From 44749c2adf214add084f01514ec2ce463c506e8a Mon Sep 17 00:00:00 2001 From: whcao <41630003+HIT-cwh@users.noreply.github.com> Date: Wed, 3 Jul 2024 20:05:13 +0800 Subject: [PATCH] Support internlm2.5 (#803) * add internlm 2.5 configs * update readme --- README.md | 31 +-- README_zh-CN.md | 29 +-- ...chat_7b_full_finetune_custom_dataset_e1.py | 226 ++++++++++++++++++ .../internlm2_5_chat_7b_qlora_alpaca_e3.py | 219 +++++++++++++++++ .../internlm2_5_chat_7b_qlora_oasst1_e3.py | 219 +++++++++++++++++ 5 files changed, 688 insertions(+), 36 deletions(-) create mode 100644 xtuner/configs/internlm/internlm2_5_chat_7b/internlm2_5_chat_7b_full_finetune_custom_dataset_e1.py create mode 100644 xtuner/configs/internlm/internlm2_5_chat_7b/internlm2_5_chat_7b_qlora_alpaca_e3.py create mode 100644 xtuner/configs/internlm/internlm2_5_chat_7b/internlm2_5_chat_7b_qlora_oasst1_e3.py diff --git a/README.md b/README.md index e0be695ef..4e729226c 100644 --- a/README.md +++ b/README.md @@ -39,6 +39,8 @@ English | [简体中文](README_zh-CN.md) ## 🎉 News +- **\[2024/07\]** Support [InternLM 2.5](xtuner/configs/internlm/internlm2_5_chat_7b/) models! +- **\[2024/06\]** Support [DeepSeek V2](xtuner/configs/deepseek/deepseek_v2_chat/) models! **2x faster!** - **\[2024/04\]** [LLaVA-Phi-3-mini](https://huggingface.co/xtuner/llava-phi-3-mini-hf) is released! Click [here](xtuner/configs/llava/phi3_mini_4k_instruct_clip_vit_large_p14_336) for details! - **\[2024/04\]** [LLaVA-Llama-3-8B](https://huggingface.co/xtuner/llava-llama-3-8b) and [LLaVA-Llama-3-8B-v1.1](https://huggingface.co/xtuner/llava-llama-3-8b-v1_1) are released! Click [here](xtuner/configs/llava/llama3_8b_instruct_clip_vit_large_p14_336) for details! - **\[2024/04\]** Support [Llama 3](xtuner/configs/llama) models! @@ -100,16 +102,15 @@ XTuner is an efficient, flexible and full-featured toolkit for fine-tuning large @@ -203,14 +204,14 @@ XTuner supports the efficient fine-tune (*e.g.*, QLoRA) for LLMs. Dataset prepar xtuner train ${CONFIG_NAME_OR_PATH} ``` - For example, we can start the QLoRA fine-tuning of InternLM2-Chat-7B with oasst1 dataset by + For example, we can start the QLoRA fine-tuning of InternLM2.5-Chat-7B with oasst1 dataset by ```shell # On a single GPU - xtuner train internlm2_chat_7b_qlora_oasst1_e3 --deepspeed deepspeed_zero2 + xtuner train internlm2_5_chat_7b_qlora_oasst1_e3 --deepspeed deepspeed_zero2 # On multiple GPUs - (DIST) NPROC_PER_NODE=${GPU_NUM} xtuner train internlm2_chat_7b_qlora_oasst1_e3 --deepspeed deepspeed_zero2 - (SLURM) srun ${SRUN_ARGS} xtuner train internlm2_chat_7b_qlora_oasst1_e3 --launcher slurm --deepspeed deepspeed_zero2 + (DIST) NPROC_PER_NODE=${GPU_NUM} xtuner train internlm2_5_chat_7b_qlora_oasst1_e3 --deepspeed deepspeed_zero2 + (SLURM) srun ${SRUN_ARGS} xtuner train internlm2_5_chat_7b_qlora_oasst1_e3 --launcher slurm --deepspeed deepspeed_zero2 ``` - `--deepspeed` means using [DeepSpeed](https://github.com/microsoft/DeepSpeed) 🚀 to optimize the training. XTuner comes with several integrated strategies including ZeRO-1, ZeRO-2, and ZeRO-3. If you wish to disable this feature, simply remove this argument. @@ -231,18 +232,10 @@ XTuner provides tools to chat with pretrained / fine-tuned LLMs. xtuner chat ${NAME_OR_PATH_TO_LLM} --adapter {NAME_OR_PATH_TO_ADAPTER} [optional arguments] ``` -For example, we can start the chat with - -InternLM2-Chat-7B with adapter trained