End of training

.gitignore

@@ -0,0 +1 @@
+checkpoint-*/

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 9.97,
+    "eval_loss": 3.5006210803985596,
+    "eval_runtime": 5.9779,
+    "eval_samples": 99,
+    "eval_samples_per_second": 16.561,
+    "eval_steps_per_second": 1.171,
+    "eval_wer": 1.0,
+    "train_loss": 5.883141581217448,
+    "train_runtime": 4312.7885,
+    "train_samples": 990,
+    "train_samples_per_second": 2.295,
+    "train_steps_per_second": 0.035
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 9.97,
+    "eval_loss": 3.5006210803985596,
+    "eval_runtime": 5.9779,
+    "eval_samples": 99,
+    "eval_samples_per_second": 16.561,
+    "eval_steps_per_second": 1.171,
+    "eval_wer": 1.0
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55429e58f4eb26fda2b961b91a4dd01fb47f8aa7726aebde52b93d4d8d08f34a
+size 1262075377

run.sh

@@ -4,6 +4,7 @@ python run_speech_recognition_ctc.py \
 	--dataset_config_name="48K_mp3" \
 	--output_dir="./" \
 	--overwrite_output_dir \
+        --hub_model_id="NbAiLab/xls-npsc" \
 	--num_train_epochs="10" \
 	--per_device_train_batch_size="8" \
 	--per_device_eval_batch_size="8" \
@@ -31,4 +32,6 @@ python run_speech_recognition_ctc.py \
 	--fp16 \
 	--group_by_length \
 	--do_train --do_eval \
+        --max_train_samples="1000" \
+        --max_eval_samples="100"  \
 	--push_to_hub

runs/Jan25_14-51-02_ficino/1643118713.0966976/events.out.tfevents.1643118713.ficino.3248974.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3555c0c0a150e385db0b5031964177d10a7c633791a951b63d8ec5312ee13509
+size 4740

runs/Jan25_14-51-02_ficino/events.out.tfevents.1643118713.ficino.3248974.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af15f65528353fa4db3871698c47b5194bf87721b62db855027380cda424ad26
+size 5182

runs/Jan25_14-51-02_ficino/events.out.tfevents.1643123033.ficino.3248974.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0eceb881945082b4d551472836fd7eda3a0627c3ef9a4bee226ef3fabfc41392
+size 358

special_tokens_map.json

@@ -1 +1 @@
-{"bos_token": "<s>", "eos_token": "</s>", "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}]}
+{"bos_token": "<s>", "eos_token": "</s>", "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}, {"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}]}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 9.97,
+    "train_loss": 5.883141581217448,
+    "train_runtime": 4312.7885,
+    "train_samples": 990,
+    "train_samples_per_second": 2.295,
+    "train_steps_per_second": 0.035
+}

trainer_state.json

@@ -0,0 +1,31 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 9.967741935483872,
+  "global_step": 150,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 6.65,
+      "learning_rate": 1.4999999999999999e-05,
+      "loss": 6.9618,
+      "step": 100
+    },
+    {
+      "epoch": 9.97,
+      "step": 150,
+      "total_flos": 2.264182160412142e+18,
+      "train_loss": 5.883141581217448,
+      "train_runtime": 4312.7885,
+      "train_samples_per_second": 2.295,
+      "train_steps_per_second": 0.035
+    }
+  ],
+  "max_steps": 150,
+  "num_train_epochs": 10,
+  "total_flos": 2.264182160412142e+18,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7469ac235d7305d53f498fff8ceaeb79ff708b37e6e935641044ed8bad4c5f81
+oid sha256:2acea8b96056a730756b356a273dd919c15332d43f021e9bac157e290e7b2389
 size 2991

vocab.json

@@ -1 +1 @@
-{"a": 1, "b": 2, "c": 3, "d": 4, "e": 5, "f": 6, "g": 7, "h": 8, "i": 9, "j": 10, "k": 11, "l": 12, "m": 13, "n": 14, "o": 15, "p": 16, "q": 17, "r": 18, "s": 19, "t": 20, "u": 21, "v": 22, "w": 23, "x": 24, "y": 25, "z": 26, "ä": 27, "å": 28, "é": 29, "ô": 30, "ö": 31, "ü": 32, "|": 0, "[UNK]": 33, "[PAD]": 34}
+{"a": 1, "b": 2, "c": 3, "d": 4, "e": 5, "f": 6, "g": 7, "h": 8, "i": 9, "j": 10, "k": 11, "l": 12, "m": 13, "n": 14, "o": 15, "p": 16, "q": 17, "r": 18, "s": 19, "t": 20, "u": 21, "v": 22, "w": 23, "x": 24, "y": 25, "z": 26, "å": 27, "æ": 28, "è": 29, "ò": 30, "ô": 31, "ø": 32, "|": 0, "[UNK]": 33, "[PAD]": 34}

