First of version Alberti 📜🖊️

README.md

@@ -0,0 +1,82 @@
+---
+language: es
+license: cc-by-4.0
+tags:
+- multilingual
+- bert
+pipeline_tag: fill-mask
+widget:
+- text: ¿Qué es la vida? Un [MASK].
+---
+
+# ALBERTI
+
+ALBERTI is a set of two BERT-based multilingual model for poetry. One for verses and another one for stanzas. This model has been further trained with the PULPO corpus for verses using [Flax](https://github.com/google/flax), including training scripts.
+
+This is part of the
+[Flax/Jax Community Week](https://discuss.huggingface.co/t/open-to-the-community-community-week-using-jax-flax-for-nlp-cv/7104), organised by [HuggingFace](https://huggingface.co/) and TPU usage sponsored by Google.
+
+## PULPO
+
+PULPO, the Prodigious Unannotated Literary Poetry Corpus, is a set of multilingual corpora of verses and stanzas with over 95M words.
+
+The following corpora has been downloaded using the [Averell](https://github.com/linhd-postdata/averell/) tool, developed by the [POSTDATA](https://postdata.linhd.uned.es/) team:
+
+### Spanish
+- [Disco v3](https://github.com/pruizf/disco)
+- [Corpus of Spanish Golden-Age Sonnets](https://github.com/bncolorado/CorpusSonetosSigloDeOro)
+- [Corpus general de poesía lírica castellana del Siglo de Oro](https://github.com/bncolorado/CorpusGeneralPoesiaLiricaCastellanaDelSigloDeOro)
+- [Gongocorpus](https://github.com/linhd-postdata/gongocorpus) - [source](http://obvil.sorbonne-universite.site/corpus/gongora/gongora_obra-poetica)
+### English
+- [Eighteenth-Century Poetry Archive (ECPA)](https://github.com/alhuber1502/ECPA)
+- [For better for verse](https://github.com/waynegraham/for_better_for_verse)
+### French
+- [Métrique en Ligne](https://crisco2.unicaen.fr/verlaine/index.php?navigation=accueil) - [source](https://github.com/linhd-postdata/metrique-en-ligne)
+### Italian
+- [Biblioteca italiana](https://github.com/linhd-postdata/biblioteca_italiana) - [source](http://www.bibliotecaitaliana.it/)
+### Czech
+- [Corpus of Czech Verse](https://github.com/versotym/corpusCzechVerse)
+### Portuguese
+- [Stichotheque](https://gitlab.com/stichotheque/stichotheque-pt)
+
+Also, we obtained the following corpora from these sources:
+### Spanish 
+- [Poesi.as](https://github.com/linhd-postdata/poesi.as) - [source](http://www.poesi.as/)
+### English
+- [A Gutenberg Poetry Corpus](https://github.com/aparrish/gutenberg-poetry-corpus)
+### Arabic
+- [Arabic Poetry dataset](https://www.kaggle.com/ahmedabelal/arabic-poetry)
+### Chinese
+- [THU Chinese Classical Poetry Corpus](https://github.com/THUNLP-AIPoet/Datasets/tree/master/CCPC)
+### Finnish
+- [SKVR](https://github.com/sks190/SKVR)
+### German
+- [TextGrid Poetry Corpus](https://github.com/linhd-postdata/textgrid-poetry) - [source](https://textgrid.de/en/digitale-bibliothek)
+- [German Rhyme Corpus](https://github.com/tnhaider/german-rhyme-corpus)
+### Hungarian
+- [verskorpusz](https://github.com/ELTE-DH/verskorpusz)
+### Portuguese
+- [Poems in Portuguese](https://www.kaggle.com/oliveirasp6/poems-in-portuguese)
+### Russian
+- [19 000 Russian poems](https://www.kaggle.com/grafstor/19-000-russian-poems)
+
+## Team members
+
+- Álvaro Pérez ([alvp](https://huggingface.co/alvp))
+- Javier de la Rosa ([versae](https://huggingface.co/versae))
+- Aitor Díaz ([aitordiaz](https://huggingface.co/aitordiaz))
+- Elena González-Blanco
+- Salvador Ros ([salva](https://huggingface.co/salva))
+
+## Useful links
+
+- [Community Week timeline](https://discuss.huggingface.co/t/open-to-the-community-community-week-using-jax-flax-for-nlp-cv/7104#summary-timeline-calendar-6)
+- [Community Week README](https://github.com/huggingface/transformers/blob/master/examples/research_projects/jax-projects/README.md)
+- [Community Week thread](https://discuss.huggingface.co/t/bertin-pretrain-roberta-large-from-scratch-in-spanish/7125)
+- [Community Week channel](https://discord.com/channels/858019234139602994/859113060068229190)
+- [Masked Language Modelling example scripts](https://github.com/huggingface/transformers/tree/master/examples/flax/language-modeling)
+- [Model Repository](https://huggingface.co/flax-community/alberti-bert-base-multilingual-cased/)
+
+## Acknowledgments
+
+This project would not have been possible without the infrastructure and resources provided by HuggingFace and Google Cloud. Moreover, we want to thank POSTDATA Project (ERC-StG-679528) and the Computational Literary Studies Infrastructure (CLS INFRA No. 101004984) of the European Union's Horizon 2020 research and innovation programme for their support and time allowance.

config.json

@@ -0,0 +1,29 @@
+{
+  "architectures": [
+    "BertForMaskedLM"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "directionality": "bidi",
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "pooler_fc_size": 768,
+  "pooler_num_attention_heads": 12,
+  "pooler_num_fc_layers": 3,
+  "pooler_size_per_head": 128,
+  "pooler_type": "first_token_transform",
+  "position_embedding_type": "absolute",
+  "transformers_version": "4.9.0.dev0",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 119547
+}

config.py

@@ -0,0 +1,7 @@
+#!/usr/bin/env python
+
+from transformers import BertConfig
+
+config = BertConfig.from_pretrained("bert-base-multilingual-cased")
+
+config.save_pretrained("./")

convert.py

@@ -0,0 +1,25 @@
+
+#!/usr/bin/env python
+import tempfile
+import jax
+from jax import numpy as jnp
+from transformers import AutoTokenizer, FlaxRobertaForMaskedLM, RobertaForMaskedLM, FlaxBertForMaskedLM, BertForMaskedLM
+
+
+def main():
+    # Saving extra files from config.json and tokenizer.json files
+    tokenizer = AutoTokenizer.from_pretrained("bert-base-multilingual-cased")
+    tokenizer.save_pretrained("./")
+    # Temporary saving bfloat16 Flax model into float32
+    tmp = tempfile.mkdtemp()
+    #flax_model = FlaxRobertaForMaskedLM.from_pretrained("./")
+    flax_model = FlaxBertForMaskedLM.from_pretrained("./")
+    flax_model.save_pretrained(tmp)
+    # Converting float32 Flax to PyTorch
+    #model = RobertaForMaskedLM.from_pretrained(tmp, from_flax=True)
+    model = BertForMaskedLM.from_pretrained(tmp, from_flax=True)
+    model.save_pretrained("./", save_config=False)
+
+
+if __name__ == "__main__":
+    main()

evals/alberti/eval_results_arabic.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.6408291005392143,
+    "eval_loss": 1.505015057336042,
+    "eval_perplexity": 4.504221451353881
+}

evals/alberti/eval_results_chinese.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.31155171271450344,
+    "eval_loss": 3.9333015512157954,
+    "eval_perplexity": 51.07532741795546
+}

evals/alberti/eval_results_czech.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.6062408166567269,
+    "eval_loss": 1.8359284606976425,
+    "eval_perplexity": 6.270953777486964
+}

evals/alberti/eval_results_english.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.5320815459645918,
+    "eval_loss": 2.394027611476076,
+    "eval_perplexity": 10.957537892911233
+}

evals/alberti/eval_results_finnish.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.6272082468928667,
+    "eval_loss": 1.845318445734281,
+    "eval_perplexity": 6.330115267679467
+}

evals/alberti/eval_results_french.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.5842149738042928,
+    "eval_loss": 2.0680841484282575,
+    "eval_perplexity": 7.909654867720871
+}

evals/alberti/eval_results_german.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.5835830474456813,
+    "eval_loss": 1.9559251674484819,
+    "eval_perplexity": 7.070457355700252
+}

evals/alberti/eval_results_hindi.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.07489781581300294,
+    "eval_loss": 9.280142339379232,
+    "eval_perplexity": 10722.958107026718
+}

evals/alberti/eval_results_hungarian.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.5566523349747832,
+    "eval_loss": 2.1803966072570926,
+    "eval_perplexity": 8.849815463823662
+}

evals/alberti/eval_results_italian.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.5162272232348571,
+    "eval_loss": 2.477469312651519,
+    "eval_perplexity": 11.91108301701437
+}

evals/alberti/eval_results_portuguese.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.4926062056586033,
+    "eval_loss": 2.7949925894154566,
+    "eval_perplexity": 16.362507496962778
+}

evals/alberti/eval_results_russian.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.5853658536585366,
+    "eval_loss": 2.086235640967887,
+    "eval_perplexity": 8.054537852308378
+}

evals/alberti/eval_results_spanish.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.5859480165147395,
+    "eval_loss": 2.084652532470857,
+    "eval_perplexity": 8.04179673294292
+}

evals/eval_results_alberti.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.5680093428409695,
+    "eval_loss": 2.2008613169662152,
+    "eval_perplexity": 9.032790245322317
+}

evals/eval_results_mbert.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.2531256913599916,
+    "eval_loss": 4.986973319130462,
+    "eval_perplexity": 146.49236618254264
+}

evals/mbert/eval_results_arabic.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.26391261358166496,
+    "eval_loss": 3.8152461392225794,
+    "eval_perplexity": 45.38792675716374
+}

evals/mbert/eval_results_chinese.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.09895315709269198,
+    "eval_loss": 6.6723325796726956,
+    "eval_perplexity": 790.2367464940897
+}

evals/mbert/eval_results_czech.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.2773844709547667,
+    "eval_loss": 4.42254573093298,
+    "eval_perplexity": 83.30809563399087
+}