from oasst1 dataset: - -```shell -xtuner chat internlm/internlm2-chat-7b --adapter xtuner/internlm2-chat-7b-qlora-oasst1 --prompt-template internlm2_chat -``` - -LLaVA-InternLM2-7B: +For example, we can start the chat with InternLM2.5-Chat-7B : ```shell -xtuner chat internlm/internlm2-chat-7b --visual-encoder openai/clip-vit-large-patch14-336 --llava xtuner/llava-internlm2-7b --prompt-template internlm2_chat --image $IMAGE_PATH +xtuner chat internlm/internlm2_5-chat-7b --prompt-template internlm2_chat ``` For more examples, please see [chat.md](./docs/en/user_guides/chat.md). diff --git a/README_zh-CN.md b/README_zh-CN.md index c5037d28c..16c1a2af2 100644 --- a/README_zh-CN.md +++ b/README_zh-CN.md @@ -39,6 +39,8 @@ ## 🎉 更新 +- **\[2024/07\]** 支持 [InternLM 2.5](xtuner/configs/internlm/internlm2_5_chat_7b/) 模型! +- **\[2024/06\]** 支持 [DeepSeek V2](xtuner/configs/deepseek/deepseek_v2_chat/) models! **训练速度提升一倍!** - **\[2024/04\]** 多模态大模型 [LLaVA-Phi-3-mini](https://huggingface.co/xtuner/llava-phi-3-mini-hf) 发布!快速开始请查阅此[文档](xtuner/configs/llava/phi3_mini_4k_instruct_clip_vit_large_p14_336)! - **\[2024/04\]** 多模态大模型 [LLaVA-Llama-3-8B](https://huggingface.co/xtuner/llava-llama-3-8b) 和 [LLaVA-Llama-3-8B-v1.1](https://huggingface.co/xtuner/llava-llama-3-8b-v1_1) 发布!快速开始请查阅此[文档](xtuner/configs/llava/llama3_8b_instruct_clip_vit_large_p14_336)! - **\[2024/04\]** 支持 [Llama 3](xtuner/configs/llama) 模型! @@ -100,16 +102,15 @@ XTuner 是一个高效、灵活、全能的轻量化大模型微调工具库。 @@ -203,14 +204,14 @@ XTuner 支持微调大语言模型。数据集预处理指南请查阅[文档](. xtuner train ${CONFIG_NAME_OR_PATH} ``` - 例如,我们可以利用 QLoRA 算法在 oasst1 数据集上微调 InternLM2-Chat-7B: + 例如,我们可以利用 QLoRA 算法在 oasst1 数据集上微调 InternLM2.5-Chat-7B: ```shell # 单卡 - xtuner train internlm2_chat_7b_qlora_oasst1_e3 --deepspeed deepspeed_zero2 + xtuner train internlm2_5_chat_7b_qlora_oasst1_e3 --deepspeed deepspeed_zero2 # 多卡 - (DIST) NPROC_PER_NODE=${GPU_NUM} xtuner train internlm2_chat_7b_qlora_oasst1_e3 --deepspeed deepspeed_zero2 - (SLURM) srun ${SRUN_ARGS} xtuner train internlm2_chat_7b_qlora_oasst1_e3 --launcher slurm --deepspeed deepspeed_zero2 + (DIST) NPROC_PER_NODE=${GPU_NUM} xtuner train internlm2_5_chat_7b_qlora_oasst1_e3 --deepspeed deepspeed_zero2 + (SLURM) srun ${SRUN_ARGS} xtuner train internlm2_5_chat_7b_qlora_oasst1_e3 --launcher slurm --deepspeed deepspeed_zero2 ``` - `--deepspeed` 表示使用 [DeepSpeed](https://github.com/microsoft/DeepSpeed) 🚀 来优化训练过程。XTuner 内置了多种策略,包括 ZeRO-1、ZeRO-2、ZeRO-3 等。如果用户期望关闭此功能,请直接移除此参数。 @@ -233,16 +234,10 @@ xtuner chat ${NAME_OR_PATH_TO_LLM} --adapter {NAME_OR_PATH_TO_ADAPTER} [optional 例如: -与 InternLM2-Chat-7B, oasst1 adapter 对话: +与 InternLM2.5-Chat-7B 对话: ```shell -xtuner chat internlm/internlm2-chat-7b --adapter xtuner/internlm2-chat-7b-qlora-oasst1 --prompt-template internlm2_chat -``` - -与 LLaVA-InternLM2-7B 对话: - -```shell -xtuner chat internlm/internlm2-chat-7b --visual-encoder openai/clip-vit-large-patch14-336 --llava xtuner/llava-internlm2-7b --prompt-template internlm2_chat --image $IMAGE_PATH +xtuner chat internlm/internlm2-chat-7b --prompt-template internlm2_chat ``` 更多示例,请查阅[文档](./docs/zh_cn/user_guides/chat.md)。 