wandb/debug-internal.log

@@ -0,0 +1 @@
+run-20220125_145153-1ghti100/logs/debug-internal.log

wandb/debug.log

@@ -0,0 +1 @@
+run-20220125_145153-1ghti100/logs/debug.log

wandb/latest-run

@@ -0,0 +1 @@
+run-20220125_145153-1ghti100

wandb/run-20220125_145153-1ghti100/files/code/run_speech_recognition_ctc.py

@@ -0,0 +1,758 @@
+#!/usr/bin/env python
+# coding=utf-8
+# Copyright 2021 The HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+
+""" Fine-tuning a 🤗 Transformers CTC model for automatic speech recognition"""
+
+import functools
+import json
+import logging
+import os
+import re
+import sys
+import warnings
+from dataclasses import dataclass, field
+from typing import Dict, List, Optional, Union
+
+import datasets
+import numpy as np
+import torch
+from datasets import DatasetDict, load_dataset, load_metric
+
+import transformers
+from transformers import (
+    AutoConfig,
+    AutoFeatureExtractor,
+    AutoModelForCTC,
+    AutoProcessor,
+    AutoTokenizer,
+    HfArgumentParser,
+    Trainer,
+    TrainingArguments,
+    Wav2Vec2Processor,
+    set_seed,
+)
+from transformers.trainer_utils import get_last_checkpoint, is_main_process
+from transformers.utils import check_min_version
+from transformers.utils.versions import require_version
+
+
+# Will error if the minimal version of Transformers is not installed. Remove at your own risks.
+check_min_version("4.16.0.dev0")
+
+require_version("datasets>=1.13.3", "To fix: pip install -r examples/pytorch/text-classification/requirements.txt")
+
+
+logger = logging.getLogger(__name__)
+
+
+def list_field(default=None, metadata=None):
+    return field(default_factory=lambda: default, metadata=metadata)
+
+
+@dataclass
+class ModelArguments:
+    """
+    Arguments pertaining to which model/config/tokenizer we are going to fine-tune from.
+    """
+
+    model_name_or_path: str = field(
+        metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models"}
+    )
+    tokenizer_name_or_path: Optional[str] = field(
+        default=None,
+        metadata={"help": "Path to pretrained tokenizer or tokenizer identifier from huggingface.co/models"},
+    )
+    cache_dir: Optional[str] = field(
+        default=None,
+        metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"},
+    )
+    freeze_feature_encoder: bool = field(
+        default=True, metadata={"help": "Whether to freeze the feature encoder layers of the model."}
+    )
+    attention_dropout: float = field(
+        default=0.0, metadata={"help": "The dropout ratio for the attention probabilities."}
+    )
+    activation_dropout: float = field(
+        default=0.0, metadata={"help": "The dropout ratio for activations inside the fully connected layer."}
+    )
+    feat_proj_dropout: float = field(default=0.0, metadata={"help": "The dropout ratio for the projected features."})
+    hidden_dropout: float = field(
+        default=0.0,
+        metadata={
+            "help": "The dropout probability for all fully connected layers in the embeddings, encoder, and pooler."
+        },
+    )
+    final_dropout: float = field(
+        default=0.0,
+        metadata={"help": "The dropout probability for the final projection layer."},
+    )
+    mask_time_prob: float = field(
+        default=0.05,
+        metadata={
+            "help": "Probability of each feature vector along the time axis to be chosen as the start of the vector"
+            "span to be masked. Approximately ``mask_time_prob * sequence_length // mask_time_length`` feature"
+            "vectors will be masked along the time axis."
+        },
+    )
+    mask_time_length: int = field(
+        default=10,
+        metadata={"help": "Length of vector span to mask along the time axis."},
+    )
+    mask_feature_prob: float = field(
+        default=0.0,
+        metadata={
+            "help": "Probability of each feature vector along the feature axis to be chosen as the start of the vector"
+            "span to be masked. Approximately ``mask_feature_prob * sequence_length // mask_feature_length`` feature bins will be masked along the time axis."
+        },
+    )
+    mask_feature_length: int = field(
+        default=10,
+        metadata={"help": "Length of vector span to mask along the feature axis."},
+    )
+    layerdrop: float = field(default=0.0, metadata={"help": "The LayerDrop probability."})
+    ctc_loss_reduction: Optional[str] = field(
+        default="mean", metadata={"help": "The way the ctc loss should be reduced. Should be one of 'mean' or 'sum'."}
+    )
+
+
+@dataclass
+class DataTrainingArguments:
+    """
+    Arguments pertaining to what data we are going to input our model for training and eval.
+
+    Using `HfArgumentParser` we can turn this class
+    into argparse arguments to be able to specify them on
+    the command line.
+    """
+
+    dataset_name: str = field(
+        metadata={"help": "The configuration name of the dataset to use (via the datasets library)."}
+    )
+    dataset_config_name: str = field(
+        default=None, metadata={"help": "The configuration name of the dataset to use (via the datasets library)."}
+    )
+    train_split_name: str = field(
+        default="train+validation",
+        metadata={
+            "help": "The name of the training data set split to use (via the datasets library). Defaults to 'train'"
+        },
+    )
+    eval_split_name: str = field(
+        default="test",
+        metadata={
+            "help": "The name of the training data set split to use (via the datasets library). Defaults to 'train'"
+        },
+    )
+    audio_column_name: str = field(
+        default="audio",
+        metadata={"help": "The name of the dataset column containing the audio data. Defaults to 'audio'"},
+    )
+    text_column_name: str = field(
+        default="text",
+        metadata={"help": "The name of the dataset column containing the text data. Defaults to 'text'"},
+    )
+    overwrite_cache: bool = field(
+        default=False, metadata={"help": "Overwrite the cached preprocessed datasets or not."}
+    )
+    preprocessing_num_workers: Optional[int] = field(
+        default=None,
+        metadata={"help": "The number of processes to use for the preprocessing."},
+    )
+    max_train_samples: Optional[int] = field(
+        default=None,
+        metadata={
+            "help": "For debugging purposes or quicker training, truncate the number of training examples to this "
+            "value if set."
+        },
+    )
+    max_eval_samples: Optional[int] = field(
+        default=None,
+        metadata={
+            "help": "For debugging purposes or quicker training, truncate the number of validation examples to this "
+            "value if set."
+        },
+    )
+    chars_to_ignore: Optional[List[str]] = list_field(
+        default=None,
+        metadata={"help": "A list of characters to remove from the transcripts."},
+    )
+    eval_metrics: List[str] = list_field(
+        default=["wer"],
+        metadata={"help": "A list of metrics the model should be evaluated on. E.g. `'wer cer'`"},
+    )
+    max_duration_in_seconds: float = field(
+        default=20.0,
+        metadata={
+            "help": "Filter audio files that are longer than `max_duration_in_seconds` seconds to 'max_duration_in_seconds`"
+        },
+    )
+    min_duration_in_seconds: float = field(
+        default=0.0, metadata={"help": "Filter audio files that are shorter than `min_duration_in_seconds` seconds"}
+    )
