run_glue_freeze.py

#!/usr/bin/env python
# coding=utf-8
# Copyright 2021. 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.
""" Finetuning a 🤗 Flax Transformers model for sequence classification on GLUE."""
import json
import logging
import os
import random
import sys
import time
from dataclasses import dataclass, field
from itertools import chain
from pathlib import Path
from typing import Any, Callable, Dict, Optional, Tuple

import datasets
import flax
import jax
import jax.numpy as jnp
import numpy as np
import optax
import wandb
from datasets import load_dataset, load_metric
from flax import struct, traverse_util
from flax.jax_utils import replicate, unreplicate
from flax.training import train_state
from flax.training.common_utils import get_metrics, onehot, shard
from huggingface_hub import Repository
from tqdm import tqdm

from BiGS.configuration_bigs import BiGSConfig
from BiGS.modeling_flax_bigs import FlaxBiGSForSequenceClassification

import transformers
from transformers import (
    AutoTokenizer,
    FlaxAutoModelForSequenceClassification,
    HfArgumentParser,
    PretrainedConfig,
    TrainingArguments,
    is_tensorboard_available,
)
from transformers.utils import check_min_version, get_full_repo_name

logger = logging.getLogger(__name__)
# Will error if the minimal version of Transformers is not installed. Remove at your own risks.
check_min_version("4.19.0.dev0")

Array = Any
Dataset = datasets.arrow_dataset.Dataset
PRNGKey = Any

task_to_keys = {
    "cola": ("sentence", None),
    "mnli": ("premise", "hypothesis"),
    "mrpc": ("sentence1", "sentence2"),
    "qnli": ("question", "sentence"),
    "qqp": ("question1", "question2"),
    "rte": ("sentence1", "sentence2"),
    "sst2": ("sentence", None),
    "stsb": ("sentence1", "sentence2"),
    "wnli": ("sentence1", "sentence2"),
}


@dataclass
class GlueTrainingArguments(TrainingArguments):
    warmup_rate: float = field(default=0.1, metadata={"help": "The warmup rate."})


@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"}
    )
    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"}
    )
    use_slow_tokenizer: Optional[bool] = field(
        default=False,
        metadata={"help": "If passed, will use a slow tokenizer (not backed by the 🤗 Tokenizers library)."},
    )
    cache_dir: Optional[str] = field(
        default=None,
        metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"},
    )
    model_revision: str = field(
        default="main",
        metadata={"help": "The specific model version to use (can be a branch name, tag name or commit id)."},
    )
    use_auth_token: bool = field(
        default=False,
        metadata={
            "help": "Will use the token generated when running `transformers-cli login` (necessary to use this script "
                    "with private models)."
        },
    )


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

    task_name: Optional[str] = field(
        default=None, metadata={"help": f"The name of the glue task to train on. choices {list(task_to_keys.keys())}"}
    )
    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 csv or JSON file)."}
    )
    validation_file: Optional[str] = field(
        default=None,
        metadata={"help": "An optional input evaluation data file to evaluate on (a csv or JSON file)."},
    )
    test_file: Optional[str] = field(
        default=None,
        metadata={"help": "An optional input test data file to predict on (a csv or JSON file)."},
    )
    text_column_name: Optional[str] = field(
        default=None, metadata={"help": "The column name of text to input in the file (a csv or JSON file)."}
    )
    label_column_name: Optional[str] = field(
        default=None, metadata={"help": "The column name of label to input in the file (a csv or JSON file)."}
    )
    overwrite_cache: bool = field(
        default=False, metadata={"help": "Overwrite the cached training and evaluation sets"}
    )
    preprocessing_num_workers: Optional[int] = field(
        default=None,
        metadata={"help": "The number of processes to use for the preprocessing."},
    )
    max_seq_length: int = field(
        default=None,
        metadata={
            "help": "The maximum total input sequence length after tokenization. If set, sequences longer "
                    "than this will be truncated, sequences shorter will be padded."
        },
    )
    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 evaluation examples to this "
                    "value if set."
        },
    )
    max_predict_samples: Optional[int] = field(
        default=None,
        metadata={
            "help": "For debugging purposes or quicker training, truncate the number of prediction examples to this "
                    "value if set."
        },
    )