evals/mbert/eval_results_english.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.2975008599745663,
+    "eval_loss": 4.490137260330782,
+    "eval_perplexity": 89.13367955737988
+}

evals/mbert/eval_results_finnish.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.22131404624517728,
+    "eval_loss": 5.228847592703858,
+    "eval_perplexity": 186.57766611729022
+}

evals/mbert/eval_results_french.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.36305560250126756,
+    "eval_loss": 3.776505989099206,
+    "eval_perplexity": 43.66321515247085
+}

evals/mbert/eval_results_german.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.27947910102919554,
+    "eval_loss": 4.477016383112791,
+    "eval_perplexity": 87.97180655711638
+}

evals/mbert/eval_results_hindi.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.3055626516796526,
+    "eval_loss": 3.95077246854643,
+    "eval_perplexity": 51.97550077126472
+}

evals/mbert/eval_results_hungarian.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.29008286594165167,
+    "eval_loss": 4.358099513515926,
+    "eval_perplexity": 78.10854903381517
+}

evals/mbert/eval_results_italian.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.2593468204267504,
+    "eval_loss": 4.857048540301865,
+    "eval_perplexity": 128.6439537945917
+}

evals/mbert/eval_results_portuguese.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.2960448389972281,
+    "eval_loss": 4.607839268924084,
+    "eval_perplexity": 100.2672648109061
+}

evals/mbert/eval_results_russian.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.28699294248879165,
+    "eval_loss": 4.441657214642727,
+    "eval_perplexity": 84.91554842410918
+}

evals/mbert/eval_results_spanish.json

@@ -0,0 +1,5 @@
+{
+    "eval_accuracy": 0.3220253743973355,
+    "eval_loss": 4.129566473031045,
+    "eval_perplexity": 62.15097297048524
+}

flax_model.msgpack

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:53d0ce75962d5d739f9bc2f5d6cdc40155753ddf8be1524112c95477237751cf
+size 711905363

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a5e2e38ef7342bc73dd33611f57c5123c491f12560fb32cec11dffe442572af
+size 711926000

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:74e76b0ba8bee646b78f7efdbdf3cd500ca183bc26439944fc1733f48a77f5ba
+size 711980011

run.sh

@@ -0,0 +1,27 @@
+# From https://arxiv.org/pdf/1907.11692.pdf
+./run_mlm_flax.py \
+    --model_name_or_path="bert-base-multilingual-cased" \
+    --output_dir="./" \
+    --model_type="bert" \
+    --config_name="./" \
+    --validation_file="./datasets/pulpo_lines_clean_val.json" \
+    --train_file="./datasets/pulpo_lines_clean_train.json" \
+    --tokenizer_name="bert-base-multilingual-cased" \
+    --max_seq_length="32" \
+    --pad_to_max_length  \
+    --per_device_train_batch_size="256" \
+    --per_device_eval_batch_size="256" \
+    --adam_beta1="0.9" \
+    --adam_beta2="0.98" \
+    --adam_epsilon="1e-6" \
+    --learning_rate="1.25e-4" \
+    --weight_decay="0.01" \
+    --save_strategy="steps" \
+    --save_steps="1000" \
+    --save_total_limit="5" \
+    --warmup_steps="10000" \
+    --num_train_epochs="40" \
+    --overwrite_output_dir \
+    --eval_steps="1000" \
+    --save_total_limit="1000" \
+    --logging_steps="500" 2>&1 | tee run.log

run_eval.sh

@@ -0,0 +1,27 @@
+# From https://arxiv.org/pdf/1907.11692.pdf
+./run_mlm_flax_eval.py \
+    --model_name_or_path="flax-community/alberti-bert-base-multilingual-cased" \
+    --output_dir="./" \
+    --model_type="bert" \
+    --config_name="./" \
+    --validation_file="./pulpo_lines_val.json" \
+    --train_file="./pulpo_lines_train.json" \
+    --tokenizer_name="bert-base-multilingual-cased" \
+    --max_seq_length="32" \
+    --pad_to_max_length  \
+    --per_device_train_batch_size="256" \
+    --per_device_eval_batch_size="256" \
+    --adam_beta1="0.9" \
+    --adam_beta2="0.98" \
+    --adam_epsilon="1e-6" \
+    --learning_rate="1.25e-4" \
+    --weight_decay="0.01" \
+    --save_steps="1000" \
+    --warmup_steps="10000" \
+    --num_train_epochs="40" \
+    --overwrite_output_dir \
+    --eval_steps="1000" \
+    --do_train="false" \
+    --do_eval="true" \
+    --preprocessing_num_workers="96" \
+    --logging_steps="500" 2>&1 | tee run.log

run_evals.sh

@@ -0,0 +1,31 @@
+for lang in "arabic" "chinese" "czech" "english" "finnish" "french" "german" "hindi" "hungarian" "italian" "portuguese" "russian" "spanish" "turkish"
+do
+  echo "";
+  echo "$lang";
+  ./run_mlm_flax_eval.py \
+    --model_name_or_path="bert-base-multilingual-cased" \
+    --output_dir="./" \
+    --model_type="bert" \
+    --config_name="./" \
+    --validation_file="./pulpo_lines_val.json" \
+    --train_file="./pulpo_lines_train.json" \
+    --tokenizer_name="bert-base-multilingual-cased" \
+    --max_seq_length="32" \
+    --pad_to_max_length  \
+    --per_device_train_batch_size="256" \
+    --per_device_eval_batch_size="256" \
+    --adam_beta1="0.9" \
+    --adam_beta2="0.98" \
+    --adam_epsilon="1e-6" \
+    --learning_rate="1.25e-4" \
+    --weight_decay="0.01" \
+    --save_steps="1000" \
+    --warmup_steps="10000" \
+    --num_train_epochs="40" \
+    --overwrite_output_dir \
+    --eval_steps="1000" \
+    --do_train="false" \
+    --do_eval="true" \
+    --lang="$lang" \
+    --logging_steps="500"
+done