+    preprocessing_only: bool = field(
+        default=False,
+        metadata={
+            "help": "Whether to only do data preprocessing and skip training. "
+            "This is especially useful when data preprocessing errors out in distributed training due to timeout. "
+            "In this case, one should run the preprocessing in a non-distributed setup with `preprocessing_only=True` "
+            "so that the cached datasets can consequently be loaded in distributed training"
+        },
+    )
+    use_auth_token: bool = field(
+        default=False,
+        metadata={
+            "help": "If :obj:`True`, will use the token generated when running"
+            ":obj:`transformers-cli login` as HTTP bearer authorization for remote files."
+        },
+    )
+    unk_token: str = field(
+        default="[UNK]",
+        metadata={"help": "The unk token for the tokenizer"},
+    )
+    pad_token: str = field(
+        default="[PAD]",
+        metadata={"help": "The padding token for the tokenizer"},
+    )
+    word_delimiter_token: str = field(
+        default="|",
+        metadata={"help": "The word delimiter token for the tokenizer"},
+    )
+    phoneme_language: Optional[str] = field(
+        default=None,
+        metadata={
+            "help": "The target language that should be used be"
+            " passed to the tokenizer for tokenization. Note that"
+            " this is only relevant if the model classifies the"
+            " input audio to a sequence of phoneme sequences."
+        },
+    )
+
+
+@dataclass
+class DataCollatorCTCWithPadding:
+    """
+    Data collator that will dynamically pad the inputs received.
+    Args:
+        processor (:class:`~transformers.AutoProcessor`)
+            The processor used for proccessing the data.
+        padding (:obj:`bool`, :obj:`str` or :class:`~transformers.tokenization_utils_base.PaddingStrategy`, `optional`, defaults to :obj:`True`):
+            Select a strategy to pad the returned sequences (according to the model's padding side and padding index)
+            among:
+            * :obj:`True` or :obj:`'longest'`: Pad to the longest sequence in the batch (or no padding if only a single
+              sequence if provided).
+            * :obj:`'max_length'`: Pad to a maximum length specified with the argument :obj:`max_length` or to the
+              maximum acceptable input length for the model if that argument is not provided.
+            * :obj:`False` or :obj:`'do_not_pad'` (default): No padding (i.e., can output a batch with sequences of
+              different lengths).
+        max_length (:obj:`int`, `optional`):
+            Maximum length of the ``input_values`` of the returned list and optionally padding length (see above).
+        max_length_labels (:obj:`int`, `optional`):
+            Maximum length of the ``labels`` returned list and optionally padding length (see above).
+        pad_to_multiple_of (:obj:`int`, `optional`):
+            If set will pad the sequence to a multiple of the provided value.
+            This is especially useful to enable the use of Tensor Cores on NVIDIA hardware with compute capability >=
+            7.5 (Volta).
+    """
+
+    processor: AutoProcessor
+    padding: Union[bool, str] = "longest"
+    pad_to_multiple_of: Optional[int] = None
+    pad_to_multiple_of_labels: Optional[int] = None
+
+    def __call__(self, features: List[Dict[str, Union[List[int], torch.Tensor]]]) -> Dict[str, torch.Tensor]:
+        # split inputs and labels since they have to be of different lenghts and need
+        # different padding methods
+        input_features = [{"input_values": feature["input_values"]} for feature in features]
+        label_features = [{"input_ids": feature["labels"]} for feature in features]
+
+        batch = self.processor.pad(
+            input_features,
+            padding=self.padding,
+            pad_to_multiple_of=self.pad_to_multiple_of,
+            return_tensors="pt",
+        )
+
+        with self.processor.as_target_processor():
+            labels_batch = self.processor.pad(
+                label_features,
+                padding=self.padding,
+                pad_to_multiple_of=self.pad_to_multiple_of_labels,
+                return_tensors="pt",
+            )
+
+        # replace padding with -100 to ignore loss correctly
+        labels = labels_batch["input_ids"].masked_fill(labels_batch.attention_mask.ne(1), -100)
+
+        batch["labels"] = labels
+
+        return batch
+
+
+def create_vocabulary_from_data(
+    datasets: DatasetDict,
+    word_delimiter_token: Optional[str] = None,
+    unk_token: Optional[str] = None,
+    pad_token: Optional[str] = None,
+):
+    # Given training and test labels create vocabulary
+    def extract_all_chars(batch):
+        all_text = " ".join(batch["target_text"])
+        vocab = list(set(all_text))
+        return {"vocab": [vocab], "all_text": [all_text]}
+
+    vocabs = datasets.map(
+        extract_all_chars,
+        batched=True,
+        batch_size=-1,
+        keep_in_memory=True,
+        remove_columns=datasets["train"].column_names,
+    )
+
+    # take union of all unique characters in each dataset
+    vocab_set = functools.reduce(
+        lambda vocab_1, vocab_2: set(vocab_1["vocab"][0]) | set(vocab_2["vocab"][0]), vocabs.values()
+    )
+
+    vocab_dict = {v: k for k, v in enumerate(sorted(list(vocab_set)))}
+
+    # replace white space with delimiter token
+    if word_delimiter_token is not None:
+        vocab_dict[word_delimiter_token] = vocab_dict[" "]
+        del vocab_dict[" "]
+
+    # add unk and pad token
+    if unk_token is not None:
+        vocab_dict[unk_token] = len(vocab_dict)
+
+    if pad_token is not None:
+        vocab_dict[pad_token] = len(vocab_dict)
+
+    return vocab_dict
+
+
+def main():
+    # See all possible arguments in src/transformers/training_args.py
+    # or by passing the --help flag to this script.
+    # We now keep distinct sets of args, for a cleaner separation of concerns.
+
+    parser = HfArgumentParser((ModelArguments, DataTrainingArguments, TrainingArguments))
+    if len(sys.argv) == 2 and sys.argv[1].endswith(".json"):
+        # If we pass only one argument to the script and it's the path to a json file,
+        # let's parse it to get our arguments.
+        model_args, data_args, training_args = parser.parse_json_file(json_file=os.path.abspath(sys.argv[1]))
+    else:
+        model_args, data_args, training_args = parser.parse_args_into_dataclasses()
+
+    # Detecting last checkpoint.
+    last_checkpoint = None
+    if os.path.isdir(training_args.output_dir) and training_args.do_train and not training_args.overwrite_output_dir:
+        last_checkpoint = get_last_checkpoint(training_args.output_dir)
+        if last_checkpoint is None and len(os.listdir(training_args.output_dir)) > 0:
+            raise ValueError(
+                f"Output directory ({training_args.output_dir}) already exists and is not empty. "
+                "Use --overwrite_output_dir to overcome."
+            )
+        elif last_checkpoint is not None:
+            logger.info(
+                f"Checkpoint detected, resuming training at {last_checkpoint}. To avoid this behavior, change "
+                "the `--output_dir` or add `--overwrite_output_dir` to train from scratch."
+            )
+
+    # Setup logging
+    logging.basicConfig(
+        format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