    def __post_init__(self):
        if self.task_name is None and 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"], "`train_file` should be a csv or a json file."
            if self.validation_file is not None:
                extension = self.validation_file.split(".")[-1]
                assert extension in ["csv", "json"], "`validation_file` should be a csv or a json file."
        self.task_name = self.task_name.lower() if type(self.task_name) == str else self.task_name


def create_train_state(
        model: FlaxAutoModelForSequenceClassification,
        tx,
        is_regression: bool,
        num_labels: int
) -> train_state.TrainState:
    """Create initial training state."""

    class TrainState(train_state.TrainState):
        """Train state with an Optax optimizer.

        The two functions below differ depending on whether the task is classification
        or regression.

        Args:
          logits_fn: Applied to last layer to obtain the logits.
          loss_fn: Function to compute the loss.
        """

        logits_fn: Callable = struct.field(pytree_node=False)
        loss_fn: Callable = struct.field(pytree_node=False)

    # 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.
    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)

    # tx = optax.adamw(
    #     learning_rate=learning_rate_fn, b1=0.9, b2=0.999, eps=1e-6, weight_decay=weight_decay, mask=decay_mask_fn
    # )

    if is_regression:

        def mse_loss(logits, labels):
            return jnp.mean((logits[..., 0] - labels) ** 2)

        return TrainState.create(
            apply_fn=model.__call__,
            params=model.params,
            tx=tx,
            logits_fn=lambda logits: logits[..., 0],
            loss_fn=mse_loss,
        )
    else:  # Classification.

        def cross_entropy_loss(logits, labels):
            xentropy = optax.softmax_cross_entropy(logits, onehot(labels, num_classes=num_labels))
            return jnp.mean(xentropy)

        return TrainState.create(
            apply_fn=model.__call__,
            params=model.params,
            tx=tx,
            logits_fn=lambda logits: logits.argmax(-1),
            loss_fn=cross_entropy_loss,
        )


def create_learning_rate_fn(
        train_ds_size: int, train_batch_size: int, num_train_epochs: int, num_warmup_rate: float, learning_rate: float
) -> Callable[[int], jnp.array]:
    """Returns a linear warmup, linear_decay learning rate function."""
    steps_per_epoch = train_ds_size // train_batch_size
    num_train_steps = steps_per_epoch * num_train_epochs
    num_warmup_steps = num_train_steps * num_warmup_rate
    warmup_fn = optax.linear_schedule(init_value=0.0, end_value=learning_rate, transition_steps=num_warmup_steps)
    decay_fn = optax.linear_schedule(
        init_value=learning_rate, end_value=0, transition_steps=num_train_steps - num_warmup_steps
    )
    schedule_fn = optax.join_schedules(schedules=[warmup_fn, decay_fn], boundaries=[num_warmup_steps])
    return schedule_fn


def glue_train_data_collator(rng: PRNGKey, dataset: Dataset, batch_size: int):
    """Returns shuffled batches of size `batch_size` from truncated `train dataset`, sharded over all local devices."""
    steps_per_epoch = len(dataset) // batch_size
    perms = jax.random.permutation(rng, len(dataset))
    perms = perms[: steps_per_epoch * batch_size]  # Skip incomplete batch.
    perms = perms.reshape((steps_per_epoch, batch_size))

    for perm in perms:
        batch = dataset[perm]
        batch = {k: np.array(v) for k, v in batch.items()}
        batch = shard(batch)

        yield batch


def glue_eval_data_collator(dataset: Dataset, batch_size: int):
    """Returns batches of size `batch_size` from `eval dataset`, sharded over all local devices."""
    for i in range(len(dataset) // batch_size):
        batch = dataset[i * batch_size: (i + 1) * batch_size]
        batch = {k: np.array(v) for k, v in batch.items()}
        batch = shard(batch)

        yield batch


# As we're using Flax, we also write a utility function to return a default TrainState object.
# This function initializes model parameters, as well as our optimizer. Note that for S4 models,
# we use a custom learning rate for parameters of the S4 kernel (lr = 0.001, no weight decay).
def map_nested_fn(fn):
    """Recursively apply `fn to the key-value pairs of a nested dict / pytree."""

