`(xt) root@autodl-container-acc940bcfe-be30ce08:~/autodl-tmp/ft# cd config
(xt) root@autodl-container-acc940bcfe-be30ce08:~/autodl-tmp/ft/config# xtuner train internlm2_chat_7b_qlora_alpaca_e3_copy.py
[2024-05-30 18:04:29,376] [INFO] [real_accelerator.py:203:get_accelerator] Setting ds_accelerator to cuda (auto detect)
05/30 18:04:29 - mmengine - WARNING - WARNING: command error: '[Errno 2] No such file or directory: '/root/autodl-tmp/ft/data/cutlass/CHANGELOG.md''!
05/30 18:04:29 - mmengine - WARNING -
Arguments received: ['xtuner', 'train', 'internlm2_chat_7b_qlora_alpaca_e3_copy.py']. xtuner commands use the following syntax:
xtuner MODE MODE_ARGS ARGS
Where MODE (required) is one of ('list-cfg', 'copy-cfg', 'log-dataset', 'check-custom-dataset', 'train', 'test', 'chat', 'convert', 'preprocess', 'mmbench', 'eval_refcoco')
MODE_ARG (optional) is the argument for specific mode
ARGS (optional) are the arguments for specific command
Some usages for xtuner commands: (See more by using -h for specific command!)
1. List all predefined configs:
xtuner list-cfg
2. Copy a predefined config to a given path:
xtuner copy-cfg $CONFIG $SAVE_FILE
3-1. Fine-tune LLMs by a single GPU:
xtuner train $CONFIG
3-2. Fine-tune LLMs by multiple GPUs:
NPROC_PER_NODE=$NGPUS NNODES=$NNODES NODE_RANK=$NODE_RANK PORT=$PORT ADDR=$ADDR xtuner dist_train $CONFIG $GPUS
4-1. Convert the pth model to HuggingFace's model:
xtuner convert pth_to_hf $CONFIG $PATH_TO_PTH_MODEL $SAVE_PATH_TO_HF_MODEL
4-2. Merge the HuggingFace's adapter to the pretrained base model:
xtuner convert merge $LLM $ADAPTER $SAVE_PATH
xtuner convert merge $CLIP $ADAPTER $SAVE_PATH --is-clip
4-3. Split HuggingFace's LLM to the smallest sharded one:
xtuner convert split $LLM $SAVE_PATH
5-1. Chat with LLMs with HuggingFace's model and adapter:
xtuner chat $LLM --adapter $ADAPTER --prompt-template $PROMPT_TEMPLATE --system-template $SYSTEM_TEMPLATE
5-2. Chat with VLMs with HuggingFace's model and LLaVA:
xtuner chat $LLM --llava $LLAVA --visual-encoder $VISUAL_ENCODER --image $IMAGE --prompt-template $PROMPT_TEMPLATE --system-template $SYSTEM_TEMPLATE
6-1. Preprocess arxiv dataset:
xtuner preprocess arxiv $SRC_FILE $DST_FILE --start-date $START_DATE --categories $CATEGORIES
6-2. Preprocess refcoco dataset:
xtuner preprocess refcoco --ann-path $RefCOCO_ANN_PATH --image-path $COCO_IMAGE_PATH --save-path $SAVE_PATH
7-1. Log processed dataset:
xtuner log-dataset $CONFIG
7-2. Verify the correctness of the config file for the custom dataset:
xtuner check-custom-dataset $CONFIG
8. MMBench evaluation:
xtuner mmbench $LLM --llava $LLAVA --visual-encoder $VISUAL_ENCODER --prompt-template $PROMPT_TEMPLATE --data-path $MMBENCH_DATA_PATH
9. Refcoco evaluation:
xtuner eval_refcoco $LLM --llava $LLAVA --visual-encoder $VISUAL_ENCODER --prompt-template $PROMPT_TEMPLATE --data-path $REFCOCO_DATA_PATH
10. List all dataset formats which are supported in XTuner
Run special commands:
xtuner help
xtuner version
GitHub: https://github.com/InternLM/xtuner`
config文件：
`# 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 openai_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
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
`(xt) root@autodl-container-acc940bcfe-be30ce08:~/autodl-tmp/ft# cd config (xt) root@autodl-container-acc940bcfe-be30ce08:~/autodl-tmp/ft/config# xtuner train internlm2_chat_7b_qlora_alpaca_e3_copy.py [2024-05-30 18:04:29,376] [INFO] [real_accelerator.py:203:get_accelerator] Setting ds_accelerator to cuda (auto detect) 05/30 18:04:29 - mmengine - WARNING - WARNING: command error: '[Errno 2] No such file or directory: '/root/autodl-tmp/ft/data/cutlass/CHANGELOG.md''! 05/30 18:04:29 - mmengine - WARNING - Arguments received: ['xtuner', 'train', 'internlm2_chat_7b_qlora_alpaca_e3_copy.py']. xtuner commands use the following syntax:
config文件： `# 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 openai_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 = './autodl-tmp/RAG-langchain/models/internlm2-chat-7b' use_varlen_attn = False
Data
alpaca_en_path = './autodl-tmp/ft/data/train_fold_1.json' prompt_template = PROMPT_TEMPLATE.internlm2_chat max_length = 1024 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 = 2 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 = 300 save_total_limit = 3 # Maximum checkpoints to keep (-1 means unlimited)
Evaluate the generation performance during the training
evaluation_freq = 300 SYSTEM = '' evaluation_inputs = ['判断以下新闻情绪，积极为1，消极为0', '判断该新闻正负面，正面为1，负面为0', '请判断以下新闻是积极还是消极，积极为1，消极为0，你的答案只有1或0']
#######################################################################
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='json', data_files=dict(train=alpaca_en_path)), tokenizer=tokenizer, max_length=max_length, dataset_map_fn=openai_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.
)
configure environment
env_cfg = dict(
whether to enable cudnn benchmark
)
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
deterministicrandomness = dict(seed=None, deterministic=False)
set log processor
log_processor = dict(by_epoch=False)`