run_mlm_flax.py

@@ -0,0 +1,745 @@
+#!/usr/bin/env python
+# coding=utf-8
+# Copyright 2021 The HuggingFace 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
+# limitations under the License.
+"""
+Fine-tuning the library models for masked language modeling (BERT, ALBERT, RoBERTa...) with whole word masking on a
+text file or a dataset.
+
+Here is the full list of checkpoints on the hub that can be fine-tuned by this script:
+https://huggingface.co/models?filter=masked-lm
+"""
+import json
+import logging
+import os
+import shutil
+import sys
+import time
+from dataclasses import dataclass, field
+
+# You can also adapt this script on your own masked language modeling task. Pointers for this are left as comments.
+from pathlib import Path
+from typing import Dict, List, Optional, Tuple
+
+import joblib
+import numpy as np
+from datasets import load_dataset
+from tqdm import tqdm
+
+import flax
+import jax
+import jax.numpy as jnp
+import optax
+from flax import jax_utils, traverse_util
+from flax.serialization import from_bytes, to_bytes
+from flax.training import train_state
+from flax.training.common_utils import get_metrics, onehot, shard
+from transformers import (
+    CONFIG_MAPPING,
+    FLAX_MODEL_FOR_MASKED_LM_MAPPING,
+    AutoConfig,
+    AutoTokenizer,
+    FlaxAutoModelForMaskedLM,
+    HfArgumentParser,
+    PreTrainedTokenizerBase,
+    TensorType,
+    TrainingArguments,
+    is_tensorboard_available,
+    set_seed,
+)
+
+print("TPU count:", jax.device_count())
+
+MODEL_CONFIG_CLASSES = list(FLAX_MODEL_FOR_MASKED_LM_MAPPING.keys())
+MODEL_TYPES = tuple(conf.model_type for conf in MODEL_CONFIG_CLASSES)
+
+
+
+@dataclass
+class ModelArguments:
+    """
+    Arguments pertaining to which model/config/tokenizer we are going to fine-tune, or train from scratch.
+    """
+
+    model_name_or_path: Optional[str] = field(
+        default=None,
+        metadata={
+            "help": "The model checkpoint for weights initialization."
+            "Don't set if you want to train a model from scratch."
+        },
+    )
+    model_type: Optional[str] = field(
+        default=None,
+        metadata={"help": "If training from scratch, pass a model type from the list: " + ", ".join(MODEL_TYPES)},
+    )
+    config_name: Optional[str] = field(
+        default=None, metadata={"help": "Pretrained config name or path if not the same as model_name"}
+    )
+    tokenizer_name: Optional[str] = field(
+        default=None, metadata={"help": "Pretrained tokenizer name or path if not the same as model_name"}
+    )
+    cache_dir: Optional[str] = field(
+        default=None, metadata={"help": "Where do you want to store the pretrained models downloaded from s3"}
+    )
+    use_fast_tokenizer: bool = field(
+        default=True,
+        metadata={"help": "Whether to use one of the fast tokenizer (backed by the tokenizers library) or not."},
+    )
+    dtype: Optional[str] = field(
+        default="float32",
+        metadata={
+            "help": "Floating-point format in which the model weights should be initialized and trained. Choose one of `[float32, float16, bfloat16]`."
+        },
+    )
+
+
+@dataclass
+class DataTrainingArguments:
+    """
+    Arguments pertaining to what data we are going to input our model for training and eval.
+    """
+
+    dataset_name: Optional[str] = field(
+        default=None, metadata={"help": "The name of the dataset to use (via the datasets library)."}
+    )
+    dataset_config_name: Optional[str] = field(
+        default=None, metadata={"help": "The configuration name of the dataset to use (via the datasets library)."}
+    )
+    train_file: Optional[str] = field(default=None, metadata={"help": "The input training data file (a text file)."})
+    validation_file: Optional[str] = field(
+        default=None,
+        metadata={"help": "An optional input evaluation data file to evaluate the perplexity on (a text file)."},
+    )
+    train_ref_file: Optional[str] = field(
+        default=None,
+        metadata={"help": "An optional input train ref data file for whole word masking in Chinese."},
+    )
+    validation_ref_file: Optional[str] = field(
+        default=None,
+        metadata={"help": "An optional input validation ref data file for whole word masking in Chinese."},
+    )
+    overwrite_cache: bool = field(
+        default=False, metadata={"help": "Overwrite the cached training and evaluation sets"}
+    )
+    validation_split_percentage: Optional[int] = field(
+        default=5,
+        metadata={
+            "help": "The percentage of the train set used as validation set in case there's no validation split"
+        },
+    )
+    max_seq_length: Optional[int] = field(
+        default=None,
+        metadata={
+            "help": "The maximum total input sequence length after tokenization. Sequences longer "
+            "than this will be truncated. Default to the max input length of the model."
+        },
+    )
+    preprocessing_num_workers: Optional[int] = field(
+        default=None,
+        metadata={"help": "The number of processes to use for the preprocessing."},
+    )
+    mlm_probability: float = field(
+        default=0.15, metadata={"help": "Ratio of tokens to mask for masked language modeling loss"}
+    )
+    pad_to_max_length: bool = field(
+        default=False,
+        metadata={
+            "help": "Whether to pad all samples to `max_seq_length`. "
+            "If False, will pad the samples dynamically when batching to the maximum length in the batch."
+        },
+    )
+    line_by_line: bool = field(
+        default=False,
+        metadata={"help": "Whether distinct lines of text in the dataset are to be handled as distinct sequences."},
+    )
+
+    def __post_init__(self):
+        if self.dataset_name is None and self.train_file is None and self.validation_file is None:
+            raise ValueError("Need either a dataset name or a training/validation file.")
+        else:
+            if self.train_file is not None:
+                extension = self.train_file.split(".")[-1]
+                assert extension in ["csv", "json", "txt"], "`train_file` should be a csv, a json or a txt file."
+            if self.validation_file is not None:
+                extension = self.validation_file.split(".")[-1]
+                assert extension in ["csv", "json", "txt"], "`validation_file` should be a csv, a json or a txt file."
+
+
+@flax.struct.dataclass
+class FlaxDataCollatorForLanguageModeling:
+    """
+    Data collator used for language modeling. Inputs are dynamically padded to the maximum length of a batch if they
+    are not all of the same length.
+
+    Args:
+        tokenizer (:class:`~transformers.PreTrainedTokenizer` or :class:`~transformers.PreTrainedTokenizerFast`):
+            The tokenizer used for encoding the data.
+        mlm_probability (:obj:`float`, `optional`, defaults to 0.15):
+            The probability with which to (randomly) mask tokens in the input.
+
+    .. note::
+
+        For best performance, this data collator should be used with a dataset having items that are dictionaries or
+        BatchEncoding, with the :obj:`"special_tokens_mask"` key, as returned by a
+        :class:`~transformers.PreTrainedTokenizer` or a :class:`~transformers.PreTrainedTokenizerFast` with the
+        argument :obj:`return_special_tokens_mask=True`.
+    """
+
+    tokenizer: PreTrainedTokenizerBase
+    mlm_probability: float = 0.15
+
+    def __post_init__(self):
+        if self.tokenizer.mask_token is None:
+            raise ValueError(
+                "This tokenizer does not have a mask token which is necessary for masked language modeling. "
+                "You should pass `mlm=False` to train on causal language modeling instead."
+            )
+
+    def __call__(self, examples: List[Dict[str, np.ndarray]], pad_to_multiple_of: int) -> Dict[str, np.ndarray]:
+        # Handle dict or lists with proper padding and conversion to tensor.
+        batch = self.tokenizer.pad(examples, pad_to_multiple_of=pad_to_multiple_of, return_tensors=TensorType.NUMPY)
+
+        # If special token mask has been preprocessed, pop it from the dict.
+        special_tokens_mask = batch.pop("special_tokens_mask", None)
+
+        batch["input_ids"], batch["labels"] = self.mask_tokens(
+            batch["input_ids"], special_tokens_mask=special_tokens_mask
+        )
+        return batch
+
+    def mask_tokens(
+        self, inputs: np.ndarray, special_tokens_mask: Optional[np.ndarray]
+    ) -> Tuple[jnp.ndarray, jnp.ndarray]:
+        """
+        Prepare masked tokens inputs/labels for masked language modeling: 80% MASK, 10% random, 10% original.
+        """
+        labels = inputs.copy()
+        # We sample a few tokens in each sequence for MLM training (with probability `self.mlm_probability`)
+        probability_matrix = np.full(labels.shape, self.mlm_probability)
+        special_tokens_mask = special_tokens_mask.astype("bool")
+
+        probability_matrix[special_tokens_mask] = 0.0
+        masked_indices = np.random.binomial(1, probability_matrix).astype("bool")
+        labels[~masked_indices] = -100  # We only compute loss on masked tokens
+
+        # 80% of the time, we replace masked input tokens with tokenizer.mask_token ([MASK])
+        indices_replaced = np.random.binomial(1, np.full(labels.shape, 0.8)).astype("bool") & masked_indices
+        inputs[indices_replaced] = self.tokenizer.convert_tokens_to_ids(self.tokenizer.mask_token)
+
+        # 10% of the time, we replace masked input tokens with random word
+        indices_random = np.random.binomial(1, np.full(labels.shape, 0.5)).astype("bool")
+        indices_random &= masked_indices & ~indices_replaced
+
+        random_words = np.random.randint(self.tokenizer.vocab_size, size=labels.shape, dtype="i4")
+        inputs[indices_random] = random_words[indices_random]
+
+        # The rest of the time (10% of the time) we keep the masked input tokens unchanged
+        return inputs, labels
+
+
+def generate_batch_splits(samples_idx: jnp.ndarray, batch_size: int) -> jnp.ndarray:
+    num_samples = len(samples_idx)
+    samples_to_remove = num_samples % batch_size
+
+    if samples_to_remove != 0:
+        samples_idx = samples_idx[:-samples_to_remove]
+    sections_split = num_samples // batch_size
+    batch_idx = np.split(samples_idx, sections_split)
+    return batch_idx
+
+
+def write_train_metric(summary_writer, train_metrics, train_time, step):
+    summary_writer.scalar("train_time", train_time, step)
+
+    train_metrics = get_metrics(train_metrics)
+    for key, vals in train_metrics.items():
+        tag = f"train_{key}"
