昇腾社区首页
支持
消息
我的论坛主页
论坛首页
/
mindie 2.0.RC1-300I-Duo-py311-openeuler24.03-lts版本镜像部署MiniCPM-V2.6-8B失败
收藏回复举报
mindie 2.0.RC1-300I-Duo-py311-openeuler24.03-lts版本镜像部署MiniCPM-V2.6-8B失败
t('forum.solved') 已解决
新人帖
发表于2025-05-22 17:25:29
0 查看

在 300I Duo 推理卡部署MiniCPM-V2.6-8B多模态模型失败，npu信息如下：

Snipaste_2025-05-22_16-45-50.png

使用的mindie镜像版本为5月8日更新的2.0.RC1-300I-Duo-py311-openeuler24.03-lts，具体操作步骤按照文档指示，大概过程如下：

  1. 使用命令启动容器，启动命令如下，将模型拷贝到/home目录下，修改模型权限，权限为750，修改MiniCPM-V-2_6中的config.json文件的"torch_dtype"为"float16", 
    docker run -dit -u root \
--name minicpm \
-e ASCEND_RUNTIME_OPTIONS=NODRV \
--privileged=true \
-v /mnt/ai/common_data/model_download:/mnt \
-v /usr/local/Ascend/driver/:/usr/local/Ascend/driver/ \
-v /usr/local/Ascend/firmware/:/usr/local/Ascend/firmware/ \
-v /usr/local/sbin/:/usr/local/sbin \
-v /etc/ascend_install.info:/etc/ascend_install.info \
-v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
--shm-size=100g \
-p 40000:22 \
--cap-add=SYS_PTRACE \
--security-opt seccomp=unconfined \
192.168.2.52:8880/ai/ascendai/mindie:2.0.RC1-300I-Duo-py311-openeuler24.03-lts \
/bin/bash
  2. 将MiniCPM-V-2_6权重目录中的resampler.py拷贝到/usr/local/Ascend/atb-models/atb_llm/models/minicpm_qwen2_v2目录下，并安装第三方库依赖 
    cd /usr/local/Ascend/atb-models/requirements
pip install -r requirements.txt
cd /usr/local/Ascend/atb-models/requirements/models
pip install -r requirements_minicpm_qwen2_v2.txt
  3. 编辑服务化推理配置文件vim /usr/local/Ascend/mindie/latest/mindie-service/conf/config.json，配置文件内容如下： 
    {
    "Version" : "1.0.0",

    "ServerConfig" :
    {
        "ipAddress" : "127.0.0.1",
        "managementIpAddress" : "127.0.0.2",
        "port" : 1025,
        "managementPort" : 1026,
        "metricsPort" : 1027,
        "allowAllZeroIpListening" : false,
        "maxLinkNum" : 1000,
        "httpsEnabled" : false,
        "fullTextEnabled" : false,
        "tlsCaPath" : "security/ca/",
        "tlsCaFile" : ["ca.pem"],
        "tlsCert" : "security/certs/server.pem",
        "tlsPk" : "security/keys/server.key.pem",
        "tlsPkPwd" : "security/pass/key_pwd.txt",
        "tlsCrlPath" : "security/certs/",
        "tlsCrlFiles" : ["server_crl.pem"],
        "managementTlsCaFile" : ["management_ca.pem"],
        "managementTlsCert" : "security/certs/management/server.pem",
        "managementTlsPk" : "security/keys/management/server.key.pem",
        "managementTlsPkPwd" : "security/pass/management/key_pwd.txt",
        "managementTlsCrlPath" : "security/management/certs/",
        "managementTlsCrlFiles" : ["server_crl.pem"],
        "kmcKsfMaster" : "tools/pmt/master/ksfa",
        "kmcKsfStandby" : "tools/pmt/standby/ksfb",
        "inferMode" : "standard",
        "interCommTLSEnabled" : true,
        "interCommPort" : 1121,
        "interCommTlsCaPath" : "security/grpc/ca/",
        "interCommTlsCaFiles" : ["ca.pem"],
        "interCommTlsCert" : "security/grpc/certs/server.pem",
        "interCommPk" : "security/grpc/keys/server.key.pem",
        "interCommPkPwd" : "security/grpc/pass/key_pwd.txt",
        "interCommTlsCrlPath" : "security/grpc/certs/",
        "interCommTlsCrlFiles" : ["server_crl.pem"],
        "openAiSupport" : "vllm",
        "tokenTimeout" : 600,
        "e2eTimeout" : 600,
        "distDPServerEnabled":false
    },