+        datefmt="%m/%d/%Y %H:%M:%S",
+        handlers=[logging.StreamHandler(sys.stdout)],
+    )
+    logger.setLevel(logging.INFO if is_main_process(training_args.local_rank) else logging.WARN)
+
+    # Log on each process the small summary:
+    logger.warning(
+        f"Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu}"
+        f"distributed training: {bool(training_args.local_rank != -1)}, 16-bits training: {training_args.fp16}"
+    )
+    # Set the verbosity to info of the Transformers logger (on main process only):
+    if is_main_process(training_args.local_rank):
+        transformers.utils.logging.set_verbosity_info()
+    logger.info("Training/evaluation parameters %s", training_args)
+
+    # Set seed before initializing model.
+    set_seed(training_args.seed)
+
+    # Pre-processing dataset
+    def preprocess_dataset(entry):
+        return (
+            "<INAUDIBLE>" not in entry["text"]
+            and entry["sentence_language_code"].lower() == "nb-no"
+        )
+    
+    def map_dataset(entry):
+        return {"text": (entry["text"]
+            .lower()
+            .replace("<ee>", "eee")
+            .replace("<mm>", "mmm")
+            .replace("<qq>", "qqq")
+            .replace("ó", "o")
+            .replace("é", "e")
+        )}
+
+    # 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,
+        )
+        raw_datasets["train"] = raw_datasets["train"].filter(preprocess_dataset)
+        raw_datasets["train"] = raw_datasets["train"].map(map_dataset)
+
+        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,
+        )
+        raw_datasets["eval"] = raw_datasets["eval"].filter(preprocess_dataset)
+        raw_datasets["eval"] = raw_datasets["eval"].map(map_dataset)
+
+        if data_args.max_eval_samples is not None:
+            raw_datasets["eval"] = raw_datasets["eval"].select(range(data_args.max_eval_samples))
+
+    # 2. We remove some special characters from the datasets
+    # that make training complicated and do not help in transcribing the speech
+    # E.g. characters, such as `,` and `.` do not really have an acoustic characteristic
+    # that could be easily picked up by the model
+    chars_to_ignore_regex = (
+        f'[{"".join(data_args.chars_to_ignore)}]' if data_args.chars_to_ignore is not None else None
+    )
+    text_column_name = data_args.text_column_name
+
+    def remove_special_characters(batch):
+        if chars_to_ignore_regex is not None:
+            batch["target_text"] = re.sub(chars_to_ignore_regex, "", batch[text_column_name]).lower() + " "
+        else:
+            batch["target_text"] = batch[text_column_name].lower() + " "
+        return batch
+
+    with training_args.main_process_first(desc="dataset map special characters removal"):
+        raw_datasets = raw_datasets.map(
+            remove_special_characters,
+            remove_columns=[text_column_name],
+            desc="remove special characters from datasets",
+        )
+
+    # save special tokens for tokenizer
+    word_delimiter_token = data_args.word_delimiter_token
+    unk_token = data_args.unk_token
+    pad_token = data_args.pad_token
+
+    # 3. Next, let's load the config as we might need it to create
+    # the tokenizer
+    # load config
+    config = AutoConfig.from_pretrained(
+        model_args.model_name_or_path, cache_dir=model_args.cache_dir, use_auth_token=data_args.use_auth_token
+    )
+
+    # 4. Next, if no tokenizer file is defined,
+    # we create the vocabulary of the model by extracting all unique characters from
+    # the training and evaluation datasets
+    # We need to make sure that only first rank saves vocabulary
+    # make sure all processes wait until vocab is created
+    tokenizer_name_or_path = model_args.tokenizer_name_or_path
+    tokenizer_kwargs = {}
+    if tokenizer_name_or_path is None:
+        # save vocab in training output dir
+        tokenizer_name_or_path = training_args.output_dir
+
+        vocab_file = os.path.join(tokenizer_name_or_path, "vocab.json")
+
+        with training_args.main_process_first():
+            if training_args.overwrite_output_dir and os.path.isfile(vocab_file):
+                os.remove(vocab_file)
+
+        with training_args.main_process_first(desc="dataset map vocabulary creation"):
+            if not os.path.isfile(vocab_file):
+                os.makedirs(tokenizer_name_or_path, exist_ok=True)
+                vocab_dict = create_vocabulary_from_data(
+                    raw_datasets,
+                    word_delimiter_token=word_delimiter_token,
+                    unk_token=unk_token,
+                    pad_token=pad_token,
+                )
+
+                # save vocab dict to be loaded into tokenizer
+                with open(vocab_file, "w") as file:
+                    json.dump(vocab_dict, file)
+
+        # if tokenizer has just been created
+        # it is defined by `tokenizer_class` if present in config else by `model_type`
+        tokenizer_kwargs = {
+            "config": config if config.tokenizer_class is not None else None,
+            "tokenizer_type": config.model_type if config.tokenizer_class is None else None,
+            "unk_token": unk_token,
+            "pad_token": pad_token,
+            "word_delimiter_token": word_delimiter_token,
+        }
+
+    # 5. Now we can instantiate the feature extractor, tokenizer and model
+    # Note for distributed training, the .from_pretrained methods guarantee that only
+    # one local process can concurrently download model & vocab.
+
+    # load feature_extractor and tokenizer
+    tokenizer = AutoTokenizer.from_pretrained(
+        tokenizer_name_or_path,
+        use_auth_token=data_args.use_auth_token,
+        **tokenizer_kwargs,
+    )
+    feature_extractor = AutoFeatureExtractor.from_pretrained(
+        model_args.model_name_or_path, cache_dir=model_args.cache_dir, use_auth_token=data_args.use_auth_token
+    )
+
+    # adapt config
+    config.update(
+        {
+            "feat_proj_dropout": model_args.feat_proj_dropout,
+            "attention_dropout": model_args.attention_dropout,
+            "hidden_dropout": model_args.hidden_dropout,
+            "final_dropout": model_args.final_dropout,
+            "mask_time_prob": model_args.mask_time_prob,
+            "mask_time_length": model_args.mask_time_length,
+            "mask_feature_prob": model_args.mask_feature_prob,
+            "mask_feature_length": model_args.mask_feature_length,
+            "gradient_checkpointing": training_args.gradient_checkpointing,
+            "layerdrop": model_args.layerdrop,
+            "ctc_loss_reduction": model_args.ctc_loss_reduction,
+            "pad_token_id": tokenizer.pad_token_id,
+            "vocab_size": len(tokenizer),
+            "activation_dropout": model_args.activation_dropout,
+        }
+    )
+
+    # create model
+    model = AutoModelForCTC.from_pretrained(
+        model_args.model_name_or_path,
+        cache_dir=model_args.cache_dir,
+        config=config,
+        use_auth_token=data_args.use_auth_token,
+    )
+
+    # freeze encoder
+    if model_args.freeze_feature_encoder:
+        model.freeze_feature_encoder()
+
+    # 6. Now we preprocess the datasets including loading the audio, resampling and normalization
+    # Thankfully, `datasets` takes care of automatically loading and resampling the audio,
+    # so that we just need to set the correct target sampling rate and normalize the input
+    # via the `feature_extractor`
+
