qwen-vl最佳实践.md

# Qwen-VL 最佳实践

## 目录
- [环境准备](#环境准备)
- [推理](#推理)
- [微调](#微调)
- [微调后推理](#微调后推理)


## 环境准备
```shell
pip install 'ms-swift[llm]' -U
```

## 推理

推理[qwen-vl-chat](https://modelscope.cn/models/qwen/Qwen-VL-Chat/summary):
```shell
# Experimental environment: 3090
# 24GB GPU memory
CUDA_VISIBLE_DEVICES=0 swift infer --model_type qwen-vl-chat
```

输出: (支持传入本地路径或URL)
```python
"""
<<< multi-line
[INFO:swift] End multi-line input with `#`.
[INFO:swift] Input `single-line` to switch to single-line input mode.
<<<[M] 你是谁？#
我是通义千问，由阿里云开发的AI助手。我被设计用来回答各种问题、提供信息和与用户进行对话。有什么我可以帮助你的吗？
--------------------------------------------------
<<<[M] Picture 1:<img>http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/animal.png</img>
Picture 2:<img>http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/cat.png</img>
这两张图片有什么区别#
两张图片的相同点是它们都是关于动物的插画，但是它们的动物不同。
第一张图片中的动物是绵羊，而第二张图片中的动物是小猫。
--------------------------------------------------
<<<[M] Picture 1:<img>http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/animal.png</img>
图中有几只羊#
图中有四只羊。
--------------------------------------------------
<<<[M] Picture 1:<img>http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/math.png</img>
计算结果是多少#
1452 + 45304 = 46756
--------------------------------------------------
<<< clear
<<<[M] Picture 1:<img>http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/poem.png</img>
根据图片中的内容写首诗#
月光如水船如星，独坐船头吹夜风。深林倒影照水面，萤火点点照船行。
"""
```

示例图片如下:

cat:

<img src="http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/cat.png" width="250" style="display: inline-block;">

animal:

<img src="http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/animal.png" width="250" style="display: inline-block;">

math:

<img src="http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/math.png" width="250" style="display: inline-block;">

poem:

<img src="http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/poem.png" width="250" style="display: inline-block;">

**单样本推理**

```python
import os
os.environ['CUDA_VISIBLE_DEVICES'] = '0'

from swift.llm import (
    get_model_tokenizer, get_template, inference, ModelType,
    get_default_template_type, inference_stream
)
from swift.utils import seed_everything
import torch

model_type = ModelType.qwen_vl_chat
template_type = get_default_template_type(model_type)
print(f'template_type: {template_type}')

model, tokenizer = get_model_tokenizer(model_type, torch.float16,
                                       model_kwargs={'device_map': 'auto'})
model.generation_config.max_new_tokens = 256
template = get_template(template_type, tokenizer)
seed_everything(42)

query = """Picture 1:<img>http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/road.png</img>
距离各城市多远？"""
response, history = inference(model, template, query)
print(f'query: {query}')
print(f'response: {response}')

# 流式
query = '距离最远的城市是哪？'
gen = inference_stream(model, template, query, history)
print_idx = 0
print(f'query: {query}\nresponse: ', end='')
for response, history in gen:
    delta = response[print_idx:]
    print(delta, end='', flush=True)
    print_idx = len(response)
print()
print(f'history: {history}')
"""
query: Picture 1:<img>http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/road.png</img>
距离各城市多远？
response: 马路边距离马路边14公里；阳江边距离马路边62公里；广州边距离马路边293公里。
query: 距离最远的城市是哪？
response: 距离最远的城市是广州，距离马路边293公里。
history: [['Picture 1:<img>http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/road.png</img>\n距离各城市多远？', '马路边距离马路边14公里；阳江边距离马路边62公里；广州边距离马路边293公里。'], ['距离最远的城市是哪？', '距离最远的城市是广州，距离马路边293公里。']]
"""
```

示例图片如下:

road:

<img src="http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/road.png" width="250" style="display: inline-block;">


## 微调
多模态大模型微调通常使用**自定义数据集**进行微调. 这里展示可直接运行的demo:

LoRA微调:

(默认只对LLM部分的qkv进行lora微调. 如果你想对所有linear含vision模型部分都进行微调, 可以指定`--lora_target_modules ALL`)
```shell
# Experimental environment: 3090
# 23GB GPU memory
CUDA_VISIBLE_DEVICES=0 swift sft \
    --model_type qwen-vl-chat \
    --dataset coco-en-mini \
```

全参数微调:
```shell
# Experimental environment: 4 * A100
# 4 * 70 GPU memory
NPROC_PER_NODE=2 CUDA_VISIBLE_DEVICES=0,1,2,3 swift sft \
    --model_type qwen-vl-chat \
    --dataset coco-en-mini \
    --sft_type full \
```

[自定义数据集](../LLM/自定义与拓展.md#-推荐命令行参数的形式)支持json, jsonl样式, 以下是自定义数据集的例子:

(支持多轮对话, 支持每轮对话含多张图片或不含图片, 支持传入本地路径或URL)

```json
[
    {"conversations": [
        {"from": "user", "value": "Picture 1:<img>img_path</img>\n11111"},
        {"from": "assistant", "value": "22222"}
    ]},
    {"conversations": [
        {"from": "user", "value": "Picture 1:<img>img_path</img>\nPicture 2:<img>img_path2</img>\nPicture 3:<img>img_path3</img>\naaaaa"},
        {"from": "assistant", "value": "bbbbb"},
        {"from": "user", "value": "Picture 1:<img>img_path</img>\nccccc"},
        {"from": "assistant", "value": "ddddd"}
    ]},
    {"conversations": [
        {"from": "user", "value": "AAAAA"},
        {"from": "assistant", "value": "BBBBB"},
        {"from": "user", "value": "CCCCC"},
        {"from": "assistant", "value": "DDDDD"}
    ]}
]
```


## 微调后推理
直接推理:
```shell
CUDA_VISIBLE_DEVICES=0 swift infer \
    --ckpt_dir output/qwen-vl-chat/vx-xxx/checkpoint-xxx \
    --load_dataset_config true \
```

**merge-lora**并推理:
```shell
CUDA_VISIBLE_DEVICES=0 swift export \
    --ckpt_dir output/qwen-vl-chat/vx-xxx/checkpoint-xxx \
    --merge_lora true

CUDA_VISIBLE_DEVICES=0 swift infer \
    --ckpt_dir output/qwen-vl-chat/vx-xxx/checkpoint-xxx-merged \
    --load_dataset_config true
```