    def map_fn(nested_dict):
        return {
            k: (map_fn(v) if hasattr(v, "keys") else fn(k, v))
            for k, v in nested_dict.items()
        }

    return map_fn


def main():
    parser = HfArgumentParser((ModelArguments, DataTrainingArguments, GlueTrainingArguments))
    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()

    # Make one log on every process with the configuration for debugging.
    logging.basicConfig(
        format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
        datefmt="%m/%d/%Y %H:%M:%S",
        level=logging.INFO,
    )
    # Setup logging, we only want one process per machine to log things on the screen.
    logger.setLevel(logging.INFO if jax.process_index() == 0 else logging.ERROR)
    if jax.process_index() == 0:
        datasets.utils.logging.set_verbosity_warning()
        transformers.utils.logging.set_verbosity_info()
    else:
        datasets.utils.logging.set_verbosity_error()
        transformers.utils.logging.set_verbosity_error()

    # 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 training and evaluation files (see below)
    # or specify a GLUE benchmark task (the dataset will be downloaded automatically from the datasets Hub).

    # For CSV/JSON files, this script will use as labels the column called 'label' and as pair of sentences the
    # sentences in columns called 'sentence1' and 'sentence2' if such column exists or the first two columns not named
    # label if at least two columns are provided.

    # If the CSVs/JSONs contain only one non-label column, the script does single sentence classification on this
    # single column. You can easily tweak this behavior (see below)

    # In distributed training, the load_dataset function guarantee that only one local process can concurrently
    # download the dataset.
    if data_args.task_name is not None:
        # Downloading and loading a dataset from the hub.
        raw_datasets = load_dataset(
            "glue",
            data_args.task_name,
            use_auth_token=True if model_args.use_auth_token else None,
        )
    elif data_args.dataset_name is not None:
        # Downloading and loading a dataset from the hub.
        raw_datasets = load_dataset(
            data_args.dataset_name,
            data_args.dataset_config_name,
            cache_dir=model_args.cache_dir,
            use_auth_token=True if model_args.use_auth_token else None,
        )
    else:
        # Loading the dataset from local csv or json file.
        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 if data_args.train_file is not None else data_args.valid_file).split(".")[-1]
        raw_datasets = load_dataset(
            extension,
            data_files=data_files,
            use_auth_token=True if model_args.use_auth_token else None,
        )
    # See more about loading any type of standard or custom dataset at
    # https://huggingface.co/docs/datasets/loading_datasets.html.

    # Labels
    if data_args.task_name is not None:
        is_regression = data_args.task_name == "stsb"
        if not is_regression:
            label_list = raw_datasets["train"].features["label"].names
            num_labels = len(label_list)
        else:
            num_labels = 1
    else:
        if data_args.dataset_name == "hyperpartisan_news_detection":
            label_list = raw_datasets["train"].unique("hyperpartisan")
            label_list.sort()  # Let's sort it for determinism
            num_labels = len(label_list)
            is_regression = 0
        else:
            # Trying to have good defaults here, don't hesitate to tweak to your needs.
            is_regression = raw_datasets["train"].features["label"].dtype in ["float32", "float64"]
            if is_regression:
                num_labels = 1
            else:
                # A useful fast method:
                # https://huggingface.co/docs/datasets/package_reference/main_classes.html#datasets.Dataset.unique
                label_list = raw_datasets["train"].unique("label")
                label_list.sort()  # Let's sort it for determinism
                num_labels = len(label_list)

