search

experiencor / keras-yolo2

Easy training on custom dataset. Various backends (MobileNet and SqueezeNet) supported. A YOLO demo to detect raccoon run entirely in brower is accessible at https://git.io/vF7vI (not on Windows).
MIT License
1.73k stars 787 forks source link

Mismatch of CUDA versions #378

Open KavyaRavulapati opened 5 years ago

KavyaRavulapati commented 5 years ago

Hello, The code works fine on my PU, but there's a mismatch of CUDA versions while running on GPU. The error is as follows:

E tensorflow/stream_executor/cuda/cuda_dnn.cc:378] Loaded runtime CuDNN library: 7104 (compatibility version 7100) but source was compiled with 7004 (compatibility version 7000). If using a binary install, upgrade your CuDNN library to match. If building from sources, make sure the library loaded at runtime matches a compatible version specified during compile configuration. 2018-11-10 15:57:56.589793: F tensorflow/core/kernels/conv_ops.cc:717] Check failed: stream->parent()->GetConvolveAlgorithms( conv_parameters.ShouldIncludeWinogradNonfusedAlgo(), &algorithms) Aborted (core dumped)

Current CUDA version on GPU: CUDA Version 9.2.148

Looks like the code runs with version 7.

Don't know how to proceed! Kindly help!!

rodrigo2019 commented 5 years ago

@KavyaRavulapati you should review your tensorflow instalation, this repo do not use any type of specific cuda version, it is pure python code that uses the modules installed in your machine.

KavyaRavulapati commented 5 years ago

@rodrigo2019 Yeah, but I'm trying to work on the GPU. Can you please elaborate

rodrigo2019 commented 5 years ago

how did you installed the tensorflow lirabry?

kdv4 commented 5 years ago

pip install tensorflow-gpu