No kernel image is available for execution on the device

docker_run.sh can be successfully run on my own device.

View post 0

No kernel image is available for execution on the device

Thank you very much for the facility and environment support of from the organizers. However, during the runtime, I encountered the following issue:

The only two modification is on

1.apt.txt (adding g++)

2.Dockerfile(https://repo.anaconda.com/miniconda/Miniconda3-py38_22.11.1-1-Linux-x86_64.sh->https://repo.anaconda.com/miniconda/Miniconda3-py310_24.3.0-0-Linux-x86_64.sh) for a different version of python

However, the following error occurred:

  File "/home/aicrowd/.conda/lib/python3.10/site-packages/vllm/model_executor/layers/layernorm.py", line 59, in forward

    out = torch.empty_like(x)

RuntimeError: CUDA error: no kernel image is available for execution on the device

CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.

For debugging consider passing CUDA_LAUNCH_BLOCKING=1.

Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions

However, it seems that the server does not support docker for different python? Or maybe I changed the miniconda in the wrong way?

detailed information can be seen at AIcrowd

View post 0

How can I use python3.10?

submission can be seen at AIcrowd

View post 0

How can I use python3.10?

Thank you very much for the facility and environment support of from the organizers. However, during the runtime, I encountered the following issue:

I tried to use VLLM to load the LoRA model. However, during loading, an issue arises in versions 3.8 and below.

  File "/home/aicrowd/.conda/lib/python3.8/site-packages/vllm/utils.py", line 54, in __init__

  File "/home/aicrowd/.conda/lib/python3.8/site-packages/vllm/utils.py", line 54, in __init__

    self.cache = OrderedDict[Hashable, T]()

    self.cache = OrderedDict[Hashable, T]()

TypeError: 'type' object is not subscriptable

TypeError: 'type' object is not subscriptable


########################################
#             build-image              #
########################################
Waiting for dockerd to be up...

Waiting for Docker daemon to become available..

This problem was resolved in Python 3.10. To change the Python version, I modified the Dockerfile to the following:

SHELL ["/bin/bash", "-c"]
RUN conda init
RUN conda install cmake -y && conda clean -y -a
COPY --chown=1001:1001 requirements.txt ${HOME_DIR}/requirements.txt
RUN conda create -y -n py310 python==3.10.14
RUN echo "conda activate py310" >> ~/.bashrc
RUN pip install -r requirements.txt --no-cache-dir

COPY --chown=1001:1001 . ${HOME_DIR}

However, from the above errors, it seems that build-image is running after the program execution, resulting in the Python version still being 3.8 at runtime (/home/aicrowd/.conda/lib/python3.8).

View post 1