 'num_adapter_layers': 3, 'output_hidden_size': 1024, 'classifier_proj_size': 256, 'tdnn_dim': [512, 512, 512, 512, 1500], 'tdnn_kernel': [5, 3, 3, 1, 1], 'tdnn_dilation': [1, 2, 3, 1, 1], 'xvector_output_dim': 512, 'output_dir': './', 'overwrite_output_dir': True, 'do_train': True, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'steps', 'prediction_loss_only': False, 'per_device_train_batch_size': 64, 'per_device_eval_batch_size': 64, 'per_gpu_train_batch_size': 'None', 'per_gpu_eval_batch_size': 'None', 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': 'None', 'learning_rate': 0.0001, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2.0, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'warmup_ratio': 0.0, 'warmup_steps': 1500, 'log_level': -1, 'log_level_replica': -1, 'log_on_each_node': True, 'logging_dir': './runs/Jan29_20-37-14_job-1abccd0a-3293-4ffe-8274-9e8f841f653f', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 100, 'logging_nan_inf_filter': True, 'save_strategy': 'steps', 'save_steps': 500, 'save_total_limit': 2, 'save_on_each_node': False, 'no_cuda': False, 'seed': 42, 'bf16': False, 'fp16': True, 'fp16_opt_level': 'O1', 'half_precision_backend': 'amp', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': 'None', 'local_rank': -1, 'xpu_backend': 'None', 'tpu_num_cores': 'None', 'tpu_metrics_debug': False, 'debug': '[]', 'dataloader_drop_last': False, 'eval_steps': 500, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': 'xls-r-300m-fr', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': 'None', 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'sharded_ddp': '[]', 'deepspeed': 'None', 'label_smoothing_factor': 0.0, 'optim': 'adamw_hf', 'adafactor': False, 'group_by_length': True, 'length_column_name': 'input_length', 'report_to': "['wandb']", 'ddp_find_unused_parameters': 'None', 'ddp_bucket_cap_mb': 'None', 'dataloader_pin_memory': True, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': 'None', 'hub_model_id': 'None', 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'gradient_checkpointing': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': 'None', 'push_to_hub_organization': 'None', 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', '_n_gpu': 1, 'mp_parameters': '', 'train_batch_size': 64, 'eval_batch_size': 64}
+2022-01-29 21:54:53,634 INFO    MainThread:7453 [wandb_watch.py:watch():43] Watching

wandb/run-20220129_215451-1vipdbow/run-1vipdbow.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:51e1b275cfd48d2b806baadf5c43eb7f3d8efffce1a2b58da8b720f5bf6e721d
+size 104538108