• Members 2 posts
    2026年7月30日 15:17

    一、软硬件信息

    镜像版本

    cr.metax-tech.com/public-ai-release/maca/deepspeed:0.16.5-maca.ai3.7.0.6-torch2.6-py310-ubuntu22.04-amd64

    启动容器命令

    CONTAINER_NAME=llamafactory_0.9.2
    DOCKER_IMAGE=cr.metax-tech.com/public-ai-release/maca/deepspeed:0.16.5-maca.ai3.7.0.6-torch2.6-py310-ubuntu22.04-amd64
    
    docker run -itd \
        --device=/dev/dri \
        --device=/dev/mxcd \
        --group-add video \
        --network=host \
        --security-opt seccomp=unconfined \
        --security-opt apparmor=unconfined \
        --shm-size 512g \
        --ulimit memlock=-1 \
        -w /workspace/LLaMA-Factory \
        --name ${CONTAINER_NAME} \
        "${DOCKER_IMAGE}"
    

    mx-msi

    mx-smi version: 2.3.1

    =================== MetaX System Management Interface Log ===================
    Timestamp : Thu Jul 30 15:06:59 2026

    Attached GPUs : 8
    +---------------------------------------------------------------------------------+
    | MX-SMI 2.3.1 Kernel Mode Driver Version: 3.3.12 |
    | MACA Version: 3.7.0.37 BIOS Version: 1.29.1.0 |
    |------------------+-----------------+---------------------+----------------------|
    | Board Name | GPU Persist-M | Bus-id | GPU-Util sGPU-M |
    | Pwr:Usage/Cap | Temp Perf | Memory-Usage | GPU-State |
    |==================+=================+=====================+======================|
    | 0 MetaX C500 | 0 Off | 0000:12:00.0 | 0% Disabled |
    | 60W / 350W | 37C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 1 MetaX C500 | 1 Off | 0000:13:00.0 | 0% Disabled |
    | 60W / 350W | 37C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 2 MetaX C500 | 2 Off | 0000:16:00.0 | 0% Disabled |
    | 60W / 350W | 37C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 3 MetaX C500 | 3 Off | 0000:1a:00.0 | 0% Disabled |
    | 61W / 350W | 36C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 4 MetaX C500 | 4 Off | 0000:8d:00.0 | 0% Disabled |
    | 60W / 350W | 37C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 5 MetaX C500 | 5 Off | 0000:8e:00.0 | 0% Disabled |
    | 62W / 350W | 38C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 6 MetaX C500 | 6 Off | 0000:95:00.0 | 0% Disabled |
    | 60W / 350W | 36C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 7 MetaX C500 | 7 Off | 0000:96:00.0 | 0% Disabled |
    | 60W / 350W | 35C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+

    +---------------------------------------------------------------------------------+
    | Process: |
    | GPU PID Process Name GPU Memory |
    | Usage(MiB) |
    |=================================================================================|
    | no process found |
    +---------------------------------------------------------------------------------+

    二、问题现象

    容器内执行 pip list | grep llamafactory,版本显示为 0.9.2,如下:

    root@server:/workspace/LLaMA-Factory# pip list | grep llamafactory
    llamafactory              0.9.2+maca3.7.0.6
    

    但是 llamafactory 的代码却不是官方版本的 0.9.2,例如 /opt/conda/lib/python3.10/site-packages/llamafactory/data/loader.py 文件,为官方 0.9.2 版本以前的代码。

    请确认该系列镜像下,llamafactory 具体是哪个版本?

    望尽快回复,感谢~

  • arrow_forward

    Thread has been moved from 公共.

  • Members 794 posts
    2026年7月30日 15:22

    尊敬的开发者您好,为定制适配,请和社区版本对比差异

  • Members 2 posts
    2026年7月30日 15:24

    你好,我具体参考 github llamafactory 哪个版本呢?

  • arrow_forward

    Thread has been moved from 解决中.