+    # make sure that dataset decodes audio with correct sampling rate
+    dataset_sampling_rate = next(iter(raw_datasets.values())).features[data_args.audio_column_name].sampling_rate
+    if dataset_sampling_rate != feature_extractor.sampling_rate:
+        raw_datasets = raw_datasets.cast_column(
+            data_args.audio_column_name, datasets.features.Audio(sampling_rate=feature_extractor.sampling_rate)
+        )
+
+    # derive max & min input length for sample rate & max duration
+    max_input_length = data_args.max_duration_in_seconds * feature_extractor.sampling_rate
+    min_input_length = data_args.min_duration_in_seconds * feature_extractor.sampling_rate
+    audio_column_name = data_args.audio_column_name
+    num_workers = data_args.preprocessing_num_workers
+
+    # `phoneme_language` is only relevant if the model is fine-tuned on phoneme classification
+    phoneme_language = data_args.phoneme_language
+
+    # Preprocessing the datasets.
+    # We need to read the audio files as arrays and tokenize the targets.
+    def prepare_dataset(batch):
+        # load audio
+        sample = batch[audio_column_name]
+
+        inputs = feature_extractor(sample["array"], sampling_rate=sample["sampling_rate"])
+        batch["input_values"] = inputs.input_values[0]
+        batch["input_length"] = len(batch["input_values"])
+
+        # encode targets
+        additional_kwargs = {}
+        if phoneme_language is not None:
+            additional_kwargs["phonemizer_lang"] = phoneme_language
+
+        batch["labels"] = tokenizer(batch["target_text"], **additional_kwargs).input_ids
+        return batch
+
+    with training_args.main_process_first(desc="dataset map preprocessing"):
+        vectorized_datasets = raw_datasets.map(
+            prepare_dataset,
+            remove_columns=next(iter(raw_datasets.values())).column_names,
+            num_proc=num_workers,
+            desc="preprocess datasets",
+        )
+
+        def is_audio_in_length_range(length):
+            return length > min_input_length and length < max_input_length
+
+        # filter data that is shorter than min_input_length
+        vectorized_datasets = vectorized_datasets.filter(
+            is_audio_in_length_range,
+            num_proc=num_workers,
+            input_columns=["input_length"],
+        )
+
+    # 7. Next, we can prepare the training.
+    # Let's use word error rate (WER) as our evaluation metric,
+    # instantiate a data collator and the trainer
+
+    # Define evaluation metrics during training, *i.e.* word error rate, character error rate
+    eval_metrics = {metric: load_metric(metric) for metric in data_args.eval_metrics}
+
+    # for large datasets it is advised to run the preprocessing on a
+    # single machine first with ``args.preprocessing_only`` since there will mostly likely
+    # be a timeout when running the script in distributed mode.
+    # In a second step ``args.preprocessing_only`` can then be set to `False` to load the
+    # cached dataset
+    if data_args.preprocessing_only:
+        logger.info(f"Data preprocessing finished. Files cached at {vectorized_datasets.cache_files}")
+        return
+
+    def compute_metrics(pred):
+        pred_logits = pred.predictions
+        pred_ids = np.argmax(pred_logits, axis=-1)
+
+        pred.label_ids[pred.label_ids == -100] = tokenizer.pad_token_id
+
+        pred_str = tokenizer.batch_decode(pred_ids)
+        # we do not want to group tokens when computing the metrics
+        label_str = tokenizer.batch_decode(pred.label_ids, group_tokens=False)
+
+        metrics = {k: v.compute(predictions=pred_str, references=label_str) for k, v in eval_metrics.items()}
+
+        return metrics
+
+    # Now save everything to be able to create a single processor later
+    if is_main_process(training_args.local_rank):
+        # save feature extractor, tokenizer and config
+        feature_extractor.save_pretrained(training_args.output_dir)
+        tokenizer.save_pretrained(training_args.output_dir)
+        config.save_pretrained(training_args.output_dir)
+
+    try:
+        processor = AutoProcessor.from_pretrained(training_args.output_dir)
+    except (OSError, KeyError):
+        warnings.warn(
+            "Loading a processor from a feature extractor config that does not"
+            " include a `processor_class` attribute is deprecated and will be removed in v5. Please add the following "
+            " attribute to your `preprocessor_config.json` file to suppress this warning: "
+            " `'processor_class': 'Wav2Vec2Processor'`",
+            FutureWarning,
+        )
+        processor = Wav2Vec2Processor.from_pretrained(training_args.output_dir)
+
+    # Instantiate custom data collator
+    data_collator = DataCollatorCTCWithPadding(processor=processor)
+
+    # Initialize Trainer
+    trainer = Trainer(
+        model=model,
+        data_collator=data_collator,
+        args=training_args,
+        compute_metrics=compute_metrics,
+        train_dataset=vectorized_datasets["train"] if training_args.do_train else None,
+        eval_dataset=vectorized_datasets["eval"] if training_args.do_eval else None,
+        tokenizer=feature_extractor,
+    )
+
+    # 8. Finally, we can start training
+
+    # Training
+    if training_args.do_train:
+
+        # use last checkpoint if exist
+        if last_checkpoint is not None:
+            checkpoint = last_checkpoint
+        elif os.path.isdir(model_args.model_name_or_path):
+            checkpoint = model_args.model_name_or_path
+        else:
+            checkpoint = None
+
+        train_result = trainer.train(resume_from_checkpoint=checkpoint)
+        trainer.save_model()
+
+        metrics = train_result.metrics
+        max_train_samples = (
+            data_args.max_train_samples
+            if data_args.max_train_samples is not None
+            else len(vectorized_datasets["train"])
+        )
+        metrics["train_samples"] = min(max_train_samples, len(vectorized_datasets["train"]))
+
+        trainer.log_metrics("train", metrics)
+        trainer.save_metrics("train", metrics)
+        trainer.save_state()
+
+    # Evaluation
+    results = {}
+    if training_args.do_eval:
+        logger.info("*** Evaluate ***")
+        metrics = trainer.evaluate()
+        max_eval_samples = (
+            data_args.max_eval_samples if data_args.max_eval_samples is not None else len(vectorized_datasets["eval"])
+        )
+        metrics["eval_samples"] = min(max_eval_samples, len(vectorized_datasets["eval"]))
+
+        trainer.log_metrics("eval", metrics)
+        trainer.save_metrics("eval", metrics)
+
+    # Write model card and (optionally) push to hub
+    config_name = data_args.dataset_config_name if data_args.dataset_config_name is not None else "na"
+    kwargs = {
+        "finetuned_from": model_args.model_name_or_path,
+        "tasks": "speech-recognition",
+        "tags": ["automatic-speech-recognition", data_args.dataset_name],
+        "dataset_args": f"Config: {config_name}, Training split: {data_args.train_split_name}, Eval split: {data_args.eval_split_name}",
+        "dataset": f"{data_args.dataset_name.upper()} - {config_name.upper()}",
+    }
+    if "common_voice" in data_args.dataset_name:
+        kwargs["language"] = config_name
+
+    if training_args.push_to_hub:
+        trainer.push_to_hub(**kwargs)
+    else:
+        trainer.create_model_card(**kwargs)
+
+    return results
+
+
+if __name__ == "__main__":
+    main()