+        for i, val in enumerate(vals):
+            summary_writer.scalar(tag, val, step - len(vals) + i + 1)
+
+
+def write_eval_metric(summary_writer, eval_metrics, step):
+    for metric_name, value in eval_metrics.items():
+        summary_writer.scalar(f"eval_{metric_name}", value, step)
+
+
+def rotate_checkpoints(path, max_checkpoints=5):
+    paths = sorted(Path(path).iterdir(), key=os.path.getmtime)[::-1]
+    if len(paths) > max_checkpoints:
+        for path_to_delete in paths[max_checkpoints:]:
+            try:
+                shutil.rmtree(path_to_delete)
+            except OSError:
+                os.remove(path_to_delete)
+
+
+def save_checkpoint_files(state, data_collator, training_args, save_dir):
+    unreplicated_state = jax_utils.unreplicate(state)
+    with open(os.path.join(save_dir, "optimizer_state.msgpack"), "wb") as f:
+        f.write(to_bytes(unreplicated_state.opt_state))
+    joblib.dump(training_args, os.path.join(save_dir, "training_args.joblib"))
+    joblib.dump(data_collator, os.path.join(save_dir, "data_collator.joblib"))
+    with open(os.path.join(save_dir, "training_state.json"), "w") as f:
+        json.dump({"step": unreplicated_state.step.item()}, f)
+
+
+
+if __name__ == "__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()
+
+    if (
+        os.path.exists(training_args.output_dir)
+        and os.listdir(training_args.output_dir)
+        and training_args.do_train
+        and not training_args.overwrite_output_dir
+    ):
+        raise ValueError(
+            f"Output directory ({training_args.output_dir}) already exists and is not empty."
+            "Use --overwrite_output_dir to overcome."
+        )
+
+    # Setup logging
+    logging.basicConfig(
+        format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
+        level="NOTSET",
+        datefmt="[%X]",
+    )
+
+    # Log on each process the small summary:
+    logger = logging.getLogger(__name__)
+    #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):
+    logger.info(f"Training/evaluation parameters {training_args}")
+
+    # Set seed before initializing model.
+    set_seed(training_args.seed)
+
+    # Get the datasets: you can either provide your own CSV/JSON/TXT training and evaluation files (see below)
+    # or just provide the name of one of the public datasets available on the hub at https://huggingface.co/datasets/
+    # (the dataset will be downloaded automatically from the datasets Hub).
+    #
+    # For CSV/JSON files, this script will use the column called 'text' or the first column if no column called
+    # 'text' is found. You can easily tweak this behavior (see below).
+    #
+    # In distributed training, the load_dataset function guarantees that only one local process can concurrently
+    # download the dataset.
+    
+    #datasets = load_dataset('json', name='averell', data_files={"train": "./datasets/lines_train.json", "validation": "./datasets/lines_val.json"})
+
+    if data_args.dataset_name is not None:
+        # Downloading and loading a dataset from the hub.
+        datasets = load_dataset(data_args.dataset_name, data_args.dataset_config_name, cache_dir=model_args.cache_dir)
+
+        if "validation" not in datasets.keys():
+            datasets["validation"] = load_dataset(
+                data_args.dataset_name,
+                data_args.dataset_config_name,
+                split=f"train[:{data_args.validation_split_percentage}%]",
+                cache_dir=model_args.cache_dir,
+            )
+            datasets["train"] = load_dataset(
+                data_args.dataset_name,
+                data_args.dataset_config_name,
+                split=f"train[{data_args.validation_split_percentage}%:]",
+                cache_dir=model_args.cache_dir,
+            )
+    else:
+        #### Custom dataset
+        data_files = {}
+        if data_args.train_file is not None:
+            data_files["train"] = data_args.train_file
+        if data_args.validation_file is not None:
+            data_files["validation"] = data_args.validation_file
+        extension = data_args.train_file.split(".")[-1]
+        if extension == "txt":
+            extension = "text"
+        datasets = load_dataset(extension, data_files=data_files, cache_dir=model_args.cache_dir)
+        #datasets = load_dataset('json', name='averell', data_files={"train": "./datasets/lines_train.json", "validation": "./datasets/lines_val.json"})
+
+    # See more about loading any type of standard or custom dataset (from files, python dict, pandas DataFrame, etc) at
+    # https://huggingface.co/docs/datasets/loading_datasets.html.
+
+    # Load pretrained model and tokenizer
+
+    # Distributed training:
+    # The .from_pretrained methods guarantee that only one local process can concurrently
+    # download model & vocab.
+    if model_args.config_name:
+        config = AutoConfig.from_pretrained(model_args.config_name, cache_dir=model_args.cache_dir)
+    elif model_args.model_name_or_path:
+        config = AutoConfig.from_pretrained(model_args.model_name_or_path, cache_dir=model_args.cache_dir)
+    else:
+        config = CONFIG_MAPPING[model_args.model_type]()
+        logger.warning("You are instantiating a new config instance from scratch.")
+
+    if model_args.tokenizer_name:
+        tokenizer = AutoTokenizer.from_pretrained(
+            model_args.tokenizer_name, cache_dir=model_args.cache_dir, use_fast=model_args.use_fast_tokenizer
+        )
+    elif model_args.model_name_or_path:
+        tokenizer = AutoTokenizer.from_pretrained(
+            model_args.model_name_or_path, cache_dir=model_args.cache_dir, use_fast=model_args.use_fast_tokenizer
+        )
+    else:
+        raise ValueError(
+            "You are instantiating a new tokenizer from scratch. This is not supported by this script."
+            "You can do it from another script, save it, and load it from here, using --tokenizer_name."
+        )
+
+    # Preprocessing the datasets.
+    # First we tokenize all the texts.
+    if training_args.do_train:
+        column_names = datasets["train"].column_names
+    else:
+        column_names = datasets["validation"].column_names
+    text_column_name = "text" if "text" in column_names else column_names[0]
+
+    max_seq_length = min(data_args.max_seq_length, tokenizer.model_max_length)
+
+    if data_args.line_by_line:
+        # When using line_by_line, we just tokenize each nonempty line.
+        padding = "max_length" if data_args.pad_to_max_length else False
+
+        def tokenize_function(examples):
+            # Remove empty lines
+            examples = [line for line in examples if len(line) > 0 and not line.isspace()]
+            return tokenizer(
+                examples,
+                return_special_tokens_mask=True,
+                padding=padding,
+                truncation=True,
+                max_length=max_seq_length,
+            )
+
+        tokenized_datasets = datasets.map(
+            tokenize_function,
+            input_columns=[text_column_name],
+            batched=True,
+            num_proc=data_args.preprocessing_num_workers,
+            remove_columns=column_names,
+            load_from_cache_file=not data_args.overwrite_cache,
+        )
+
+    else:
+        # Otherwise, we tokenize every text, then concatenate them together before splitting them in smaller parts.
+        # We use `return_special_tokens_mask=True` because DataCollatorForLanguageModeling (see below) is more
+        # efficient when it receives the `special_tokens_mask`.
+        def tokenize_function(examples):
+            return tokenizer(examples[text_column_name], return_special_tokens_mask=True)
+
+        tokenized_datasets = datasets.map(
+            tokenize_function,
+            batched=True,
+            num_proc=data_args.preprocessing_num_workers,
+            remove_columns=column_names,
+            load_from_cache_file=not data_args.overwrite_cache,
+        )
+
+        # Main data processing function that will concatenate all texts from our dataset and generate chunks of
+        # max_seq_length.
+        def group_texts(examples):
+            # Concatenate all texts.
+            concatenated_examples = {k: sum(examples[k], []) for k in examples.keys()}
+            total_length = len(concatenated_examples[list(examples.keys())[0]])
+            # We drop the small remainder, we could add padding if the model supported it instead of this drop, you can
+            # customize this part to your needs.
+            total_length = (total_length // max_seq_length) * max_seq_length
+            # Split by chunks of max_len.
+            result = {
+                k: [t[i : i + max_seq_length] for i in range(0, total_length, max_seq_length)]
+                for k, t in concatenated_examples.items()
+            }
+            return result
+
+        # Note that with `batched=True`, this map processes 1,000 texts together, so group_texts throws away a
+        # remainder for each of those groups of 1,000 texts. You can adjust that batch_size here but a higher value
+        # might be slower to preprocess.
+        #
+        # To speed up this part, we use multiprocessing. See the documentation of the map method for more information:
+        # https://huggingface.co/docs/datasets/package_reference/main_classes.html#datasets.Dataset.map
+        tokenized_datasets = tokenized_datasets.map(
+            group_texts,
+            batched=True,
+            num_proc=data_args.preprocessing_num_workers,
+            load_from_cache_file=not data_args.overwrite_cache,
+        )
+
+    # Enable tensorboard only on the master node
+    has_tensorboard = is_tensorboard_available()
+    if has_tensorboard and jax.process_index() == 0:
+        try:
+            import wandb
+            wandb.init(
+                entity='wandb',
+                project='hf-flax-alberti-poetry',
+                sync_tensorboard=True,
+            )
+            wandb.config.update(training_args)
+            wandb.config.update(model_args)
+            wandb.config.update(data_args)
+
+            from flax.metrics.tensorboard import SummaryWriter
+
+            summary_writer = SummaryWriter(log_dir=Path(training_args.output_dir) / "logs")
+
+
+        except ImportError as ie:
+            has_tensorboard = False
+            logger.warning(
+                f"Unable to display metrics through TensorBoard because some package are not installed: {ie}"
+            )
+    else:
+        logger.warning(
+            "Unable to display metrics through TensorBoard because the package is not installed: "
+            "Please run pip install tensorboard to enable."
