statement

    2024-04-22 简化
    2023-10-19 微调完成
    2023-10-18 初始化

update

• deep_training

install

• pip install -U -r requirements.txt
• 如果无法安装 , 可以切换官方源 pip install -i https://pypi.org/simple -U -r requirements.txt

# flash-attention对显卡算例要求算力7.5 以上 ， 下面可选安装 ，如果卡不支持可以不安装。
git clone -b v1.0.8 https://github.com/Dao-AILab/flash-attention
cd flash-attention && pip install .
pip install csrc/layer_norm
pip install csrc/rotary

weight

• Qwen-VL
• Qwen-VL-Chat

data sample

open_data https://github.com/ssbuild/open_data

单条数据示例

p prefix  optional
q question optional
a answer   must

 {
    "id": 0, 
    "p": null,
    "paragraph": [
         {
             "q": "Picture 1: <img>https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg</img>\n图中的狗是什么品种？",
             "a":  "图中是一只拉布拉多犬。"
         },
        {
           "q": "框出图中的格子衬衫",
           "a": "<ref>格子衬衫</ref><box>(588,499),(725,789)</box>"
        }
     ]
 }

或者

{
    "id": 0,
    "conversations": [
      {
        "from": "user",
        "value": "Picture 1: <img>https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg</img>\n图中的狗是什么品种？"
      },
      {
        "from": "assistant",
        "value": "图中是一只拉布拉多犬。"
      },
      {
        "from": "user",
        "value": "框出图中的格子衬衫"
      },
      {
        "from": "assistant",
        "value": "<ref>格子衬衫</ref><box>(588,499),(725,789)</box>"
      }
    ]
  }

infer

# infer.py 推理预训练模型
# infer_finetuning.py 推理微调模型
# infer_lora_finetuning.py 推理lora微调模型
 python infer.py

量化等级	最低 GPU 显存
FP16（无量化）	13 GB
INT8	10 GB
INT4	6 GB

training

    # 制作数据
    cd scripts
    bash train_full.sh -m dataset 
    or
    bash train_lora.sh -m dataset 
    or
    bash train_ptv2.sh -m dataset 
    
    注: num_process_worker 为多进程制作数据 ， 如果数据量较大 ， 适当调大至cpu数量
    dataHelper.make_dataset_with_args(data_args.train_file,mixed_data=False, shuffle=True,mode='train',num_process_worker=0)
    
    # 全参数训练 
        bash train_full.sh -m train
        
    # lora adalora ia3 
        bash train_lora.sh -m train
        
    # ptv2
        bash train_ptv2.sh -m train

训练参数

训练参数

友情链接

• pytorch-task-example
• chatmoss_finetuning
• chatglm_finetuning
• chatglm2_finetuning
• t5_finetuning
• llm_finetuning
• llm_rlhf
• chatglm_rlhf
• t5_rlhf
• rwkv_finetuning
• baichuan_finetuning
• baichuan2_finetuning
• internlm_finetuning
• qwen_finetuning

纯粹而干净的代码

Reference

https://github.com/QwenLM/Qwen-VL

Star History