wandb/run-20220125_145153-1ghti100/files/diff.patch

@@ -0,0 +1,50 @@
+diff --git a/config.json b/config.json
+index e2b341d..c6e8001 100644
+--- a/config.json
++++ b/config.json
+@@ -6,7 +6,7 @@
+   "add_adapter": false,
+   "apply_spec_augment": true,
+   "architectures": [
+-    "Wav2Vec2ForCTC"
++    "Wav2Vec2ForPreTraining"
+   ],
+   "attention_dropout": 0.0,
+   "bos_token_id": 1,
+diff --git a/run.sh b/run.sh
+index f854ea3..24b2efe 100755
+--- a/run.sh
++++ b/run.sh
+@@ -4,6 +4,7 @@ python run_speech_recognition_ctc.py \
+ 	--dataset_config_name="48K_mp3" \
+ 	--output_dir="./" \
+ 	--overwrite_output_dir \
++        --hub_model_id="NbAiLab/xls-npsc" \
+ 	--num_train_epochs="10" \
+ 	--per_device_train_batch_size="8" \
+ 	--per_device_eval_batch_size="8" \
+@@ -31,4 +32,6 @@ python run_speech_recognition_ctc.py \
+ 	--fp16 \
+ 	--group_by_length \
+ 	--do_train --do_eval \
++        --max_train_samples="1000" \
++        --max_eval_samples="100"  \
+ 	--push_to_hub
+diff --git a/special_tokens_map.json b/special_tokens_map.json
+index 623bcb0..59f35a2 100644
+--- a/special_tokens_map.json
++++ b/special_tokens_map.json
+@@ -1 +1 @@
+-{"bos_token": "<s>", "eos_token": "</s>", "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}]}
+
++{"bos_token": "<s>", "eos_token": "</s>", "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}, {"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}]}
+
+diff --git a/vocab.json b/vocab.json
+index 6328bc8..5211c51 100644
+--- a/vocab.json
++++ b/vocab.json
+@@ -1 +1 @@
+-{"a": 1, "b": 2, "c": 3, "d": 4, "e": 5, "f": 6, "g": 7, "h": 8, "i": 9, "j": 10, "k": 11, "l": 12, "m": 13, "n": 14, "o": 15, "p": 16, "q": 17, "r": 18, "s": 19, "t": 20, "u": 21, "v": 22, "w": 23, "x": 24, "y": 25, "z": 26, "ä": 27, "å": 28, "é": 29, "ô": 30, "ö": 31, "ü": 32, "|": 0, "[UNK]": 33, "[PAD]": 34}
+
++{"a": 1, "b": 2, "c": 3, "d": 4, "e": 5, "f": 6, "g": 7, "h": 8, "i": 9, "j": 10, "k": 11, "l": 12, "m": 13, "n": 14, "o": 15, "p": 16, "q": 17, "r": 18, "s": 19, "t": 20, "u": 21, "v": 22, "w": 23, "x": 24, "y": 25, "z": 26, "å": 27, "æ": 28, "è": 29, "ò": 30, "ô": 31, "ø": 32, "|": 0, "[UNK]": 33, "[PAD]": 34}
+

wandb/run-20220125_145153-1ghti100/files/output.log

@@ -0,0 +1,185 @@
+
+  0%|                                                                                                                                                                                                                                                             | 0/150 [00:00<?, ?it/s]/home/javierr/venvs/notram/lib/python3.9/site-packages/torch/nn/parallel/_functions.py:68: UserWarning: Was asked to gather along dimension 0, but all input tensors were scalars; will instead unsqueeze and return a vector.
+  warnings.warn('Was asked to gather along dimension 0, but all '
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 66%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████                                                                                   | 99/150 [47:27<23:20, 27.46s/it]
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 99%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████▍ | 149/150 [1:11:24<00:27, 27.93s/it]
+100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 150/150 [1:11:49<00:00, 27.18s/it]
+Training completed. Do not forget to share your model on huggingface.co/models =)
+100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 150/150 [1:11:49<00:00, 28.73s/it]
+Saving model checkpoint to ./
+Configuration saved in ./config.json
+***** train metrics *****
+  epoch                    =       9.97
+  train_loss               =     5.8831
+  train_runtime            = 1:11:52.78
+  train_samples            =        990
+  train_samples_per_second =      2.295
+  train_steps_per_second   =      0.035
+01/25/2022 16:03:47 - INFO - __main__ - *** Evaluate ***
+Model weights saved in ./pytorch_model.bin
+Configuration saved in ./preprocessor_config.json
+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 = 99
+  Batch size = 16
+
+
+100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████��████████████████████████████████████████████████████████████████████████| 7/7 [00:05<00:00,  1.38it/s]
+***** eval metrics *****
+  epoch                   =       9.97
+  eval_loss               =     3.5006
+  eval_runtime            = 0:00:05.97
+  eval_samples            =         99
+  eval_samples_per_second =     16.561
+  eval_steps_per_second   =      1.171
+100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 7/7 [00:05<00:00,  1.28it/s]
+Saving model checkpoint to ./
+Configuration saved in ./config.json
+Model weights saved in ./pytorch_model.bin