diff --git a/xtuner/configs/internlm/internlm2_5_chat_7b/internlm2_5_chat_7b_full_finetune_custom_dataset_e1.py b/xtuner/configs/internlm/internlm2_5_chat_7b/internlm2_5_chat_7b_full_finetune_custom_dataset_e1.py new file mode 100644 index 000000000..bc8a2816a --- /dev/null +++ b/xtuner/configs/internlm/internlm2_5_chat_7b/internlm2_5_chat_7b_full_finetune_custom_dataset_e1.py @@ -0,0 +1,226 @@ +# Copyright (c) OpenMMLab. All rights reserved. +"""Data format: +[ + { + "conversation": [ + { + "system": "", + "input": "xxx", + "output": "xxx" + }, + { + "input": "xxx", + "output": "xxx" + } + ] + }, +... +] +Please refer to https://github.com/InternLM/xtuner/blob/main/docs/en/user_guides/dataset_format.md for details. +""" # noqa: E501 +from datasets import load_dataset +from mmengine.hooks import (CheckpointHook, DistSamplerSeedHook, IterTimerHook, + LoggerHook, ParamSchedulerHook) +from mmengine.optim import AmpOptimWrapper, CosineAnnealingLR +from torch.optim import AdamW +from torch.utils.data import BatchSampler +from transformers import AutoModelForCausalLM, AutoTokenizer + +from xtuner.dataset import process_hf_dataset +from xtuner.dataset.collate_fns import default_collate_fn +from xtuner.dataset.map_fns import template_map_fn_factory +from xtuner.dataset.samplers import InternRepoSampler +from xtuner.engine import (DatasetInfoHook, EvaluateChatHook, ThroughputHook, + VarlenAttnArgsToMessageHubHook) +from xtuner.engine.runner import TrainLoop +from xtuner.model import SupervisedFinetune +from xtuner.utils import PROMPT_TEMPLATE + +####################################################################### +# PART 1 Settings # +####################################################################### +# Model +pretrained_model_name_or_path = 'internlm/internlm2_5-7b-chat' +use_varlen_attn = True + +# Data +data_files = ['/path/to/json/file.json'] +prompt_template = PROMPT_TEMPLATE.internlm2_chat +max_length = 32768 +pack_to_max_length = True + +# parallel +sequence_parallel_size = 1 + +# Scheduler & Optimizer +# batch size per device, set to 1 if `use_varlen_attn` = True +# To clarify, enlarging the batch size essentially enlarges the `max_length`. +# For example, doubling the max length is tantamount to doubling the batch size +batch_size = 1 +accumulative_counts = 1 # 1bs * 1acc * 64gpu = 64 batchsize +accumulative_counts *= sequence_parallel_size +dataloader_num_workers = 4 +max_epochs = 1 +optim_type = AdamW +lr = 4e-5 +betas = (0.9, 0.95) +weight_decay = 0.01 +max_norm = 1 # grad clip +warm_up_ratio = 0.025 + +# Save +save_steps = 500 +save_total_limit = 2 # Maximum checkpoints to keep (-1 means unlimited) + +# Evaluate the generation performance during the training +evaluation_freq = 500 +SYSTEM = '' +evaluation_inputs = [ + '请给我介绍五个上海的景点', 'Please tell me five scenic spots in Shanghai' +] + +####################################################################### +# 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)) + +####################################################################### +# PART 3 Dataset & Dataloader # +####################################################################### +train_dataset = dict( + type=process_hf_dataset, + use_varlen_attn=use_varlen_attn, + dataset=dict(type=load_dataset, path='json', data_files=data_files), + 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) + +train_dataloader = dict( + batch_size=batch_size, + num_workers=dataloader_num_workers, + dataset=train_dataset, + sampler=dict(type=InternRepoSampler, shuffle=True, seed=1024), + batch_sampler=dict( + type=BatchSampler, drop_last=True, batch_size=batch_size), + 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', +) + +# 