+        )
+
+    # Data collator
+    # This one will take care of randomly masking the tokens.
+    data_collator = FlaxDataCollatorForLanguageModeling(tokenizer=tokenizer, mlm_probability=data_args.mlm_probability)
+
+    # Initialize our training
+    rng = jax.random.PRNGKey(training_args.seed)
+    dropout_rngs = jax.random.split(rng, jax.local_device_count())
+
+    if model_args.model_name_or_path:
+        model = FlaxAutoModelForMaskedLM.from_pretrained(
+            model_args.model_name_or_path, config=config, seed=training_args.seed, dtype=getattr(jnp, model_args.dtype)
+        )
+    else:
+        model = FlaxAutoModelForMaskedLM.from_config(
+            config, seed=training_args.seed, dtype=getattr(jnp, model_args.dtype)
+        )
+
+    # Store some constant
+    num_epochs = int(training_args.num_train_epochs)
+    train_batch_size = int(training_args.per_device_train_batch_size) * jax.device_count()
+    eval_batch_size = int(training_args.per_device_eval_batch_size) * jax.device_count()
+
+    num_train_steps = len(tokenized_datasets["train"]) // train_batch_size * num_epochs
+
+    # Create learning rate schedule
+    warmup_fn = optax.linear_schedule(
+        init_value=0.0, end_value=training_args.learning_rate, transition_steps=training_args.warmup_steps
+    )
+    decay_fn = optax.linear_schedule(
+        init_value=training_args.learning_rate,
+        end_value=0,
+        transition_steps=num_train_steps - training_args.warmup_steps,
+    )
+    linear_decay_lr_schedule_fn = optax.join_schedules(
+        schedules=[warmup_fn, decay_fn], boundaries=[training_args.warmup_steps]
+    )
+
+    # We use Optax's "masking" functionality to not apply weight decay
+    # to bias and LayerNorm scale parameters. decay_mask_fn returns a
+    # mask boolean with the same structure as the parameters.
+    # The mask is True for parameters that should be decayed.
+    # Note that this mask is specifically adapted for FlaxBERT-like models.
+    # For other models, one should correct the layer norm parameter naming
+    # accordingly.
+    def decay_mask_fn(params):
+        flat_params = traverse_util.flatten_dict(params)
+        flat_mask = {path: (path[-1] != "bias" and path[-2:] != ("LayerNorm", "scale")) for path in flat_params}
+        return traverse_util.unflatten_dict(flat_mask)
+
+    # create adam optimizer
+    if training_args.adafactor:
+        # We use the default parameters here to initialize adafactor,
+        # For more details about the parameters please check https://github.com/deepmind/optax/blob/ed02befef9bf81cbbf236be3d2b0e032e9ed4a40/optax/_src/alias.py#L74
+        optimizer = optax.adafactor(
+            learning_rate=linear_decay_lr_schedule_fn,
+        )
+    else:
+        optimizer = optax.adamw(
+            learning_rate=linear_decay_lr_schedule_fn,
+            b1=training_args.adam_beta1,
+            b2=training_args.adam_beta2,
+            eps=training_args.adam_epsilon,
+            weight_decay=training_args.weight_decay,
+            mask=decay_mask_fn,
+        )
+
+    # Setup train state
+    state = train_state.TrainState.create(apply_fn=model.__call__, params=model.params, tx=optimizer)
+
+    # Define gradient update step fn
+    def train_step(state, batch, dropout_rng):
+        dropout_rng, new_dropout_rng = jax.random.split(dropout_rng)
+
+        def loss_fn(params):
+            labels = batch.pop("labels")
+
+            logits = state.apply_fn(**batch, params=params, dropout_rng=dropout_rng, train=True)[0]
+
+            # compute loss, ignore padded input tokens
+            label_mask = jnp.where(labels > 0, 1.0, 0.0)
+            loss = optax.softmax_cross_entropy(logits, onehot(labels, logits.shape[-1])) * label_mask
+
+            # take average
+            loss = loss.sum() / label_mask.sum()
+
+            return loss
+
+        grad_fn = jax.value_and_grad(loss_fn)
+        loss, grad = grad_fn(state.params)
+        grad = jax.lax.pmean(grad, "batch")
+        new_state = state.apply_gradients(grads=grad)
+
+        metrics = jax.lax.pmean(
+            {"loss": loss, "learning_rate": linear_decay_lr_schedule_fn(state.step)}, axis_name="batch"
+        )
+
+        return new_state, metrics, new_dropout_rng
+
+    # Create parallel version of the train step
+    p_train_step = jax.pmap(train_step, "batch", donate_argnums=(0,))
+
+    # Define eval fn
+    def eval_step(params, batch):
+        labels = batch.pop("labels")
+
+        logits = model(**batch, params=params, train=False)[0]
+
+        # compute loss, ignore padded input tokens
+        label_mask = jnp.where(labels > 0, 1.0, 0.0)
+        loss = optax.softmax_cross_entropy(logits, onehot(labels, logits.shape[-1])) * label_mask
+
+        # compute accuracy
+        accuracy = jnp.equal(jnp.argmax(logits, axis=-1), labels) * label_mask
+
+        # summarize metrics
+        metrics = {"loss": loss.sum(), "accuracy": accuracy.sum(), "normalizer": label_mask.sum()}
+        metrics = jax.lax.psum(metrics, axis_name="batch")
+
+        return metrics
+
+    p_eval_step = jax.pmap(eval_step, "batch", donate_argnums=(0,))
+
+    # Replicate the train state on each device
+    state = jax_utils.replicate(state)
+
+    train_time = 0
+    epochs = tqdm(range(num_epochs), desc=f"Epoch ... (1/{num_epochs})", position=0)
+    for epoch in epochs:
+        # ======================== Training ================================
+        train_start = time.time()
+        train_metrics = []
+
+        # Create sampling rng
+        rng, input_rng = jax.random.split(rng)
+
+        # Generate an epoch by shuffling sampling indices from the train dataset
+        num_train_samples = len(tokenized_datasets["train"])
+        train_samples_idx = jax.random.permutation(input_rng, jnp.arange(num_train_samples))
+        train_batch_idx = generate_batch_splits(train_samples_idx, train_batch_size)
+
+        # Gather the indexes for creating the batch and do a training step
+        for step, batch_idx in enumerate(tqdm(train_batch_idx, desc="Training...", position=1)):
+            samples = [tokenized_datasets["train"][int(idx)] for idx in batch_idx]
+            model_inputs = data_collator(samples, pad_to_multiple_of=16)
+
+            # Model forward
+            model_inputs = shard(model_inputs.data)
+            state, train_metric, dropout_rngs = p_train_step(state, model_inputs, dropout_rngs)
+            train_metrics.append(train_metric)
+
+            cur_step = epoch * (num_train_samples // train_batch_size) + step
+
+            if cur_step % training_args.logging_steps == 0 and cur_step > 0:
+                # Save metrics
+                train_metric = jax_utils.unreplicate(train_metric)
+                train_time += time.time() - train_start
+                if has_tensorboard and jax.process_index() == 0:
+                    write_train_metric(summary_writer, train_metrics, train_time, cur_step)
+
+                epochs.write(
+                    f"Step... ({cur_step} | Loss: {train_metric['loss']}, Learning Rate: {train_metric['learning_rate']})"
+                )
+
+                train_metrics = []
+
+            if cur_step % training_args.eval_steps == 0 and cur_step > 0:
+                # ======================== Evaluating ==============================
+                num_eval_samples = len(tokenized_datasets["validation"])
+                eval_samples_idx = jnp.arange(num_eval_samples)
+                eval_batch_idx = generate_batch_splits(eval_samples_idx, eval_batch_size)
+
+                eval_metrics = []
+                for i, batch_idx in enumerate(tqdm(eval_batch_idx, desc="Evaluating ...", position=2)):
+                    samples = [tokenized_datasets["validation"][int(idx)] for idx in batch_idx]
+                    model_inputs = data_collator(samples, pad_to_multiple_of=16)
+
+                    # Model forward
+                    model_inputs = shard(model_inputs.data)
+                    metrics = p_eval_step(state.params, model_inputs)
+                    eval_metrics.append(metrics)
+
+                # normalize eval metrics
+                eval_metrics = get_metrics(eval_metrics)
+                eval_metrics = jax.tree_map(jnp.sum, eval_metrics)
+                eval_normalizer = eval_metrics.pop("normalizer")
+                eval_metrics = jax.tree_map(lambda x: x / eval_normalizer, eval_metrics)
+
+                # Update progress bar
+                epochs.desc = f"Step... ({cur_step} | Loss: {eval_metrics['loss']}, Acc: {eval_metrics['accuracy']})"
+
+                # Save metrics
+                if has_tensorboard and jax.process_index() == 0:
+                    cur_step = epoch * (len(tokenized_datasets["train"]) // train_batch_size)
+                    write_eval_metric(summary_writer, eval_metrics, cur_step)
+
+           # if cur_step % training_args.save_steps == 0 and cur_step > 0:
+           #     # save checkpoint after each epoch and push checkpoint to the hub
+           #     if jax.process_index() == 0:
+           #         params = jax.device_get(jax.tree_map(lambda x: x[0], state.params))
+           #         model.save_pretrained(
+           #             training_args.output_dir,
+           #             params=params,
+           #             push_to_hub=training_args.push_to_hub,
+           #             commit_message=f"Saving weights and logs of step {cur_step}",
+           #         )
+        # save checkpoint after eval_steps
+            if step % training_args.save_steps == 0 and step > 0 and jax.process_index() == 0:
+                logger.info(f"Saving checkpoint at {step + 1} steps")
+                params = jax.device_get(jax.tree_map(lambda x: x[0], state.params))
+                model.save_pretrained(
+                    training_args.output_dir,
+                    params=params,
+                    push_to_hub=training_args.push_to_hub,
+                    commit_message=f"Saving weights and logs of step {step}",
+                )
+                save_checkpoint_files(state, data_collator, training_args, training_args.output_dir)
+                checkpoints_dir = Path(training_args.output_dir) / "checkpoints" / f"checkpoint-{step}"
+                checkpoints_dir.mkdir(parents=True, exist_ok=True)
+                model.save_pretrained(checkpoints_dir, params=params,)
+                save_checkpoint_files(state, data_collator, training_args, checkpoints_dir)
+                rotate_checkpoints(
+                    Path(training_args.output_dir) / "checkpoints",
+                    max_checkpoints=training_args.save_total_limit
+                )