wandb/run-20220125_145153-1ghti100/files/requirements.txt

@@ -0,0 +1,185 @@
+absl-py==0.15.0
+aiohttp==3.7.4.post0
+altair==4.1.0
+appdirs==1.4.4
+argon2-cffi==21.1.0
+astor==0.8.1
+astunparse==1.6.3
+async-timeout==3.0.1
+attrs==21.2.0
+audioread==2.1.9
+backcall==0.2.0
+base58==2.1.0
+bleach==4.1.0
+blinker==1.4
+cachetools==4.2.4
+certifi==2021.10.8
+cffi==1.15.0
+chardet==4.0.0
+charset-normalizer==2.0.7
+chex==0.0.8
+clang==5.0
+click==7.1.2
+colorama==0.4.4
+configparser==5.0.2
+conllu==4.4.1
+cycler==0.10.0
+datasets==1.18.1.dev0
+debugpy==1.5.1
+decorator==5.1.0
+defusedxml==0.7.1
+dill==0.3.4
+dm-tree==0.1.6
+docker-pycreds==0.4.0
+entrypoints==0.3
+filelock==3.3.0
+flatbuffers==1.12
+flax==0.3.5
+fsspec==2021.10.0
+gast==0.4.0
+gitdb==4.0.9
+gitpython==3.1.24
+google-auth-oauthlib==0.4.6
+google-auth==2.3.1
+google-pasta==0.2.0
+grpcio==1.41.1
+h5py==3.1.0
+huggingface-hub==0.4.0.dev0
+idna==3.3
+ipykernel==6.4.2
+ipython-genutils==0.2.0
+ipython==7.28.0
+ipywidgets==7.6.5
+jax==0.2.24
+jaxlib==0.1.73
+jedi==0.18.0
+jinja2==3.0.2
+jiwer==2.3.0
+joblib==1.1.0
+jsonschema==4.1.2
+jupyter-client==7.0.6
+jupyter-console==6.4.0
+jupyter-core==4.8.1
+jupyter==1.0.0
+jupyterlab-pygments==0.1.2
+jupyterlab-widgets==1.0.2
+keras-preprocessing==1.1.2
+keras==2.6.0
+kiwisolver==1.3.2
+librosa==0.8.1
+libtpu-nightly==0.1.dev20211018
+llvmlite==0.38.0
+markdown==3.3.4
+markupsafe==2.0.1
+matplotlib-inline==0.1.3
+matplotlib==3.4.3
+mistune==0.8.4
+msgpack==1.0.2
+multidict==5.2.0
+multiprocess==0.70.12.2
+nbclient==0.5.4
+nbconvert==6.2.0
+nbformat==5.1.3
+nest-asyncio==1.5.1
+nltk==3.6.5
+notebook==6.4.5
+numba==0.55.0
+numpy==1.19.5
+oauthlib==3.1.1
+opt-einsum==3.3.0
+optax==0.0.9
+packaging==21.0
+pandas==1.3.4
+pandocfilters==1.5.0
+parso==0.8.2
+pathtools==0.1.2
+pexpect==4.8.0
+pickleshare==0.7.5
+pillow==8.3.2
+pip==21.3.1
+pkg-resources==0.0.0
+pooch==1.6.0
+portalocker==2.3.2
+prometheus-client==0.11.0
+promise==2.3
+prompt-toolkit==3.0.20
+protobuf==3.19.0
+psutil==5.8.0
+ptyprocess==0.7.0
+pyarrow==5.0.0
+pyasn1-modules==0.2.8
+pyasn1==0.4.8
+pycparser==2.20
+pydeck==0.7.1
+pygments==2.10.0
+pymupdf==1.19.1
+pyparsing==2.4.7
+pyrsistent==0.18.0
+python-dateutil==2.8.2
+python-levenshtein==0.12.2
+python-slugify==5.0.2
+pytz-deprecation-shim==0.1.0.post0
+pytz==2021.3
+pyyaml==6.0
+pyzmq==22.3.0
+qtconsole==5.1.1
+qtpy==1.11.2
+regex==2021.10.8
+requests-oauthlib==1.3.0
+requests==2.26.0
+resampy==0.2.2
+rsa==4.7.2
+sacrebleu==2.0.0
+sacremoses==0.0.46
+scikit-learn==1.0.1
+scipy==1.7.1
+send2trash==1.8.0
+sentence-splitter==1.4
+sentencepiece==0.1.96
+sentry-sdk==1.4.3
+seqeval==1.2.2
+setuptools==44.1.1
+shortuuid==1.0.1
+simpletransformers==0.62.2
+six==1.15.0
+smmap==5.0.0
+soundfile==0.10.3.post1
+streamlit==1.1.0
+subprocess32==3.5.4
+tabulate==0.8.9
+tensorboard-data-server==0.6.1
+tensorboard-plugin-wit==1.8.0
+tensorboard==2.7.0
+tensorboardx==2.4
+tensorflow-estimator==2.6.0
+tensorflow==2.6.0
+termcolor==1.1.0
+terminado==0.12.1
+testpath==0.5.0
+text-unidecode==1.3
+threadpoolctl==3.0.0
+tokenizers==0.10.3
+toml==0.10.2
+toolz==0.11.1
+torch==1.10.1
+torchaudio==0.10.1
+tornado==6.1
+tqdm==4.62.3
+traitlets==5.1.0
+transformers==4.16.0.dev0
+typing-extensions==3.7.4.3
+tzdata==2021.5
+tzlocal==4.0.2
+urllib3==1.26.7
+validators==0.18.2
+wandb==0.12.7
+watchdog==2.1.6
+wcwidth==0.2.5
+webencodings==0.5.1
+werkzeug==2.0.2
+wheel==0.37.0
+widgetsnbextension==3.5.1
+wrapt==1.12.1
+xxhash==2.0.2
+yarl==1.7.0
+yaspin==2.1.0

