internlm2_20b_chat_lora_alpaca_e3.py

# Copyright (c) OpenMMLab. All rights reserved.
import torch
from datasets import load_dataset
from mmengine.dataset import DefaultSampler
from mmengine.hooks import (CheckpointHook, DistSamplerSeedHook, IterTimerHook,
                            LoggerHook, ParamSchedulerHook)
from mmengine.optim import AmpOptimWrapper, CosineAnnealingLR, LinearLR
from peft import LoraConfig
from torch.optim import AdamW
from transformers import (AutoModelForCausalLM, AutoTokenizer,
                          BitsAndBytesConfig)

from xtuner.dataset import process_hf_dataset
from xtuner.dataset import ConcatDataset, process_hf_dataset
from xtuner.dataset.collate_fns import default_collate_fn
from xtuner.dataset.map_fns import alpaca_map_fn, template_map_fn_factory
from xtuner.engine.hooks import (DatasetInfoHook, EvaluateChatHook,
                                 VarlenAttnArgsToMessageHubHook)
from xtuner.engine.runner import TrainLoop
from xtuner.model import SupervisedFinetune
from xtuner.utils import PROMPT_TEMPLATE, SYSTEM_TEMPLATE
from mmengine.visualization import Visualizer,WandbVisBackend, TensorboardVisBackend
#######################################################################
#                          PART 1  Settings                           #
#######################################################################
# Model
#存储模型的路径
pretrained_model_name_or_path = '/root/share/model_repos/internlm2-chat-20b'
use_varlen_attn = False

# Data
#alpaca_en_path = 'tatsu-lab/alpaca'
#两个数据集，存储数据集的路径
data_path1 = '../datasets/data.json'
data_path2 = '../datasets/result.json'

prompt_template = PROMPT_TEMPLATE.internlm2_chat
max_length = 3096
pack_to_max_length = True

# Scheduler & Optimizer
batch_size =  4 # per_device
accumulative_counts = 4
dataloader_num_workers = 10
max_epochs = 3
optim_type = AdamW
lr = 3e-5
betas = (0.9, 0.999)
weight_decay = 0.001
max_norm = 1  # grad clip
warmup_ratio = 0.03

# Save
save_steps = 300
save_total_limit = 2  # Maximum checkpoints to keep (-1 means unlimited)

# Evaluate the generation performance during the training
evaluation_freq = 100
SYSTEM = "你是心理健康助手EmoLLM，由EmoLLM团队打造。你旨在通过专业心理咨询，协助来访者完成心理诊断。请充分利用专业心理学知识与咨询技术，一步步帮助来访者解决心理问题。"
evaluation_inputs = [
    '我最近总是感到很焦虑，尤其是在学业上。我有个特别崇拜的同学，他好像在各方面都比我优秀，我总觉得自己怎么努力也追不上他，这让我压力特别大。', '我知道应该理性看待，但就是忍不住会去比较。我甚至晚上会因为这个睡不着觉，总想着怎样才能像他那样出色。'
]

#######################################################################
#                      PART 2  Model & Tokenizer                      #
#######################################################################
tokenizer = dict(
    type=AutoTokenizer.from_pretrained,
    pretrained_model_name_or_path=pretrained_model_name_or_path,
    trust_remote_code=True,
    padding_side='right')

model = dict(
    type=SupervisedFinetune,
    use_varlen_attn=use_varlen_attn,
    llm=dict(
        type=AutoModelForCausalLM.from_pretrained,
        pretrained_model_name_or_path=pretrained_model_name_or_path,
        trust_remote_code=True,
        torch_dtype=torch.float16),
    lora=dict(
        type=LoraConfig,
        r=64,
        lora_alpha=128,
        lora_dropout=0.1,
        bias='none',
        task_type='CAUSAL_LM'))