run_mlm_flax_eval.py

@@ -0,0 +1,828 @@
+#!/usr/bin/env python
+# coding=utf-8
+# Copyright 2021 The HuggingFace 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
+# limitations under the License.
+"""
+Fine-tuning the library models for masked language modeling (BERT, ALBERT, RoBERTa...) with whole word masking on a
+text file or a dataset.
+
+Here is the full list of checkpoints on the hub that can be fine-tuned by this script:
+https://huggingface.co/models?filter=fill-mask
+"""
+import json
+import logging
+import math
+import os
+import sys
+import time
+from dataclasses import asdict, dataclass, field
+from enum import Enum
+from itertools import chain
+
+# You can also adapt this script on your own masked language modeling task. Pointers for this are left as comments.
+from pathlib import Path
+from typing import Dict, List, Optional, Tuple
+
+import numpy as np
+from datasets import load_dataset
+from tqdm import tqdm
+
+import flax
+import jax
+import jax.numpy as jnp
+import optax
+from flax import jax_utils, traverse_util
+from flax.training import train_state
+from flax.training.common_utils import get_metrics, onehot, shard
+from huggingface_hub import Repository
+from transformers import (
+    CONFIG_MAPPING,
+    FLAX_MODEL_FOR_MASKED_LM_MAPPING,
+    AutoConfig,
+    AutoTokenizer,
+    FlaxAutoModelForMaskedLM,
+    HfArgumentParser,
+    PreTrainedTokenizerBase,
+    TensorType,
+    is_tensorboard_available,
+    set_seed,
+)
+from transformers.file_utils import get_full_repo_name
+
+MODEL_CONFIG_CLASSES = list(FLAX_MODEL_FOR_MASKED_LM_MAPPING.keys())
+MODEL_TYPES = tuple(conf.model_type for conf in MODEL_CONFIG_CLASSES)
+
+
+@dataclass
+class TrainingArguments:
+    output_dir: str = field(
+        metadata={"help": "The output directory where the model predictions and checkpoints will be written."},
+    )
+    overwrite_output_dir: bool = field(
+        default=False,
+        metadata={
+            "help": (
+                "Overwrite the content of the output directory. "
+                "Use this to continue training if output_dir points to a checkpoint directory."
+            )
+        },
+    )
+    do_train: bool = field(default=False, metadata={"help": "Whether to run training."})
+    do_eval: bool = field(default=False, metadata={"help": "Whether to run eval on the dev set."})
+    per_device_train_batch_size: int = field(
+        default=8, metadata={"help": "Batch size per GPU/TPU core/CPU for training."}
+    )
+    per_device_eval_batch_size: int = field(
+        default=8, metadata={"help": "Batch size per GPU/TPU core/CPU for evaluation."}
+    )
+    learning_rate: float = field(default=5e-5, metadata={"help": "The initial learning rate for AdamW."})
+    weight_decay: float = field(default=0.0, metadata={"help": "Weight decay for AdamW if we apply some."})
+    adam_beta1: float = field(default=0.9, metadata={"help": "Beta1 for AdamW optimizer"})
+    adam_beta2: float = field(default=0.999, metadata={"help": "Beta2 for AdamW optimizer"})
+    adam_epsilon: float = field(default=1e-8, metadata={"help": "Epsilon for AdamW optimizer."})
+    adafactor: bool = field(default=False, metadata={"help": "Whether or not to replace AdamW by Adafactor."})
+    num_train_epochs: float = field(default=3.0, metadata={"help": "Total number of training epochs to perform."})
+    warmup_steps: int = field(default=0, metadata={"help": "Linear warmup over warmup_steps."})
+    logging_steps: int = field(default=500, metadata={"help": "Log every X updates steps."})
+    save_steps: int = field(default=500, metadata={"help": "Save checkpoint every X updates steps."})
+    eval_steps: int = field(default=None, metadata={"help": "Run an evaluation every X steps."})
+    seed: int = field(default=42, metadata={"help": "Random seed that will be set at the beginning of training."})
+    push_to_hub: bool = field(
+        default=False, metadata={"help": "Whether or not to upload the trained model to the model hub after training."}
+    )
+    hub_model_id: str = field(
+        default=None, metadata={"help": "The name of the repository to keep in sync with the local `output_dir`."}
+    )
+    hub_token: str = field(default=None, metadata={"help": "The token to use to push to the Model Hub."})
+
+    def __post_init__(self):
+        if self.output_dir is not None:
+            self.output_dir = os.path.expanduser(self.output_dir)
+
+    def to_dict(self):
+        """
+        Serializes this instance while replace `Enum` by their values (for JSON serialization support). It obfuscates
+        the token values by removing their value.
+        """
+        d = asdict(self)
+        for k, v in d.items():
+            if isinstance(v, Enum):
+                d[k] = v.value
+            if isinstance(v, list) and len(v) > 0 and isinstance(v[0], Enum):
+                d[k] = [x.value for x in v]
+            if k.endswith("_token"):
+                d[k] = f"<{k.upper()}>"
+        return d
+
+
+@dataclass
+class ModelArguments:
+    """
+    Arguments pertaining to which model/config/tokenizer we are going to fine-tune, or train from scratch.
+    """
+
+    model_name_or_path: Optional[str] = field(
+        default=None,
+        metadata={
+            "help": "The model checkpoint for weights initialization."
+            "Don't set if you want to train a model from scratch."
+        },
+    )
+    model_type: Optional[str] = field(
+        default=None,
+        metadata={"help": "If training from scratch, pass a model type from the list: " + ", ".join(MODEL_TYPES)},
+    )
+    config_name: Optional[str] = field(
+        default=None, metadata={"help": "Pretrained config name or path if not the same as model_name"}
+    )
+    tokenizer_name: Optional[str] = field(
+        default=None, metadata={"help": "Pretrained tokenizer name or path if not the same as model_name"}
+    )
+    cache_dir: Optional[str] = field(
+        default=None, metadata={"help": "Where do you want to store the pretrained models downloaded from s3"}
+    )
+    use_fast_tokenizer: bool = field(
+        default=True,
+        metadata={"help": "Whether to use one of the fast tokenizer (backed by the tokenizers library) or not."},
+    )
+    dtype: Optional[str] = field(
+        default="float32",
+        metadata={
+            "help": "Floating-point format in which the model weights should be initialized and trained. Choose one of `[float32, float16, bfloat16]`."
+        },
+    )
+
+
+@dataclass
+class DataTrainingArguments:
+    """
+    Arguments pertaining to what data we are going to input our model for training and eval.
+    """
+
+    dataset_name: Optional[str] = field(
+        default=None, metadata={"help": "The name of the dataset to use (via the datasets library)."}
+    )
+    dataset_config_name: Optional[str] = field(
+        default=None, metadata={"help": "The configuration name of the dataset to use (via the datasets library)."}
+    )
+    train_file: Optional[str] = field(default=None, metadata={"help": "The input training data file (a text file)."})
+    validation_file: Optional[str] = field(
+        default=None,
+        metadata={"help": "An optional input evaluation data file to evaluate the perplexity on (a text file)."},
+    )
+    train_ref_file: Optional[str] = field(
+        default=None,
+        metadata={"help": "An optional input train ref data file for whole word masking in Chinese."},
+    )
+    validation_ref_file: Optional[str] = field(
+        default=None,
+        metadata={"help": "An optional input validation ref data file for whole word masking in Chinese."},
+    )
+    overwrite_cache: bool = field(
+        default=False, metadata={"help": "Overwrite the cached training and evaluation sets"}
+    )
+    validation_split_percentage: Optional[int] = field(
+        default=5,
+        metadata={
+            "help": "The percentage of the train set used as validation set in case there's no validation split"
+        },
+    )
+    max_seq_length: Optional[int] = field(
+        default=None,
+        metadata={
+            "help": "The maximum total input sequence length after tokenization. Sequences longer "
+            "than this will be truncated. Default to the max input length of the model."
+        },
+    )
+    preprocessing_num_workers: Optional[int] = field(
+        default=None,
+        metadata={"help": "The number of processes to use for the preprocessing."},
+    )
+    mlm_probability: float = field(
+        default=0.15, metadata={"help": "Ratio of tokens to mask for masked language modeling loss"}
+    )
+    pad_to_max_length: bool = field(
+        default=False,
+        metadata={
+            "help": "Whether to pad all samples to `max_seq_length`. "
+            "If False, will pad the samples dynamically when batching to the maximum length in the batch."
+        },
+    )
+    line_by_line: bool = field(
+        default=False,
+        metadata={"help": "Whether distinct lines of text in the dataset are to be handled as distinct sequences."},
+    )
+    lang: Optional[str] = field(
+        default=None,
+        metadata={"help": "Langauge to filter a dataset entries by."},
+    )
+
+
+    def __post_init__(self):
+        if self.dataset_name is None and self.train_file is None and self.validation_file is None:
+            raise ValueError("Need either a dataset name or a training/validation file.")
+        else:
+            if self.train_file is not None:
+                extension = self.train_file.split(".")[-1]
+                assert extension in ["csv", "json", "txt"], "`train_file` should be a csv, a json or a txt file."
+            if self.validation_file is not None:
+                extension = self.validation_file.split(".")[-1]
+                assert extension in ["csv", "json", "txt"], "`validation_file` should be a csv, a json or a txt file."
+
+
+@flax.struct.dataclass
+class FlaxDataCollatorForLanguageModeling:
+    """
+    Data collator used for language modeling. Inputs are dynamically padded to the maximum length of a batch if they
+    are not all of the same length.
+
+    Args:
+        tokenizer (:class:`~transformers.PreTrainedTokenizer` or :class:`~transformers.PreTrainedTokenizerFast`):
+            The tokenizer used for encoding the data.
+        mlm_probability (:obj:`float`, `optional`, defaults to 0.15):
+            The probability with which to (randomly) mask tokens in the input.
+
+    .. note::
+
+        For best performance, this data collator should be used with a dataset having items that are dictionaries or
+        BatchEncoding, with the :obj:`"special_tokens_mask"` key, as returned by a
+        :class:`~transformers.PreTrainedTokenizer` or a :class:`~transformers.PreTrainedTokenizerFast` with the
+        argument :obj:`return_special_tokens_mask=True`.
+    """
+
+    tokenizer: PreTrainedTokenizerBase
+    mlm_probability: float = 0.15
+
+    def __post_init__(self):
+        if self.tokenizer.mask_token is None:
+            raise ValueError(
+                "This tokenizer does not have a mask token which is necessary for masked language modeling. "
+                "You should pass `mlm=False` to train on causal language modeling instead."
+            )
+
+    def __call__(self, examples: List[Dict[str, np.ndarray]], pad_to_multiple_of: int) -> Dict[str, np.ndarray]:
+        # Handle dict or lists with proper padding and conversion to tensor.
+        batch = self.tokenizer.pad(examples, pad_to_multiple_of=pad_to_multiple_of, return_tensors=TensorType.NUMPY)
+
+        # If special token mask has been preprocessed, pop it from the dict.
+        special_tokens_mask = batch.pop("special_tokens_mask", None)
+
+        batch["input_ids"], batch["labels"] = self.mask_tokens(
+            batch["input_ids"], special_tokens_mask=special_tokens_mask
+        )
+        return batch
+
+    def mask_tokens(
+        self, inputs: np.ndarray, special_tokens_mask: Optional[np.ndarray]
+    ) -> Tuple[np.ndarray, np.ndarray]:
+        """