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=1 / 40, + by_epoch=True, + begin=0, + end=warm_up_ratio * max_epochs, + convert_to_iter_based=True), + dict( + type=CosineAnnealingLR, + eta_min=lr * 0.15, + by_epoch=True, + begin=warm_up_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, + is_intern_repo_dataset=True), + dict( + type=EvaluateChatHook, + tokenizer=tokenizer, + every_n_iters=evaluation_freq, + evaluation_inputs=evaluation_inputs, + system=SYSTEM, + prompt_template=prompt_template), + dict(type=ThroughputHook) +] + +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 100 iterations. + logger=dict(type=LoggerHook, log_metric_by_epoch=False, interval=1), + # 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 = None + +# 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) + +log_processor = dict( + by_epoch=False, + window_size=1, + mean_pattern=r'.*(loss|time|data_time|grad_norm|tflops).*') diff --git a/xtuner/configs/internlm/internlm2_5_chat_7b/internlm2_5_chat_7b_qlora_alpaca_e3.py b/xtuner/configs/internlm/internlm2_5_chat_7b/internlm2_5_chat_7b_qlora_alpaca_e3.py new file mode 100644 index 000000000..7dfc92617 --- /dev/null +++ b/xtuner/configs/internlm/internlm2_5_chat_7b/internlm2_5_chat_7b_qlora_alpaca_e3.py @@ -0,0 +1,219 @@ +# 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.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.parallel.sequence import SequenceParallelSampler +from xtuner.utils import PROMPT_TEMPLATE, SYSTEM_TEMPLATE + +####################################################################### +# PART 1 Settings # +####################################################################### +# Model +pretrained_model_name_or_path = 'internlm/internlm2_5-7b-chat' +use_varlen_attn = False + +# Data +alpaca_en_path = 'tatsu-lab/alpaca' +prompt_template = PROMPT_TEMPLATE.internlm2_chat +max_length = 2048 +pack_to_max_length = True + +# parallel +sequence_parallel_size = 1 + +# Scheduler & Optimizer +batch_size = 1 # per_device +accumulative_counts = 1 +accumulative_counts *= sequence_parallel_size +dataloader_num_workers = 0 +max_epochs = 3 +optim_type = AdamW +lr = 2e-4 +betas = (0.9, 0.999) +weight_decay = 0 +max_norm = 1 # grad clip +warmup_ratio = 0.03 + +# Save +save_steps = 500 +save_total_limit = 2 # Maximum checkpoints to keep (-1 means unlimited) + +# Evaluate the generation performance during the training +evaluation_freq = 500 +SYSTEM = SYSTEM_TEMPLATE.alpaca +evaluation_inputs = [ + '请给我介绍五个上海的景点', 'Please tell me five scenic spots in Shanghai' +] + +####################################################################### +# 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, + quantization_config=dict( + type=BitsAndBytesConfig, + load_in_4bit=True, + load_in_8bit=False, + llm_int8_threshold=6.0, + llm_int8_has_fp16_weight=False, + bnb_4bit_compute_dtype=torch.float16, + bnb_4bit_use_double_quant=True, + bnb_4bit_quant_type='nf4')), + lora=dict( + type=LoraConfig, + r=64, + lora_alpha=16, + lora_dropout=0.1, + bias='none', + task_type='CAUSAL_LM')) + +####################################################################### +# PART 3 Dataset & Dataloader # +####################################################################### +alpaca_en = dict( + type=process_hf_dataset, + dataset=dict(type=load_dataset, path=alpaca_en_path), + tokenizer=tokenizer, + max_length=max_length, + dataset_map_fn=alpaca_map_fn, + 