    "BackendConfig" : {
        "backendName" : "mindieservice_llm_engine",
        "modelInstanceNumber" : 1,
        "npuDeviceIds" : [[0]],
        "tokenizerProcessNumber" : 8,
        "multiNodesInferEnabled" : false,
        "multiNodesInferPort" : 1120,
        "interNodeTLSEnabled" : true,
        "interNodeTlsCaPath" : "security/grpc/ca/",
        "interNodeTlsCaFiles" : ["ca.pem"],
        "interNodeTlsCert" : "security/grpc/certs/server.pem",
        "interNodeTlsPk" : "security/grpc/keys/server.key.pem",
        "interNodeTlsPkPwd" : "security/grpc/pass/mindie_server_key_pwd.txt",
        "interNodeTlsCrlPath" : "security/grpc/certs/",
        "interNodeTlsCrlFiles" : ["server_crl.pem"],
        "interNodeKmcKsfMaster" : "tools/pmt/master/ksfa",
        "interNodeKmcKsfStandby" : "tools/pmt/standby/ksfb",
        "ModelDeployConfig" :
        {
            "maxSeqLen" : 2048,
            "maxInputTokenLen" : 2048,
            "truncation" : false,
            "ModelConfig" : [
                {
                    "modelInstanceType" : "Standard",
                    "modelName" : "minicpm_qwen2_v2",
                    "modelWeightPath" : "/home/MiniCPM-V-2_6",
                    "worldSize" : 1,
                    "cpuMemSize" : 5,
                    "npuMemSize" : 1,
                    "backendType" : "atb",
                    "trustRemoteCode" : false
                }
            ]
        },

        "ScheduleConfig" :
        {
            "templateType" : "Standard",
            "templateName" : "Standard_LLM",
            "cacheBlockSize" : 128,

            "maxPrefillBatchSize" : 50,
            "maxPrefillTokens" : 4096,
            "prefillTimeMsPerReq" : 150,
            "prefillPolicyType" : 0,

            "decodeTimeMsPerReq" : 50,
            "decodePolicyType" : 0,

            "maxBatchSize" : 200,
            "maxIterTimes" : 512,
            "maxPreemptCount" : 0,
            "supportSelectBatch" : false,
            "maxQueueDelayMicroseconds" : 5000
        }
    }
}
  4. 拉起服务化部署 
    cd /usr/local/Ascend/mindie/latest/mindie-service/bin
./mindieservice_daemon
  5. 打印信息如下： 
    [root@c71538e6230a debug]# cd /usr/local/Ascend/mindie/latest/mindie-service/bin
[root@c71538e6230a bin]# ./mindieservice_daemon
The old environment variable ATB_LOG_TO_STDOUT will be deprecated on 2025/12/31. Please use the new environment variable MINDIE_LOG_TO_STDOUT as soon as possible.
The old environment variable ATB_LOG_LEVEL will be deprecated on 2025/12/31. Please use the new environment variable MINDIE_LOG_LEVEL as soon as possible.
The old environment variable MINDIE_LLM_LOG_LEVEL will be deprecated on 2025/12/31. Please use the new environment variable MINDIE_LOG_LEVEL as soon as possible.
The old environment variable MINDIE_LLM_PYTHON_LOG_TO_STDOUT will be deprecated on 2025/12/31. Please use the new environment variable MINDIE_LOG_TO_STDOUT as soon as possible.
The old environment variable MINDIE_LLM_LOG_TO_STDOUT will be deprecated on 2025/12/31. Please use the new environment variable MINDIE_LOG_TO_STDOUT as soon as possible.
The old environment variable MINDIE_LLM_PYTHON_LOG_PATH will be deprecated on 2025/12/31. Please use the new environment variable MINDIE_LOG_PATH as soon as possible.
The old environment variable ATB_LOG_TO_FILE will be deprecated on 2025/12/31. Please use the new environment variable MINDIE_LOG_TO_FILE as soon as possible.
The old environment variable MINDIE_LLM_PYTHON_LOG_LEVEL will be deprecated on 2025/12/31. Please use the new environment variable MINDIE_LOG_LEVEL as soon as possible.
The old environment variable MINDIE_LLM_PYTHON_LOG_TO_FILE will be deprecated on 2025/12/31. Please use the new environment variable MINDIE_LOG_TO_FILE as soon as possible.
The old environment variable MINDIE_LLM_LOG_TO_FILE will be deprecated on 2025/12/31. Please use the new environment variable MINDIE_LOG_TO_FILE as soon as possible.
The old environment variable OCK_LOG_LEVEL will be deprecated on 2025/12/31. Please use the new environment variable MINDIE_LOG_LEVEL as soon as possible.
The old environment variable OCK_LOG_TO_STDOUT will be deprecated on 2025/12/31. Please use the new environment variable MINDIE_LOG_TO_STDOUT as soon as possible.
LLMInferEngine failed to init LLMInferModels
Killed
  6. 进入路径/root/mindie/log/debug，日志内容如下：