+        Prepare masked tokens inputs/labels for masked language modeling: 80% MASK, 10% random, 10% original.
+        """
+        labels = inputs.copy()
+        # We sample a few tokens in each sequence for MLM training (with probability `self.mlm_probability`)
+        probability_matrix = np.full(labels.shape, self.mlm_probability)
+        special_tokens_mask = special_tokens_mask.astype("bool")
+
+        probability_matrix[special_tokens_mask] = 0.0
+        masked_indices = np.random.binomial(1, probability_matrix).astype("bool")
+        labels[~masked_indices] = -100  # We only compute loss on masked tokens
+
+        # 80% of the time, we replace masked input tokens with tokenizer.mask_token ([MASK])
+        indices_replaced = np.random.binomial(1, np.full(labels.shape, 0.8)).astype("bool") & masked_indices
+        inputs[indices_replaced] = self.tokenizer.convert_tokens_to_ids(self.tokenizer.mask_token)
+
+        # 10% of the time, we replace masked input tokens with random word
+        indices_random = np.random.binomial(1, np.full(labels.shape, 0.5)).astype("bool")
+        indices_random &= masked_indices & ~indices_replaced
+
+        random_words = np.random.randint(self.tokenizer.vocab_size, size=labels.shape, dtype="i4")
+        inputs[indices_random] = random_words[indices_random]
+
+        # The rest of the time (10% of the time) we keep the masked input tokens unchanged
+        return inputs, labels
+
+
+def generate_batch_splits(samples_idx: jnp.ndarray, batch_size: int) -> jnp.ndarray:
+    num_samples = len(samples_idx)
+    samples_to_remove = num_samples % batch_size
+
+    if samples_to_remove != 0:
+        samples_idx = samples_idx[:-samples_to_remove]
+    sections_split = num_samples // batch_size
+    batch_idx = np.split(samples_idx, sections_split)
+    return batch_idx
+
+
+def write_train_metric(summary_writer, train_metrics, train_time, step):
+    summary_writer.scalar("train_time", train_time, step)
+
+    train_metrics = get_metrics(train_metrics)
+    for key, vals in train_metrics.items():
+        tag = f"train_{key}"
+        for i, val in enumerate(vals):
+            summary_writer.scalar(tag, val, step - len(vals) + i + 1)
+
+
+def write_eval_metric(summary_writer, eval_metrics, step):
+    for metric_name, value in eval_metrics.items():
+        summary_writer.scalar(f"eval_{metric_name}", value, step)
+
+
+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()
+
+    if (
+        os.path.exists(training_args.output_dir)
+        and os.listdir(training_args.output_dir)
+        and training_args.do_train
+        and not training_args.overwrite_output_dir
+    ):
+        raise ValueError(
+            f"Output directory ({training_args.output_dir}) already exists and is not empty."
+            "Use --overwrite_output_dir to overcome."
+        )
+
+    # Setup logging
+    logging.basicConfig(
+        format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
+        level=logging.INFO,
+        datefmt="[%X]",
+    )
+
+    # Log on each process the small summary:
+    logger = logging.getLogger(__name__)
+
+    # Set the verbosity to info of the Transformers logger (on main process only):
+    logger.info(f"Training/evaluation parameters {training_args}")
+
+    # Set seed before initializing model.
+    set_seed(training_args.seed)
+
+    # Handle the repository creation
+    if training_args.push_to_hub:
+        if training_args.hub_model_id is None:
+            repo_name = get_full_repo_name(
+                Path(training_args.output_dir).absolute().name, token=training_args.hub_token
+            )
+        else:
+            repo_name = training_args.hub_model_id
+        repo = Repository(training_args.output_dir, clone_from=repo_name)
+
+    # Get the datasets: you can either provide your own CSV/JSON/TXT training and evaluation files (see below)
+    # or just provide the name of one of the public datasets available on the hub at https://huggingface.co/datasets/
+    # (the dataset will be downloaded automatically from the datasets Hub).
+    #
+    # For CSV/JSON files, this script will use the column called 'text' or the first column if no column called
+    # 'text' is found. You can easily tweak this behavior (see below).
+    #
+    # In distributed training, the load_dataset function guarantees that only one local process can concurrently
+    # download the dataset.
+    if data_args.dataset_name is not None:
+        # Downloading and loading a dataset from the hub.
+        datasets = load_dataset(data_args.dataset_name, data_args.dataset_config_name, cache_dir=model_args.cache_dir)
+
+        if "validation" not in datasets.keys():
+            datasets["validation"] = load_dataset(
+                data_args.dataset_name,
+                data_args.dataset_config_name,
+                split=f"train[:{data_args.validation_split_percentage}%]",
+                cache_dir=model_args.cache_dir,
+            )
+            datasets["train"] = load_dataset(
+                data_args.dataset_name,
+                data_args.dataset_config_name,
+                split=f"train[{data_args.validation_split_percentage}%:]",
+                cache_dir=model_args.cache_dir,
+            )
+    else:
+        data_files = {}
+        if data_args.train_file is not None:
+            data_files["train"] = data_args.train_file
+        if data_args.validation_file is not None:
+            data_files["validation"] = data_args.validation_file
+        extension = data_args.train_file.split(".")[-1]
+        if extension == "txt":
+            extension = "text"
+        datasets = load_dataset(extension, data_files=data_files, cache_dir=model_args.cache_dir)
+
+        if "validation" not in datasets.keys():
+            datasets["validation"] = load_dataset(
+                extension,
+                data_files=data_files,
+                split=f"train[:{data_args.validation_split_percentage}%]",
+                cache_dir=model_args.cache_dir,
+            )
+            datasets["train"] = load_dataset(
+                extension,
+                data_files=data_files,
+                split=f"train[{data_args.validation_split_percentage}%:]",
+                cache_dir=model_args.cache_dir,
+            )
+    if data_args.lang is not None:
+        datasets = datasets.filter(lambda x: x["lang"] == data_args.lang)
+    # See more about loading any type of standard or custom dataset (from files, python dict, pandas DataFrame, etc) at
+    # https://huggingface.co/docs/datasets/loading_datasets.html.
+
+    # Load pretrained model and tokenizer
+
+    # Distributed training:
+    # The .from_pretrained methods guarantee that only one local process can concurrently
+    # download model & vocab.
+    if model_args.config_name:
+        config = AutoConfig.from_pretrained(model_args.config_name, cache_dir=model_args.cache_dir)
+    elif model_args.model_name_or_path:
+        config = AutoConfig.from_pretrained(model_args.model_name_or_path, cache_dir=model_args.cache_dir)
+    else:
+        config = CONFIG_MAPPING[model_args.model_type]()
+        logger.warning("You are instantiating a new config instance from scratch.")
+
+    if model_args.tokenizer_name:
+        tokenizer = AutoTokenizer.from_pretrained(
+            model_args.tokenizer_name, cache_dir=model_args.cache_dir, use_fast=model_args.use_fast_tokenizer
+        )
+    elif model_args.model_name_or_path:
+        tokenizer = AutoTokenizer.from_pretrained(
+            model_args.model_name_or_path, cache_dir=model_args.cache_dir, use_fast=model_args.use_fast_tokenizer
+        )
+    else:
+        raise ValueError(
+            "You are instantiating a new tokenizer from scratch. This is not supported by this script."
+            "You can do it from another script, save it, and load it from here, using --tokenizer_name."
+        )
+
+    # Preprocessing the datasets.
+    # First we tokenize all the texts.
+    if training_args.do_train:
+        column_names = datasets["train"].column_names
+    else:
+        column_names = datasets["validation"].column_names
+    text_column_name = "text" if "text" in column_names else column_names[0]
+
+    max_seq_length = min(data_args.max_seq_length, tokenizer.model_max_length)
+
+    if data_args.line_by_line:
+        # When using line_by_line, we just tokenize each nonempty line.
+        padding = "max_length" if data_args.pad_to_max_length else False
+
+        def tokenize_function(examples):
+            # Remove empty lines
+            examples = [line for line in examples if len(line) > 0 and not line.isspace()]
+            return tokenizer(
+                examples,
+                return_special_tokens_mask=True,
+                padding=padding,
+                truncation=True,
+                max_length=max_seq_length,
+            )
+
+        tokenized_datasets = datasets.map(
+            tokenize_function,
+            input_columns=[text_column_name],
+            batched=True,
+            num_proc=data_args.preprocessing_num_workers,
+            remove_columns=column_names,
+            load_from_cache_file=not data_args.overwrite_cache,
+        )
+
+    else:
+        # Otherwise, we tokenize every text, then concatenate them together before splitting them in smaller parts.
+        # We use `return_special_tokens_mask=True` because DataCollatorForLanguageModeling (see below) is more
+        # efficient when it receives the `special_tokens_mask`.
+        def tokenize_function(examples):
+            return tokenizer(examples[text_column_name], return_special_tokens_mask=True)
+
+        tokenized_datasets = datasets.map(
+            tokenize_function,
+            batched=True,
+            num_proc=data_args.preprocessing_num_workers,
+            remove_columns=column_names,
+            load_from_cache_file=not data_args.overwrite_cache,
+        )
+
+        # Main data processing function that will concatenate all texts from our dataset and generate chunks of
+        # max_seq_length.
+        def group_texts(examples):
+            # Concatenate all texts.
+            concatenated_examples = {k: list(chain(*examples[k])) for k in examples.keys()}
+            total_length = len(concatenated_examples[list(examples.keys())[0]])
+            # We drop the small remainder, we could add padding if the model supported it instead of this drop, you can
+            # customize this part to your needs.
+            if total_length >= max_seq_length:
+                total_length = (total_length // max_seq_length) * max_seq_length
+            # Split by chunks of max_len.
+            result = {
+                k: [t[i : i + max_seq_length] for i in range(0, total_length, max_seq_length)]
+                for k, t in concatenated_examples.items()
+            }
+            return result
+
+        # Note that with `batched=True`, this map processes 1,000 texts together, so group_texts throws away a
+        # remainder for each of those groups of 1,000 texts. You can adjust that batch_size here but a higher value
+        # might be slower to preprocess.
+        #
+        # To speed up this part, we use multiprocessing. See the documentation of the map method for more information:
+        # https://huggingface.co/docs/datasets/package_reference/main_classes.html#datasets.Dataset.map
+        tokenized_datasets = tokenized_datasets.map(
+            group_texts,
+            batched=True,
+            num_proc=data_args.preprocessing_num_workers,
+            load_from_cache_file=not data_args.overwrite_cache,
+        )
+
+    # Enable tensorboard only on the master node
+    has_tensorboard = is_tensorboard_available()
+    if has_tensorboard and jax.process_index() == 0:
+        try:
+            # Enable Weight&Biases
+            #import wandb
+            #wandb.init(
+            #    entity='versae',
+            #    project='roberta-base-ncc',
+            #    sync_tensorboard=True,
+            #)
+            #wandb.config.update(training_args)
+            #wandb.config.update(model_args)
+            #wandb.config.update(data_args)
+