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) + +sampler = SequenceParallelSampler \ + if sequence_parallel_size > 1 else DefaultSampler +train_dataloader = dict( + batch_size=batch_size, + num_workers=dataloader_num_workers, + dataset=alpaca_en, + sampler=dict(type=sampler, 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 = None + +# 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) diff --git a/xtuner/configs/internlm/internlm2_5_chat_7b/internlm2_5_chat_7b_qlora_oasst1_e3.py b/xtuner/configs/internlm/internlm2_5_chat_7b/internlm2_5_chat_7b_qlora_oasst1_e3.py new file mode 100644 index 000000000..98b097efb --- /dev/null +++ b/xtuner/configs/internlm/internlm2_5_chat_7b/internlm2_5_chat_7b_qlora_oasst1_e3.py @@ -0,0 +1,219 @@ +# 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.collate_fns import default_collate_fn +from xtuner.dataset.map_fns import oasst1_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.parallel.sequence import SequenceParallelSampler +from xtuner.utils import PROMPT_TEMPLATE + +####################################################################### +# PART 1 Settings # +####################################################################### +# Model +pretrained_model_name_or_path = 'internlm/internlm2_5-7b-chat' +use_varlen_attn = False + +# Data +data_path = 'timdettmers/openassistant-guanaco' +prompt_template = PROMPT_TEMPLATE.internlm2_chat +max_length = 2048 +pack_to_max_length = True + +# parallel +sequence_parallel_size = 1 + +# Scheduler & Optimizer +batch_size = 1 # per_device +accumulative_counts = 16 +accumulative_counts *= sequence_parallel_size +dataloader_num_workers = 0 +max_epochs = 3 +optim_type = AdamW +lr = 2e-4 +betas = (0.9, 0.999) +weight_decay = 0 +max_norm = 1 # grad clip +warmup_ratio = 0.03 + +# Save +save_steps = 500 +save_total_limit = 2 # Maximum checkpoints to keep (-1 means unlimited) + +# Evaluate the generation performance during the training +evaluation_freq = 500 +SYSTEM = '' +evaluation_inputs = [ + '请给我介绍五个上海的景点', 'Please tell me five scenic spots in Shanghai' +] + +####################################################################### +# 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, + quantization_config=dict( + type=BitsAndBytesConfig, + load_in_4bit=True, + load_in_8bit=False, + llm_int8_threshold=6.0, + llm_int8_has_fp16_weight=False, + bnb_4bit_compute_dtype=torch.float16, + bnb_4bit_use_double_quant=True, + bnb_4bit_quant_type='nf4')), + lora=dict( + type=LoraConfig, + r=64, + lora_alpha=16, + lora_dropout=0.1, + bias='none', + task_type='CAUSAL_LM')) + +####################################################################### +# PART 3 Dataset & Dataloader # +####################################################################### +train_dataset = dict( + type=process_hf_dataset, + dataset=dict(type=load_dataset, path=data_path), + tokenizer=tokenizer, + max_length=max_length, + dataset_map_fn=oasst1_map_fn, + 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) + +sampler = SequenceParallelSampler \ + if sequence_parallel_size > 1 else DefaultSampler +train_dataloader = dict( + batch_size=batch_size, + num_workers=dataloader_num_workers, + dataset=train_dataset, + sampler=dict(type=sampler, 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 = None + +# 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)