    # Load pretrained model and tokenizer
    config = BiGSConfig.from_pretrained(
        model_args.model_name_or_path,
        num_labels=num_labels,
        finetuning_task=data_args.task_name,
        use_auth_token=True if model_args.use_auth_token else None,
    )
    tokenizer = AutoTokenizer.from_pretrained(
        model_args.model_name_or_path,
        use_fast=not model_args.use_slow_tokenizer,
        use_auth_token=True if model_args.use_auth_token else None,
    )
    model = FlaxBiGSForSequenceClassification.from_pretrained(
        model_args.model_name_or_path,
        ignore_mismatched_sizes=True,
        config=config,
        use_auth_token=True if model_args.use_auth_token else None,
    )

    # Preprocessing the datasets
    if data_args.task_name is not None:
        sentence1_key, sentence2_key = task_to_keys[data_args.task_name]
    else:
        # Again, we try to have some nice defaults but don't hesitate to tweak to your use case.
        non_label_column_names = [name for name in raw_datasets["train"].column_names if name != "label"]
        if "sentence1" in non_label_column_names and "sentence2" in non_label_column_names:
            sentence1_key, sentence2_key = "sentence1", "sentence2"
        else:
            if len(non_label_column_names) >= 2:
                sentence1_key, sentence2_key = non_label_column_names[:2]
            else:
                sentence1_key, sentence2_key = non_label_column_names[0], None

    # Some models have set the order of the labels to use, so let's make sure we do use it.
    label_to_id = None
    if (
            model.config.label2id != PretrainedConfig(num_labels=num_labels).label2id
            and data_args.task_name is not None
            and not is_regression
    ):
        # Some have all caps in their config, some don't.
        label_name_to_id = {k.lower(): v for k, v in model.config.label2id.items()}
        if list(sorted(label_name_to_id.keys())) == list(sorted(label_list)):
            logger.info(
                f"The configuration of the model provided the following label correspondence: {label_name_to_id}. "
                "Using it!"
            )
            label_to_id = {i: label_name_to_id[label_list[i]] for i in range(num_labels)}
        else:
            logger.warning(
                "Your model seems to have been trained with labels, but they don't match the dataset: ",
                f"model labels: {list(sorted(label_name_to_id.keys()))}, dataset labels: {list(sorted(label_list))}."
                "\nIgnoring the model labels as a result.",
            )
    elif data_args.task_name is None:
        label_to_id = {v: i for i, v in enumerate(label_list)}

    def preprocess_function(examples):
        # Tokenize the texts
        if data_args.dataset_name == "tau/scrolls" and data_args.dataset_config_name == "contract_nli":
            contexts1 = [example.split("\n\n")[0].strip() for example in examples[sentence1_key]]
            contexts2 = [example.split("\n\n")[1].strip() for example in examples[sentence1_key]]
            texts = ((contexts1, contexts2))
            result = tokenizer(*texts, padding="max_length", max_length=data_args.max_seq_length, truncation=True)
        else:
            if sentence2_key is not None:
                texts = [examples[sentence1_key][idx] + ' [CLS], ' + examples[sentence2_key][idx] for idx in
                         range(len(examples[sentence1_key]))]
            else:
                texts = [examples[sentence1_key][idx] for idx in range(len(examples[sentence1_key]))]
            # tokenizer texts
            result = tokenizer(texts, padding="max_length", max_length=data_args.max_seq_length, truncation=True)
        if data_args.dataset_name == "hyperpartisan_news_detection":
            result['labels'] = [int(example_label) for example_label in examples['hyperpartisan']]
        else:
            if "label" in examples:
                if label_to_id is not None:
                    # Map labels to IDs (not necessary for GLUE tasks)
                    result["labels"] = [(label_to_id[example_label] if example_label != -1 else -1) for example_label in
                                        examples["label"]]
                else:
                    # In all cases, rename the column to labels because the model will expect that.
                    result["labels"] = examples["label"]
        return result

    processed_datasets = raw_datasets.map(
        preprocess_function, batched=True, remove_columns=raw_datasets["train"].column_names
    )