    mindie-llm_1712_20250522172134237.log

    2025-05-22 17:21:34.237 1714 LLM log default format: [yyyy-mm-dd hh:mm:ss.uuuuuu] [processid] [threadid] [llm] [loglevel] [file:line] [status code] msg
[2025-05-22 17:21:34.237] [1712] [281471257145312] [llm] [INFO] [llm_manager_impl.cpp:76] LLMRuntime init success!
[2025-05-22 17:21:34.238] [1712] [281471257145312] [llm] [DEBUG] [model_deploy_config.cpp:80] ModelDeployConfig::pySonamelibpython3.11.so
[2025-05-22 17:21:38.449] [1712] [281471257145312] [llm] [ERROR] [model_deploy_config.cpp:102] Failed to get vocab size from tokenizer wrapper with exception: ValueError: Get instance from AutoProcessor failed. Please check the input parameters model_path and kwargs. If the input parameters are valid and the required files exist in model_path, make sure the folder's owner has execute permission. Otherwise, please check the function stack for detailed exception information.

At:
  /usr/local/Ascend/atb-models/atb_llm/models/base/model_utils.py(312): wrapper
  /usr/local/Ascend/atb-models/atb_llm/models/minicpm_qwen2_v2/router_minicpm_qwen2_v2.py(20): __post_init__
  <string>(23): __init__
  /usr/local/Ascend/atb-models/atb_llm/models/__init__.py(110): get_model
  /usr/local/Ascend/atb-models/atb_llm/runner/tokenizer_wrapper.py(18): __init__
  /usr/local/lib/python3.11/site-packages/mindie_llm/modeling/model_wrapper/atb/atb_tokenizer_wrapper.py(8): __init__
  /usr/local/lib/python3.11/site-packages/mindie_llm/modeling/model_wrapper/__init__.py(28): get_tokenizer_wrapper

[2025-05-22 17:21:38.450] [1712] [281471257145312] [llm] [ERROR] [llm_manager_impl.cpp:784] Config manager init exception: ValueError: Get instance from AutoProcessor failed. Please check the input parameters model_path and kwargs. If the input parameters are valid and the required files exist in model_path, make sure the folder's owner has execute permission. Otherwise, please check the function stack for detailed exception information.

At:
  /usr/local/Ascend/atb-models/atb_llm/models/base/model_utils.py(312): wrapper
  /usr/local/Ascend/atb-models/atb_llm/models/minicpm_qwen2_v2/router_minicpm_qwen2_v2.py(20): __post_init__
  <string>(23): __init__
  /usr/local/Ascend/atb-models/atb_llm/models/__init__.py(110): get_model
  /usr/local/Ascend/atb-models/atb_llm/runner/tokenizer_wrapper.py(18): __init__
  /usr/local/lib/python3.11/site-packages/mindie_llm/modeling/model_wrapper/atb/atb_tokenizer_wrapper.py(8): __init__
  /usr/local/lib/python3.11/site-packages/mindie_llm/modeling/model_wrapper/__init__.py(28): get_tokenizer_wrapper

    mindie-server_1712_202505221721450.log

    [2025-05-22 17:21:38.450+08:00] [1712] [1714] [server] [ERROR] [llm_infer_model_instance.cpp:284] : MIE05E040001[llm_backend] [LLMInferModelInstance::Init] llmManager_ init fail!
[2025-05-22 17:21:38.450+08:00] [1712] [1714] [server] [ERROR] [llm_infer_model.cpp:28] : MIE05E040001[llm_backend] ModelInstance init failed
[2025-05-22 17:21:38.450+08:00] [1712] [1714] [server] [ERROR] [llm_infer_engine.cpp:41] : MIE05E040001[llm_backend] LLMInferEngine initialize LLMInferModels failed. please check the error stacks for more details
[2025-05-22 17:21:38.450+08:00] [1712] [1714] [server] [ERROR] [infer_backend_manager.cpp:216] : [MIE04E06011C] [infer_backend_manager] Failed to init engine. backendName is mindieservice_llm_engine
[2025-05-22 17:21:38.450+08:00] [1712] [1714] [server] [ERROR] [endpoint.cpp:45] : [MIE04E02011C] [endpoint] Failed to init engine! Please check in the env MINDIE_LOG_PATH (default ~/mindie/log) or console output.

    mindie-llmmodels_1712_20250522172136174.log日志为空。

    如果需要更详细日志，请回复我

我要发帖子

我们使用cookie来确保您的高速浏览体验。继续浏览本站，即表示您同意我们使用cookie。查看详情