fix EngineGenerationConfig importError of lmdeploy

[irexyc]

PR type

• Bug Fix

PR information

EngineGenerationConfig is removed from lmdeploy and swift can't use inference_lmdeploy due to importError

https://github.com/InternLM/lmdeploy/pull/2375/files
InternLM/lmdeploy#2435

Experiment results

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

# from swift.hub import HubApi
# _api = HubApi()
# _api.login('<your-sdk-token>')  # https://modelscope.cn/my/myaccesstoken

from swift.llm import (
    ModelType, get_lmdeploy_engine, get_default_template_type,
    get_template, inference_lmdeploy, inference_stream_lmdeploy
)

model_type = ModelType.internvl2_2b
model_id_or_path = None
lmdeploy_engine = get_lmdeploy_engine(model_type, model_id_or_path=model_id_or_path)
template_type = get_default_template_type(model_type)
template = get_template(template_type, lmdeploy_engine.hf_tokenizer)
lmdeploy_engine.generation_config.max_new_tokens = 256
generation_info = {}

request_list = [{'query': '<image>描述图片', 'images': ['http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/animal.png']},
                {'query': '你是谁？'},
                {'query': (
                    '<img>http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/cat.png</img>'
                    '<img>http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/animal.png</img>'
                    'What is the difference bewteen the two images?'
                )}]
resp_list = inference_lmdeploy(lmdeploy_engine, template, request_list, generation_info=generation_info)
for request, resp in zip(request_list, resp_list):
    print(f"query: {request['query']}")
    print(f"response: {resp['response']}")

query: <image>描述图片
response: 这是一幅卡通风格的插画，描绘了四只绵羊，从左到右依次排列，背景是一片绿色的草地和山峦。绵羊们表情可爱，有的头上顶着蓬松的羊毛，有的则没有。它们的四肢是黑色的，身体是白色的，耳朵是棕色的，眼睛大而圆，鼻子是棕色的。背景的天空是浅蓝色，点缀着几朵白云，整体画面给人一种宁静、和谐的感觉。
query: 你是谁？
response: 我是InternVL，是由上海人工智能实验室的通用视觉团队（OpenGVLab）和商汤科技联合开发的模型。
query: <img>http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/cat.png</img><img>http://modelscope-open.oss-cn-hangzhou.aliyuncs.com/images/animal.png</img>What is the difference bewteen the two images?
response: The two images are of the same cartoon sheep, but there are subtle differences between them. Let's break down the differences step by step:

1. **Positioning**:
   - In the first image, the sheep is positioned slightly to the left of the center. 
   - In the second image, the sheep is positioned a bit closer to the center but slightly more to the right.

2. **Background and Surroundings**:
   - The background looks slightly blurred in the first image, giving it a soft appearance.
   - The background in the second image is more focused and sharp, giving it a clearer look.

3. **Feathers**:
   - In the second image, the sheep has feathers on its head, which adds to the sense of a natural, well-groomed look.
   - In the first image, the sheep’s head lacks feathers, giving it a more simplistic appearance.

4. **Eye Color**:
   - The eyes in the second image are lighter in color, enhancing the cheerful character of the sheep.

5. **Texture and Fur**:
   - The fur in the second image appears slightly drier or has a different color gradient compared to the first image, giving a more distinct character.

6.

[Jintao-Huang]

thank you very much!