    train_dataset = processed_datasets["train"]
    if data_args.dataset_name == "imdb":
        eval_dataset = processed_datasets["test"]
    elif data_args.dataset_name == "hyperpartisan_news_detection":
        train_testvalid = train_dataset.train_test_split(test_size=0.2)
        test_valid = train_testvalid["test"].train_test_split(test_size=0.5)
        train_dataset = train_testvalid["train"]
        eval_dataset = test_valid["train"]
        # test_dataset = test_valid["test"]
    else:
        eval_dataset = processed_datasets["validation_matched" if data_args.task_name == "mnli" else "validation"]

    if training_args.do_predict or data_args.task_name is not None or data_args.test_file is not None:
        if "test" not in raw_datasets and "test_matched" not in raw_datasets:
            raise ValueError("--do_predict requires a test dataset")
        predict_dataset = processed_datasets["test_matched" if data_args.task_name == "mnli" else "test"]
        if data_args.max_predict_samples is not None:
            max_predict_samples = min(len(predict_dataset), data_args.max_predict_samples)
            predict_dataset = predict_dataset.select(range(max_predict_samples))

    # Log a few random samples from the training set:
    for index in random.sample(range(len(train_dataset)), 3):
        logger.info(f"Sample {index} of the training set: {train_dataset[index]}.")

    # Define a summary writer
    has_tensorboard = is_tensorboard_available()
    if has_tensorboard and jax.process_index() == 0:
        try:
            from flax.metrics.tensorboard import SummaryWriter

            summary_writer = SummaryWriter(training_args.output_dir)
            summary_writer.hparams({**training_args.to_dict(), **vars(model_args), **vars(data_args)})
        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."
        )

    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)

    num_epochs = int(training_args.num_train_epochs)
    rng = jax.random.PRNGKey(training_args.seed)
    dropout_rngs = jax.random.split(rng, jax.local_device_count())

    train_batch_size = training_args.per_device_train_batch_size * jax.local_device_count()
    eval_batch_size = training_args.per_device_eval_batch_size * jax.local_device_count()

    schedule_fn = create_learning_rate_fn(
        len(train_dataset),
        train_batch_size,
        training_args.num_train_epochs,
        training_args.warmup_rate,
        training_args.learning_rate,
    )

    lr_layer = {
        "A_re": 0,
        "A_im": 0,
        "C": 0,
        "D": 0,
        "log_step": 0,
    }
    print("lr_layer:", lr_layer.keys())

    def flattened_traversal(fn):
        """Returns function that is called with `(path, param)` instead of pytree."""

        def mask(tree):
            flat = flax.traverse_util.flatten_dict(tree)
            for k, v in flat.items():
                print(k)
            mfun = flax.traverse_util.unflatten_dict({k: fn(k, v) for k, v in flat.items()})
            return mfun

        return mask

    # Freezes all but the last layer.
    label_fn = flattened_traversal(lambda path, _: 'none' if path[-1] in lr_layer.keys() else 'adamw')

    tx = optax.multi_transform(
        {'adamw': optax.adamw(learning_rate=schedule_fn, weight_decay=training_args.weight_decay),
         'none': optax.set_to_zero()}, label_fn)

    state = create_train_state(
        model, tx, is_regression, num_labels=num_labels
    )

    # define step functions
    def train_step(
            state: train_state.TrainState, batch: Dict[str, Array], dropout_rng: PRNGKey
    ) -> Tuple[train_state.TrainState, float]:
        """Trains model with an optimizer (both in `state`) on `batch`, returning a pair `(new_state, loss)`."""
        dropout_rng, new_dropout_rng = jax.random.split(dropout_rng)
        targets = batch.pop("labels")

        def loss_fn(params):
            logits = state.apply_fn(**batch, params=params, dropout_rng=dropout_rng, train=True)[0]
            loss = state.loss_fn(logits, targets)
            return loss, logits

        grad_fn = jax.value_and_grad(loss_fn, has_aux=True)
        (loss, logits), grad = grad_fn(state.params)
        grad = jax.lax.pmean(grad, "batch")
        accuracy = jnp.sum(jnp.equal(jnp.argmax(logits, axis=-1), targets)) / jnp.size(targets)
        new_state = state.apply_gradients(grads=grad)
        metrics = jax.lax.pmean(
            {"train loss": loss, "train accuracy": accuracy, "learning_rate": schedule_fn(state.step)},
            axis_name="batch")
        return new_state, metrics, new_dropout_rng

