【大模型】fineturn Q-wen

github上下载qwen1_5源码

修改finetun.sh

然后在路径qwen1_5/examples/sft下修改finetun.sh, 内容如下

#!/bin/bash
export CUDA_DEVICE_MAX_CONNECTIONS=1
DIR=`pwd`

# Guide:
# This script supports distributed training on multi-gpu workers (as well as single-worker training).
# Please set the options below according to the comments.
# For multi-gpu workers training, these options should be manually set for each worker.
# After setting the options, please run the script on each worker.

# Number of GPUs per GPU worker
GPUS_PER_NODE=$(python -c 'import torch; print(torch.cuda.device_count())')

# Number of GPU workers, for single-worker training, please set to 1
NNODES=${NNODES:-1}

# The rank of this worker, should be in {0, ..., WORKER_CNT-1}, for single-worker training, please set to 0
NODE_RANK=${NODE_RANK:-0}

# The ip address of the rank-0 worker, for single-worker training, please set to localhost
MASTER_ADDR=${MASTER_ADDR:-localhost}

# The port for communication
MASTER_PORT=${MASTER_PORT:-6010}

MODEL="Qwen/Qwen1.5-7B" # Set the path if you do not want to load from huggingface directly
# ATTENTION: specify the path to your training data, which should be a json file consisting of a list of conversations.
# See the section for finetuning in README for more information.
DATA="path_to_data"
DS_CONFIG_PATH="finetune/ds_config_zero3.json"
USE_LORA=False
Q_LORA=False

function usage() {
    echo '
Usage: bash finetune/finetune_lora_ds.sh [-m MODEL_PATH] [-d DATA_PATH] [--deepspeed DS_CONFIG_PATH] [--use_lora USE_LORA] [--q_lora Q_LORA]
'
}

while [[ "$1" != "" ]]; do
    case $1 in
        -m | --model )
            shift
            MODEL=$1
            ;;
        -d | --data )
            shift
            DATA=$1
            ;;
        --deepspeed )
            shift
            DS_CONFIG_PATH=$1
            ;;
        --use_lora  )
            shift
            USE_LORA=$1
            ;;
        --q_lora    )
            shift
            Q_LORA=$1
            ;;
        -h | --help )
            usage
            exit 0
            ;;
        * )
            echo "Unknown argument ${1}"
            exit 1
            ;;
    esac
    shift
done

DISTRIBUTED_ARGS="
    --nproc_per_node $GPUS_PER_NODE \
    --nnodes $NNODES \
    --node_rank $NODE_RANK \
    --master_addr $MASTER_ADDR \
    --master_port $MASTER_PORT
"

torchrun $DISTRIBUTED_ARGS finetune.py \
    --model_name_or_path $MODEL \
    --data_path $DATA \
    --bf16 True \
    --output_dir output_qwen \
    --num_train_epochs 5 \
    --per_device_train_batch_size 2 \
    --per_device_eval_batch_size 1 \
    --gradient_accumulation_steps 8 \
    --evaluation_strategy "no" \
    --save_strategy "steps" \
    --save_steps 10 \
    --save_total_limit 10 \
    --learning_rate 3e-4 \
    --weight_decay 0.01 \
    --adam_beta2 0.95 \
    --warmup_ratio 0.01 \
    --lr_scheduler_type "cosine" \
    --logging_steps 1 \
    --report_to "none" \
    --model_max_length 512 \
    --lazy_preprocess True \
    --use_lora ${USE_LORA} \
    --q_lora ${Q_LORA} \
    --gradient_checkpointing \
    --deepspeed ${DS_CONFIG_PATH}

训练

（在qwen1_5/examples/sft路径下开个bash里运行finetune.sh，不要在jupyter里跑）

pip install transformers==4.37.0

# 要用命令行运行
# 不想用多卡训练的时候，先 export CUDA_VISIBLE_DEVICE=0
bash finetune.sh -m "/opt/app-root/src/Qwen1.5-14B-Chat" -d "./data/traindata.jsonl" --deepspeed "ds_config_zero3.json" --use_lora True

预测

（在qwen1_5/examples/sft路径下建个inference.py）

pip install transformers==4.33.0

from transformers import AutoModelForCausalLM, AutoTokenizer
import os
os.environ["CUDA_VISIBLE_DEVICES"] = "0"
device = "cuda" # the device to load the model onto
path = "output_qwen/checkpoint-70"
model = AutoModelForCausalLM.from_pretrained(
    path,
    torch_dtype="auto",
    device_map="cuda:0"
)
tokenizer = AutoTokenizer.from_pretrained(path)

    
def predict_answer(messages):
    text = tokenizer.apply_chat_template(
        messages,
        tokenize=False,
        add_generation_prompt=True
    )
    model_inputs = tokenizer([text], return_tensors="pt").to(device)

    generated_ids = model.generate(
        model_inputs.input_ids,
        max_new_tokens=512,
    )
    generated_ids = [
        output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
    ]

    response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
    return response
    

text = "xxxx"
messages = [{"role": "user", "content": "我需要起草投标文件中的一个章节，章节内容为：\n\n\n{}\n\n\n\n请将章节内容拆分成多个小节，每个小节覆盖一个信息点，形成一份本章节的提纲。注意，要覆盖所有信息点，不要使用‘同上、略’等省略表述，尽可能保持原文的措词。".format(text)}]
response = predict_answer(messages)
print(response)

训练数据格式

格式为jsonl，每行一条json，位于qwen1_5/examples/sft/data下，不妨命名为traindata.jsonl

{"type": "chatml", "messages": [{"role": "user", "content": "PROMPT"}, {"role": "assistant", "content": "ANSWER"}], "source": "self-made"}