tests/collections/llm/megatron_t5_pretraining.py

# Copyright (c) 2025, NVIDIA CORPORATION.  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.

## NOTE: This script is present for github-actions testing only.
## There are no guarantees that this script is up-to-date with latest NeMo.

import argparse

import torch
from lightning.pytorch.loggers import WandbLogger
from megatron.core.optimizer import OptimizerConfig

from nemo import lightning as nl
from nemo.collections import llm
from nemo.collections.llm.api import pretrain
from nemo.collections.llm.t5.data import MockDataModule, PreTrainingDataModule
from nemo.collections.nlp.modules.common.tokenizer_utils import get_nmt_tokenizer
from nemo.lightning import NeMoLogger
from nemo.lightning.pytorch.callbacks import ModelCheckpoint
from nemo.lightning.pytorch.optim.lr_scheduler import WarmupAnnealingScheduler
from nemo.lightning.pytorch.optim.megatron import MegatronOptimizerModule
from tests.collections.llm.common import AssertOptimizerParamGroupsHaveAtLeastTwoWeightDecays


def get_args():
    parser = argparse.ArgumentParser(description='Train a small T5 model using NeMo 2.0')
    parser.add_argument('--devices', type=int, help="Number of devices to use for training")
    parser.add_argument('--max-steps', type=int, help="Number of steps to train for")
    parser.add_argument('--experiment-dir', type=str, help="directory to write results and checkpoints to")
    parser.add_argument('--experiment-name', type=str, help="name of experiment")
    parser.add_argument('--wandb-project', type=str, default=None, help="wandb project name")
    parser.add_argument('--data-path', type=str, help="Path to data file")
    parser.add_argument('--vocab-path', type=str, default=None, help="Path to vocab file")
    parser.add_argument('--index-mapping-dir', type=str, help="directory to write index mappings to")

    return parser.parse_args()


if __name__ == '__main__':

    args = get_args()

    special_tokens = {}
    special_tokens['additional_special_tokens'] = [f'<extra_id_{i}>' for i in range(100)]
    tokenizer = get_nmt_tokenizer(
        "megatron",
        "BertWordPieceCase",
        vocab_file=args.vocab_path,
        special_tokens=special_tokens,
    )
    if args.data_path == "mock":
        data = MockDataModule(
            seq_length=512,
            seq_length_dec=128,
            micro_batch_size=64,
            global_batch_size=512,
            tokenizer=tokenizer,
            num_train_samples=10_000,
            num_val_samples=10_000,
            num_test_samples=10_000,
        )
    else:
        data = PreTrainingDataModule(
            paths=args.data_path,
            seq_length=512,
            seq_length_dec=128,
            micro_batch_size=64,
            global_batch_size=512,
            seed=1234,
            tokenizer=tokenizer,
            split="99982,9,9",
            index_mapping_dir=args.index_mapping_dir,
        )

    t5_config = llm.t5.model.t5.T5Config(
        num_layers=12,
        encoder_num_layers=12,
        hidden_size=768,
        ffn_hidden_size=3072,
        num_attention_heads=12,
        kv_channels=64,
        init_method_std=0.015,
        hidden_dropout=0.1,
        attention_dropout=0.1,
        layernorm_epsilon=1e-5,
        make_vocab_size_divisible_by=128,
        max_position_embeddings=512,
        bf16=True,
        params_dtype=torch.bfloat16,
        pipeline_dtype=torch.bfloat16,
    )
    model = llm.t5.model.t5.T5Model(t5_config, tokenizer=data.tokenizer)
    strategy = nl.MegatronStrategy(
        tensor_model_parallel_size=1,
        pipeline_model_parallel_size=1,
        pipeline_dtype=None,
    )
    checkpoint_callback = ModelCheckpoint(
        every_n_train_steps=5000,
        save_optim_on_train_end=True,
    )
    callbacks = [checkpoint_callback, AssertOptimizerParamGroupsHaveAtLeastTwoWeightDecays()]

    resume = nl.AutoResume(
        resume_if_exists=True,
        resume_ignore_no_checkpoint=True,
    )

    opt_config = OptimizerConfig(
        optimizer='adam',
        lr=0.0001,
        use_distributed_optimizer=False,
        bf16=True,
        weight_decay=0.01,
    )
    lr_scheduler = WarmupAnnealingScheduler(
        warmup_steps=None,
        warmup_ratio=0.01,
        max_steps=args.max_steps,
        min_lr=0.00001,
    )
    opt = MegatronOptimizerModule(
        config=opt_config,
        lr_scheduler=lr_scheduler,
    )

    trainer = nl.Trainer(
        devices=args.devices,
        max_steps=args.max_steps,
        accelerator="gpu",
        strategy=strategy,
        callbacks=callbacks,
        log_every_n_steps=1,
        limit_val_batches=2,
        val_check_interval=2,
        plugins=nl.MegatronMixedPrecision(precision="bf16-mixed"),
    )

    if args.wandb_project is not None:
        wandb_logger = WandbLogger(
            name=args.experiment_name,
            project=args.wandb_project,
            log_model="all",
        )
    else:
        wandb_logger = None
    nemo_logger = NeMoLogger(
        name=args.experiment_name,
        use_datetime_version=False,
        log_dir=args.experiment_dir,
        wandb=wandb_logger,
    )

    pretrain(
        model=model,
        resume=resume,
        data=data,
        trainer=trainer,
        log=nemo_logger,
        optim=opt,
    )