    p_train_step = jax.pmap(train_step, axis_name="batch", donate_argnums=(0,))

    def eval_step(state, batch):
        targets = batch.pop("labels")
        logits = state.apply_fn(**batch, params=state.params, train=False)[0]
        loss = state.loss_fn(logits, targets)
        accuracy = jnp.sum(jnp.equal(jnp.argmax(logits, axis=-1), targets)) / jnp.size(targets)
        eval_additional_metrics = jax.lax.pmean({"test loss": loss, "test accuracy": accuracy}, axis_name="batch")

        return state.logits_fn(logits), eval_additional_metrics

    p_eval_step = jax.pmap(eval_step, axis_name="batch")

    if data_args.task_name is not None:
        metric = load_metric("glue", data_args.task_name)
    elif data_args.dataset_name == "hyperpartisan_news_detection":
        metric = load_metric("f1")
    else:
        metric = load_metric("accuracy")

    if data_args.task_name is not None:
        wandb.init(project=f"{data_args.task_name}_BiGS_freeze_kernel")
    elif data_args.dataset_name is not None:
        wandb.init(project=f"{data_args.dataset_name}_BiGS_freeze_kernel")

    wandb.config.update(training_args)

    logger.info(f"===== Starting training ({num_epochs} epochs) =====")
    train_time = 0

    # make sure weights are replicated on each device
    state = replicate(state)

    steps_per_epoch = len(train_dataset) // train_batch_size
    total_steps = steps_per_epoch * num_epochs
    epochs = tqdm(range(num_epochs), desc=f"Epoch ... (0/{num_epochs})", position=0)
    for epoch in epochs:

        train_start = time.time()
        train_metrics = []

        # Create sampling rng
        rng, input_rng = jax.random.split(rng)

        # train
        train_loader = glue_train_data_collator(input_rng, train_dataset, train_batch_size)
        for step, batch in enumerate(
                tqdm(
                    train_loader,
                    total=steps_per_epoch,
                    desc="Training...",
                    position=1,
                ),
        ):
            state, train_metric, dropout_rngs = p_train_step(state, batch, dropout_rngs)
            train_metrics.append(train_metric)

            cur_step = (epoch * steps_per_epoch) + (step + 1)

            if cur_step % training_args.logging_steps == 0 and cur_step > 0:
                # Save metrics
                train_metric = 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}/{total_steps} | Training Loss: {train_metric['train loss']}, Learning Rate: {train_metric['learning_rate']})"
                )

                if data_args.task_name is not None:
                    wandb.log(
                        {
                            f"{data_args.task_name} Step": cur_step,
                            f"{data_args.task_name} Train Loss": train_metric['train loss'],
                            f"{data_args.task_name} Train Accuracy": train_metric['train accuracy'],
                            f'{data_args.task_name} Learning Rate': train_metric['learning_rate']
                        }
                    )
                else:
                    wandb.log(
                        {
                            "Step": cur_step,
                            "Train Loss": train_metric['train loss'],
                            "Train Accuracy": train_metric['train accuracy'],
                            'Learning Rate': train_metric['learning_rate']
                        }
                    )

                train_metrics = []

            if (cur_step % training_args.eval_steps == 0 or cur_step % steps_per_epoch == 0) and cur_step > 0:

                # evaluate
                eval_loader = glue_eval_data_collator(eval_dataset, eval_batch_size)
                eval_additional_metrics = []
                for batch in tqdm(
                        eval_loader,
                        total=len(eval_dataset) // eval_batch_size,
                        desc="Evaluating ...",
                        position=2,
                ):
                    labels = batch["labels"]
                    predictions, eval_additional_metric = p_eval_step(state, batch)
                    metric.add_batch(predictions=chain(*predictions), references=chain(*labels))
                    eval_additional_metrics.append(eval_additional_metric)