wandb/run-20220125_145153-1ghti100/files/wandb-metadata.json

@@ -0,0 +1,79 @@
+{
+    "os": "Linux-5.13.0-25-generic-x86_64-with-glibc2.34",
+    "python": "3.9.7",
+    "heartbeatAt": "2022-01-25T13:51:55.755641",
+    "startedAt": "2022-01-25T13:51:53.776560",
+    "docker": null,
+    "gpu": "Quadro RTX 6000",
+    "gpu_count": 2,
+    "cpu_count": 96,
+    "cuda": null,
+    "args": [
+        "--dataset_name=NbAiLab/NPSC",
+        "--model_name_or_path=facebook/wav2vec2-xls-r-300m",
+        "--dataset_config_name=48K_mp3",
+        "--output_dir=./",
+        "--overwrite_output_dir",
+        "--hub_model_id=NbAiLab/xls-npsc",
+        "--num_train_epochs=10",
+        "--per_device_train_batch_size=8",
+        "--per_device_eval_batch_size=8",
+        "--gradient_accumulation_steps=4",
+        "--learning_rate=7.5e-5",
+        "--warmup_steps=500",
+        "--length_column_name=input_length",
+        "--evaluation_strategy=steps",
+        "--text_column_name=text",
+        "--chars_to_ignore",
+        ",",
+        "?",
+        ".",
+        "!",
+        "-",
+        ";",
+        ":",
+        "\"",
+        "\u201c",
+        "%",
+        "\u2018",
+        "\u201d",
+        "\ufffd",
+        "\u2014",
+        "\u2019",
+        "\u2026",
+        "\u2013",
+        "--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=0.25",
+        "--mask_feature_length=64",
+        "--gradient_checkpointing",
+        "--use_auth_token",
+        "--fp16",
+        "--group_by_length",
+        "--do_train",
+        "--do_eval",
+        "--max_train_samples=1000",
+        "--max_eval_samples=100",
+        "--push_to_hub"
+    ],
+    "state": "running",
+    "program": "/mnt/lv_ai_1_ficino/ml/xls-npsc/run_speech_recognition_ctc.py",
+    "codePath": "run_speech_recognition_ctc.py",
+    "git": {
+        "remote": "https://huggingface.co/NbAiLab/xls-npsc",
+        "commit": "64b71f2107460016dd1f487d6425c855f4f7dfe5"
+    },
+    "email": "versae@gmail.com",
+    "root": "/mnt/lv_ai_1_ficino/ml/xls-npsc",
+    "host": "ficino",
+    "username": "javierr",
+    "executable": "/home/javierr/venvs/notram/bin/python"
+}

wandb/run-20220125_145153-1ghti100/logs/debug.log

@@ -0,0 +1,25 @@
+2022-01-25 14:51:53,779 INFO    MainThread:3248974 [wandb_setup.py:_flush():71] setting env: {}
+2022-01-25 14:51:53,780 INFO    MainThread:3248974 [wandb_setup.py:_flush():71] setting login settings: {}
+2022-01-25 14:51:53,780 INFO    MainThread:3248974 [wandb_init.py:_log_setup():371] Logging user logs to /mnt/lv_ai_1_ficino/ml/xls-npsc/wandb/run-20220125_145153-1ghti100/logs/debug.log
+2022-01-25 14:51:53,780 INFO    MainThread:3248974 [wandb_init.py:_log_setup():372] Logging internal logs to /mnt/lv_ai_1_ficino/ml/xls-npsc/wandb/run-20220125_145153-1ghti100/logs/debug-internal.log
+2022-01-25 14:51:53,780 INFO    MainThread:3248974 [wandb_init.py:init():404] calling init triggers
+2022-01-25 14:51:53,780 INFO    MainThread:3248974 [wandb_init.py:init():409] wandb.init called with sweep_config: {}
+config: {}
+2022-01-25 14:51:53,780 INFO    MainThread:3248974 [wandb_init.py:init():449] starting backend
+2022-01-25 14:51:53,780 INFO    MainThread:3248974 [backend.py:_multiprocessing_setup():95] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
+2022-01-25 14:51:53,835 INFO    MainThread:3248974 [backend.py:ensure_launched():199] starting backend process...
+2022-01-25 14:51:53,865 INFO    MainThread:3248974 [backend.py:ensure_launched():204] started backend process with pid: 3249965
+2022-01-25 14:51:53,867 INFO    MainThread:3248974 [wandb_init.py:init():458] backend started and connected
+2022-01-25 14:51:53,869 INFO    MainThread:3248974 [wandb_init.py:init():519] updated telemetry
+2022-01-25 14:51:53,944 INFO    MainThread:3248974 [wandb_init.py:init():549] communicating current version
+2022-01-25 14:51:54,574 INFO    MainThread:3248974 [wandb_init.py:init():554] got version response upgrade_message: "wandb version 0.12.9 is available!  To upgrade, please run:\n $ pip install wandb --upgrade"
+
+2022-01-25 14:51:54,574 INFO    MainThread:3248974 [wandb_init.py:init():564] communicating run to backend with 30 second timeout
+2022-01-25 14:51:54,883 INFO    MainThread:3248974 [wandb_init.py:init():592] starting run threads in backend
+2022-01-25 14:51:55,945 INFO    MainThread:3248974 [wandb_run.py:_console_start():1816] atexit reg
+2022-01-25 14:51:55,946 INFO    MainThread:3248974 [wandb_run.py:_redirect():1690] redirect: SettingsConsole.REDIRECT
+2022-01-25 14:51:55,947 INFO    MainThread:3248974 [wandb_run.py:_redirect():1695] Redirecting console.
+2022-01-25 14:51:55,950 INFO    MainThread:3248974 [wandb_run.py:_redirect():1751] Redirects installed.
+2022-01-25 14:51:55,950 INFO    MainThread:3248974 [wandb_init.py:init():619] run started, returning control to user process
+2022-01-25 14:51:55,954 INFO    MainThread:3248974 [wandb_run.py:_config_callback():962] 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': 34, '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': '4.16.0.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': 37, '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': 0.25, 'mask_feature_length': 64, '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': 8, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': 'None', 'per_gpu_eval_batch_size': 'None', 'gradient_accumulation_steps': 4, 'eval_accumulation_steps': 'None', 'learning_rate': 7.5e-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': 10.0, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'warmup_ratio': 0.0, 'warmup_steps': 500, 'log_level': -1, 'log_level_replica': -1, 'log_on_each_node': True, 'logging_dir': './runs/Jan25_14-51-02_ficino', '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': './', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': 'None', 'load_best_model_at_end': False, 'metric_for_best_model': 'None', 'greater_is_better': 'None', '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': "['tensorboard', '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': 'NbAiLab/xls-npsc', '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': 2, 'mp_parameters': '', 'train_batch_size': 16, 'eval_batch_size': 16}
+2022-01-25 14:51:55,960 INFO    MainThread:3248974 [wandb_watch.py:watch():43] Watching