#######################################################################
#                      PART 3  Dataset & Dataloader                   #
#######################################################################
data1 = dict(
    type=process_hf_dataset,
    dataset=dict(type=load_dataset, path='json', data_files=dict(train=data_path1)),
    tokenizer=tokenizer,
    max_length=max_length,
    dataset_map_fn=None,
    template_map_fn=dict(
        type=template_map_fn_factory, template=prompt_template),
    remove_unused_columns=True,
    shuffle_before_pack=True,
    pack_to_max_length=pack_to_max_length,
    use_varlen_attn=use_varlen_attn)

data2 = dict(
    type=process_hf_dataset,
    dataset=dict(type=load_dataset, path='json', data_files=dict(train=data_path2)),
    tokenizer=tokenizer,
    max_length=max_length,
    dataset_map_fn=None,
    template_map_fn=dict(
        type=template_map_fn_factory, template=prompt_template),
    remove_unused_columns=True,
    shuffle_before_pack=True,
    pack_to_max_length=pack_to_max_length,
    use_varlen_attn=use_varlen_attn)


train_dataset = dict(
    type=ConcatDataset, datasets=[data1, data2])

train_dataloader = dict(
    batch_size=batch_size,
    num_workers=dataloader_num_workers,
    dataset=train_dataset,
    sampler=dict(type=DefaultSampler, shuffle=True),
    collate_fn=dict(type=default_collate_fn, use_varlen_attn=use_varlen_attn))

#######################################################################
#                    PART 4  Scheduler & Optimizer                    #
#######################################################################
# optimizer
optim_wrapper = dict(
    type=AmpOptimWrapper,
    optimizer=dict(
        type=optim_type, lr=lr, betas=betas, weight_decay=weight_decay),
    clip_grad=dict(max_norm=max_norm, error_if_nonfinite=False),
    accumulative_counts=accumulative_counts,
    loss_scale='dynamic',
    dtype='float16')

# learning policy
# More information: https://github.com/open-mmlab/mmengine/blob/main/docs/en/tutorials/param_scheduler.md  # noqa: E501
param_scheduler = [
    dict(
        type=LinearLR,
        start_factor=1e-5,
        by_epoch=True,
        begin=0,
        end=warmup_ratio * max_epochs,
        convert_to_iter_based=True),
    dict(
        type=CosineAnnealingLR,
        eta_min=0.0,
        by_epoch=True,
        begin=warmup_ratio * max_epochs,
        end=max_epochs,
        convert_to_iter_based=True)
]

# train, val, test setting
train_cfg = dict(type=TrainLoop, max_epochs=max_epochs)

#######################################################################
#                           PART 5  Runtime                           #
#######################################################################
# Log the dialogue periodically during the training process, optional
custom_hooks = [
    dict(type=DatasetInfoHook, tokenizer=tokenizer),
    dict(
        type=EvaluateChatHook,
        tokenizer=tokenizer,
        every_n_iters=evaluation_freq,
        evaluation_inputs=evaluation_inputs,
        system=SYSTEM,
        prompt_template=prompt_template)
]

if use_varlen_attn:
    custom_hooks += [dict(type=VarlenAttnArgsToMessageHubHook)]

# configure default hooks
default_hooks = dict(
    # record the time of every iteration.
    timer=dict(type=IterTimerHook),
    # print log every 10 iterations.
    logger=dict(type=LoggerHook, log_metric_by_epoch=False, interval=10),
    # enable the parameter scheduler.
    param_scheduler=dict(type=ParamSchedulerHook),
    # save checkpoint per `save_steps`.
    checkpoint=dict(
        type=CheckpointHook,
        by_epoch=False,
        interval=save_steps,
        max_keep_ckpts=save_total_limit),
    # set sampler seed in distributed evrionment.
    sampler_seed=dict(type=DistSamplerSeedHook),
)

# configure environment
env_cfg = dict(
    # whether to enable cudnn benchmark
    cudnn_benchmark=False,
    # set multi process parameters
    mp_cfg=dict(mp_start_method='fork', opencv_num_threads=0),
    # set distributed parameters
    dist_cfg=dict(backend='nccl'),
)

# set visualizer
visualizer = dict(
    type=Visualizer, 
    vis_backends=[dict(type=WandbVisBackend)]
)


# set log level
log_level = 'INFO'

# load from which checkpoint
load_from = None

# whether to resume training from the loaded checkpoint
resume = False

# Defaults to use random seed and disable `deterministic`
randomness = dict(seed=None, deterministic=False)

# set log processor
log_processor = dict(by_epoch=False)