+            from flax.metrics.tensorboard import SummaryWriter
+
+            summary_writer = SummaryWriter(log_dir=Path(training_args.output_dir))
+        except ImportError as ie:
+            has_tensorboard = False
+            logger.warning(
+                f"Unable to display metrics through TensorBoard because some package are not installed: {ie}"
+            )
+    else:
+        logger.warning(
+            "Unable to display metrics through TensorBoard because the package is not installed: "
+            "Please run pip install tensorboard to enable."
+        )
+
+    # Data collator
+    # This one will take care of randomly masking the tokens.
+    data_collator = FlaxDataCollatorForLanguageModeling(tokenizer=tokenizer, mlm_probability=data_args.mlm_probability)
+
+    # Initialize our training
+    rng = jax.random.PRNGKey(training_args.seed)
+    dropout_rngs = jax.random.split(rng, jax.local_device_count())
+
+    if model_args.model_name_or_path:
+        model = FlaxAutoModelForMaskedLM.from_pretrained(
+            model_args.model_name_or_path, config=config, seed=training_args.seed, dtype=getattr(jnp, model_args.dtype)
+        )
+    else:
+        model = FlaxAutoModelForMaskedLM.from_config(
+            config, seed=training_args.seed, dtype=getattr(jnp, model_args.dtype)
+        )
+
+    # Store some constant
+    num_epochs = int(training_args.num_train_epochs)
+    train_batch_size = int(training_args.per_device_train_batch_size) * jax.device_count()
+    eval_batch_size = int(training_args.per_device_eval_batch_size) * jax.device_count()
+
+    num_train_steps = len(tokenized_datasets["train"]) // train_batch_size * num_epochs
+
+    # Create learning rate schedule
+    warmup_fn = optax.linear_schedule(
+        init_value=0.0, end_value=training_args.learning_rate, transition_steps=training_args.warmup_steps
+    )
+    decay_fn = optax.linear_schedule(
+        init_value=training_args.learning_rate,
+        end_value=0,
+        transition_steps=num_train_steps - training_args.warmup_steps,
+    )
+    linear_decay_lr_schedule_fn = optax.join_schedules(
+        schedules=[warmup_fn, decay_fn], boundaries=[training_args.warmup_steps]
+    )
+
+    # We use Optax's "masking" functionality to not apply weight decay
+    # to bias and LayerNorm scale parameters. decay_mask_fn returns a
+    # mask boolean with the same structure as the parameters.
+    # The mask is True for parameters that should be decayed.
+    # Note that this mask is specifically adapted for FlaxBERT-like models.
+    # For other models, one should correct the layer norm parameter naming
+    # accordingly.
+    def decay_mask_fn(params):
+        flat_params = traverse_util.flatten_dict(params)
+        flat_mask = {path: (path[-1] != "bias" and path[-2:] != ("LayerNorm", "scale")) for path in flat_params}
+        return traverse_util.unflatten_dict(flat_mask)
+
+    # create adam optimizer
+    if training_args.adafactor:
+        # We use the default parameters here to initialize adafactor,
+        # For more details about the parameters please check https://github.com/deepmind/optax/blob/ed02befef9bf81cbbf236be3d2b0e032e9ed4a40/optax/_src/alias.py#L74
+        optimizer = optax.adafactor(
+            learning_rate=linear_decay_lr_schedule_fn,
+        )
+    else:
+        optimizer = optax.adamw(
+            learning_rate=linear_decay_lr_schedule_fn,
+            b1=training_args.adam_beta1,
+            b2=training_args.adam_beta2,
+            eps=training_args.adam_epsilon,
+            weight_decay=training_args.weight_decay,
+            mask=decay_mask_fn,
+        )
+
+    # Setup train state
+    state = train_state.TrainState.create(apply_fn=model.__call__, params=model.params, tx=optimizer)
+
+    # Define gradient update step fn
+    def train_step(state, batch, dropout_rng):
+        dropout_rng, new_dropout_rng = jax.random.split(dropout_rng)
+
+        def loss_fn(params):
+            labels = batch.pop("labels")
+
+            logits = state.apply_fn(**batch, params=params, dropout_rng=dropout_rng, train=True)[0]
+
+            # compute loss, ignore padded input tokens
+            label_mask = jnp.where(labels > 0, 1.0, 0.0)
+            loss = optax.softmax_cross_entropy(logits, onehot(labels, logits.shape[-1])) * label_mask
+
+            # take average
+            loss = loss.sum() / label_mask.sum()
+
+            return loss
+
+        grad_fn = jax.value_and_grad(loss_fn)
+        loss, grad = grad_fn(state.params)
+        grad = jax.lax.pmean(grad, "batch")
+        new_state = state.apply_gradients(grads=grad)
+
+        metrics = jax.lax.pmean(
+            {"loss": loss, "learning_rate": linear_decay_lr_schedule_fn(state.step)}, axis_name="batch"
+        )
+
+        return new_state, metrics, new_dropout_rng
+
+    # Create parallel version of the train step
+    p_train_step = jax.pmap(train_step, "batch", donate_argnums=(0,))
+
+    # Define eval fn
+    def eval_step(params, batch):
+        labels = batch.pop("labels")
+
+        logits = model(**batch, params=params, train=False)[0]
+
+        # compute loss, ignore padded input tokens
+        label_mask = jnp.where(labels > 0, 1.0, 0.0)
+        loss = optax.softmax_cross_entropy(logits, onehot(labels, logits.shape[-1])) * label_mask
+
+        # compute accuracy
+        accuracy = jnp.equal(jnp.argmax(logits, axis=-1), labels) * label_mask
+
+        # summarize metrics
+        metrics = {"loss": loss.sum(), "accuracy": accuracy.sum(), "normalizer": label_mask.sum()}
+        metrics = jax.lax.psum(metrics, axis_name="batch")
+
+        return metrics
+
+    p_eval_step = jax.pmap(eval_step, "batch", donate_argnums=(0,))
+
+    # Replicate the train state on each device
+    state = jax_utils.replicate(state)
+
+    train_time = 0
+    if training_args.do_train:
+        epochs = tqdm(range(num_epochs), desc=f"Epoch ... (1/{num_epochs})", position=0)
+    else:
+        epochs = []
+    for epoch in epochs:
+        # ======================== Training ================================
+        train_start = time.time()
+        train_metrics = []
+
+        # Create sampling rng
+        rng, input_rng = jax.random.split(rng)
+
+        # Generate an epoch by shuffling sampling indices from the train dataset
+        num_train_samples = len(tokenized_datasets["train"])
+        train_samples_idx = jax.random.permutation(input_rng, jnp.arange(num_train_samples))
+        train_batch_idx = generate_batch_splits(train_samples_idx, train_batch_size)
+
+        # Gather the indexes for creating the batch and do a training step
+        for step, batch_idx in enumerate(tqdm(train_batch_idx, desc="Training...", position=1)):
+            samples = [tokenized_datasets["train"][int(idx)] for idx in batch_idx]
+            model_inputs = data_collator(samples, pad_to_multiple_of=16)
+
+            # Model forward
+            model_inputs = shard(model_inputs.data)
+            state, train_metric, dropout_rngs = p_train_step(state, model_inputs, dropout_rngs)
+            train_metrics.append(train_metric)
+
+            cur_step = epoch * (num_train_samples // train_batch_size) + step
+
+            if cur_step % training_args.logging_steps == 0 and cur_step > 0:
+                # Save metrics
+                train_metric = jax_utils.unreplicate(train_metric)
+                train_time += time.time() - train_start
+                if has_tensorboard and jax.process_index() == 0:
+                    write_train_metric(summary_writer, train_metrics, train_time, cur_step)
+
+                epochs.write(
+                    f"Step... ({cur_step} | Loss: {train_metric['loss']}, Learning Rate: {train_metric['learning_rate']})"
+                )
+
+                train_metrics = []
+
+            if cur_step % training_args.eval_steps == 0 and cur_step > 0:
+                # ======================== Evaluating ==============================
+                num_eval_samples = len(tokenized_datasets["validation"])
+                eval_samples_idx = jnp.arange(num_eval_samples)
+                eval_batch_idx = generate_batch_splits(eval_samples_idx, eval_batch_size)
+
+                eval_metrics = []
+                for i, batch_idx in enumerate(tqdm(eval_batch_idx, desc="Evaluating ...", position=2)):
+                    samples = [tokenized_datasets["validation"][int(idx)] for idx in batch_idx]
+                    model_inputs = data_collator(samples, pad_to_multiple_of=16)
+
+                    # Model forward
+                    model_inputs = shard(model_inputs.data)
+                    metrics = p_eval_step(state.params, model_inputs)
+                    eval_metrics.append(metrics)
+
+                # normalize eval metrics
+                eval_metrics = get_metrics(eval_metrics)
+                eval_metrics = jax.tree_map(jnp.sum, eval_metrics)
+                eval_normalizer = eval_metrics.pop("normalizer")
+                eval_metrics = jax.tree_map(lambda x: x / eval_normalizer, eval_metrics)
+
+                # Update progress bar
+                epochs.desc = f"Step... ({cur_step} | Loss: {eval_metrics['loss']}, Acc: {eval_metrics['accuracy']})"
+
+                # Save metrics
+                if has_tensorboard and jax.process_index() == 0:
+                    write_eval_metric(summary_writer, eval_metrics, cur_step)
+
+            if cur_step % training_args.save_steps == 0 and cur_step > 0:
+                # save checkpoint after each epoch and push checkpoint to the hub
+                if jax.process_index() == 0:
+                    params = jax.device_get(jax.tree_map(lambda x: x[0], state.params))
+                    model.save_pretrained(training_args.output_dir, params=params)
+                    tokenizer.save_pretrained(training_args.output_dir)
+                    if training_args.push_to_hub:
+                        repo.push_to_hub(commit_message=f"Saving weights and logs of step {cur_step}", blocking=False)
+
+    # Eval after training
+    if training_args.do_eval:
+        num_eval_samples = len(tokenized_datasets["validation"])
+        eval_samples_idx = jnp.arange(num_eval_samples)
+        eval_batch_idx = generate_batch_splits(eval_samples_idx, eval_batch_size)
+
+        eval_metrics = []
+        for _, batch_idx in enumerate(tqdm(eval_batch_idx, desc="Evaluating ...", position=2)):
+            samples = [tokenized_datasets["validation"][int(idx)] for idx in batch_idx]
+            model_inputs = data_collator(samples, pad_to_multiple_of=16)
+
+            # Model forward
+            model_inputs = shard(model_inputs.data)
+            metrics = p_eval_step(state.params, model_inputs)
+            eval_metrics.append(metrics)
+
+        # normalize eval metrics
+        eval_metrics = get_metrics(eval_metrics)
+        eval_metrics = jax.tree_map(lambda metric: jnp.sum(metric).item(), eval_metrics)
+        eval_normalizer = eval_metrics.pop("normalizer")
+        eval_metrics = jax.tree_map(lambda x: x / eval_normalizer, eval_metrics)
+
+        try:
+            perplexity = math.exp(eval_metrics["loss"])
+        except OverflowError:
+            perplexity = float("inf")
+        eval_metrics["perplexity"] = perplexity
+
+        if data_args.lang is not None:
+            eval_filename = f"eval_results_{data_args.lang}.json"
+        else:
+            eval_filename = "eval_results.json"
+        if jax.process_index() == 0:
+            eval_metrics = {f"eval_{metric_name}": value for metric_name, value in eval_metrics.items()}
+            path = os.path.join(training_args.output_dir, eval_filename)
+            with open(path, "w") as f:
+                json.dump(eval_metrics, f, indent=4, sort_keys=True)
+
+
+if __name__ == "__main__":
+    main()

special_tokens_map.json

@@ -0,0 +1 @@
+{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}

tokenizer_config.json

@@ -0,0 +1 @@
+{"do_lower_case": false, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "bert-base-multilingual-cased", "tokenizer_class": "BertTokenizer"}

training_state.json

@@ -0,0 +1 @@
+{"step": 111279}