                # evaluate also on leftover examples (not divisible by batch_size)
                num_leftover_samples = len(eval_dataset) % eval_batch_size

                # make sure leftover batch is evaluated on one device
                if num_leftover_samples > 0 and jax.process_index() == 0:
                    # take leftover samples
                    batch = eval_dataset[-num_leftover_samples:]
                    batch = {k: np.array(v) for k, v in batch.items()}
                    master_state = unreplicate(state)
                    labels = batch.pop("labels")
                    logits = master_state.apply_fn(**batch, params=master_state.params, train=False)[0]
                    predictions = master_state.logits_fn(logits)
                    metric.add_batch(predictions=predictions, references=labels)

                eval_metric = metric.compute()

                logger.info(f"Step... ({cur_step}/{total_steps} | Eval metrics: {eval_metric})")

                if data_args.task_name is None:
                    dict_eval_metric = {f"eval_{metric_name}": value for metric_name, value in
                                        eval_metric.items()}
                else:
                    dict_eval_metric = {f"{data_args.task_name} eval_{metric_name}": value for metric_name, value in
                                        eval_metric.items()}

                eval_additional_metrics = get_metrics(eval_additional_metrics)
                print(eval_additional_metrics)

                dict_eval_metric["Test Loss"] = jnp.mean(eval_additional_metrics['test loss'])
                dict_eval_metric["Test Accuracy"] = jnp.mean(eval_additional_metrics['test accuracy'])
                wandb.log(dict_eval_metric)

                if has_tensorboard and jax.process_index() == 0:
                    write_eval_metric(summary_writer, eval_metric, cur_step)

            if (cur_step % training_args.save_steps == 0 and cur_step > 0) or (cur_step == total_steps):
                # save checkpoint after each epoch and push checkpoint to the hub
                if jax.process_index() == 0:
                    params = jax.device_get(unreplicate(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)
            epochs.desc = f"Epoch ... {epoch + 1}/{num_epochs}"

    # save the eval metrics in json
    if jax.process_index() == 0:
        eval_metric = {f"eval_{metric_name}": value for metric_name, value in eval_metric.items()}
        path = os.path.join(training_args.output_dir, "eval_results.json")
        with open(path, "w") as f:
            json.dump(eval_metric, f, indent=4, sort_keys=True)

    if training_args.do_predict:
        logger.info("*** Predict ***")

        # Loop to handle MNLI double evaluation (matched, mis-matched)
        tasks = [data_args.task_name]
        predict_datasets = [predict_dataset]
        if data_args.task_name == "mnli":
            tasks.append("mnli-mm")
            predict_datasets.append(raw_datasets["test_mismatched"])

        for predict_dataset, task in zip(predict_datasets, tasks):
            # Removing the `label` columns because it contains -1 and Trainer won't like that.
            predict_dataset = predict_dataset.remove_columns("label")
            predict_loader = glue_eval_data_collator(predict_dataset, eval_batch_size)
            predictions = []
            for batch in tqdm(
                    predict_loader,
                    total=len(predict_dataset) // eval_batch_size,
                    desc="Predict ...",
                    position=2,
            ):
                labels = batch.pop("labels")
                batch_predictions = p_eval_step(state, batch)
                predictions = jnp.vstack([predictions, batch_predictions]) if predictions.size else batch_predictions

                # predictions = trainer.predict(predict_dataset, metric_key_prefix="predict").predictions
            predictions = np.squeeze(predictions) if is_regression else np.argmax(predictions, axis=1)

            output_predict_file = os.path.join(training_args.output_dir, f"predict_results_{task}.txt")
            if jax.process_index() == 0:
                with open(output_predict_file, "w") as writer:
                    logger.info(f"***** Predict results {task} *****")
                    writer.write("index\tprediction\n")
                    for index, item in enumerate(predictions):
                        if is_regression:
                            writer.write(f"{index}\t{item:3.3f}\n")
                        else:
                            item = label_list[item]
                            writer.write(f"{index}\t{item}\n")


if __name__ == "__main__":
    main()