碼上快樂

相關內容    簡體    繁體

MXNetError: [05:53:50] src/operator/nn/./cudnn/cudnn_convolution-inl.h:287

本文轉載自   查看原文   2019-09-19 14:05   408    報錯記錄/ insightface

insightface train.py 報錯：mxnet.base.MXNetError: [05:53:50] src/operator/nn/./cudnn/cudnn_convolution-inl.h:287: Check failed: e == CUDNN_STATUS_SUCCESS (8 vs. 0) : cuDNN: CUDNN_STATUS_EXECUTION_FAILED

解決方案：可能是超過顯存大小了，改小bach_size=36可以正常運行，本人用的顯卡是RTX2080TI


免責聲明！

本站轉載的文章為個人學習借鑒使用，本站對版權不負任何法律責任。如果侵犯了您的隱私權益，請聯系本站郵箱yoyou2525@163.com刪除。



猜您在找 Failed to get convolution algorithm. This is probably because cuDNN failed to initialize 安裝cudnn CUDA和cuDNN cudnn安裝 cudnn.hpp中 cudnnGetErrorString與cudnn.h不兼容問題 “Failed to get convolution algorithm. This is probably because cuDNN failed to initialize”錯誤的解決辦法 cuDNN概述 Error : Failed to get convolution algorithm. This is probably because cuDNN failed to initialize 模型訓練錯誤：Failed to get convolution algorithm，This is probably because cuDNN failed to initialize. win10+keras 跑ResNet 完美解決Failed to get convolution algorithm. This is probably because cuDNN failed to initialize報錯
 
粵ICP備18138465號   © 2018-2025 CODEPRJ.COM