[Bug] qwen2-vl-7b docker delpoy bugs

[jnzbfgjd]

Checklist

• 1. I have searched related issues but cannot get the expected help.
• 2. The bug has not been fixed in the latest version.
• 3. Please note that if the bug-related issue you submitted lacks corresponding environment info and a minimal reproducible demo, it will be challenging for us to reproduce and resolve the issue, reducing the likelihood of receiving feedback.

Describe the bug

I'm using the docker version 0.6.1-cu12 to deploy qwen2-vl-7b local

Reproduction

docker run --runtime nvidia --gpus all
-v ~/.cache/huggingface:/root/.cache/huggingface
--env "HUGGING_FACE_HUB_TOKEN="
-p 23333:23333
--ipc=host
openmmlab/lmdeploy:latest
lmdeploy serve api_server qwen2-vl-7b

Environment

cuda12 A100 40g

Error traceback

Unrecognized keys in `rope_scaling` for 'rope_type'='default': {'mrope_section'}
Unrecognized keys in `rope_scaling` for 'rope_type'='default': {'mrope_section'}
2024-10-22 01:36:08,354 - lmdeploy - �[33mWARNING�[0m - archs.py:53 - Fallback to pytorch engine because `/data/qwen-vl-7b` not supported by turbomind engine.
Unrecognized keys in `rope_scaling` for 'rope_type'='default': {'mrope_section'}
Unrecognized keys in `rope_scaling` for 'rope_type'='default': {'mrope_section'}
2024-10-22 01:36:09,521 - lmdeploy - �[31mERROR�[0m - builder.py:58 - matching vision model: Qwen2VLModel failed
Traceback (most recent call last):
  File "/opt/lmdeploy/lmdeploy/vl/model/qwen2.py", line 15, in check_qwen_vl_deps_install
    import qwen_vl_utils  # noqa: F401
ModuleNotFoundError: No module named 'qwen_vl_utils'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/opt/py3/bin/lmdeploy", line 33, in <module>
    sys.exit(load_entry_point('lmdeploy', 'console_scripts', 'lmdeploy')())
  File "/opt/lmdeploy/lmdeploy/cli/entrypoint.py", line 42, in run
    args.run(args)
  File "/opt/lmdeploy/lmdeploy/cli/serve.py", line 329, in api_server
    run_api_server(args.model_path,
  File "/opt/lmdeploy/lmdeploy/serve/openai/api_server.py", line 1045, in serve
    VariableInterface.async_engine = pipeline_class(
  File "/opt/lmdeploy/lmdeploy/serve/vl_async_engine.py", line 24, in __init__
    self.vl_encoder = ImageEncoder(model_path,
  File "/opt/lmdeploy/lmdeploy/vl/engine.py", line 90, in __init__
    self.model = load_vl_model(model_path, backend_config=backend_config)
  File "/opt/lmdeploy/lmdeploy/vl/model/builder.py", line 56, in load_vl_model
    return module(**kwargs)
  File "/opt/lmdeploy/lmdeploy/vl/model/base.py", line 31, in __init__
    self.build_model()
  File "/opt/lmdeploy/lmdeploy/vl/model/qwen2.py", line 35, in build_model
    check_qwen_vl_deps_install()
  File "/opt/lmdeploy/lmdeploy/vl/model/qwen2.py", line 17, in check_qwen_vl_deps_install
    raise ImportError(
ImportError: please install qwen_vl_utils by pip install qwen_vl_utils

[lvhan028]

openmmlab/lmdeploy:latest doesn't contain the dependent packages of VLM models.
May build the image using this dockfile: https://github.com/InternLM/lmdeploy/blob/main/docker/Qwen2VL_Dockerfile