Closed moiseevigor closed 7 years ago
I've had one problem with Jetpack 3.1, solved it with this small change
diff --git a/setTensorFlowEV.sh b/setTensorFlowEV.sh index 9b21dfa..0e6c82c 100755 --- a/setTensorFlowEV.sh +++ b/setTensorFlowEV.sh @@ -27,7 +27,7 @@ TF_CUDA_VERSION=8.0 default_cuda_path=/usr/local/cuda CUDA_TOOLKIT_PATH=$default_cuda_path # cuDNN -TF_CUDNN_VERSION=5.1.10 +TF_CUDNN_VERSION=6.0.21 default_cudnn_path=/usr/lib/aarch64-linux-gnu CUDNN_INSTALL_PATH=$default_cudnn_path # CUDA compute capability
the latest Jetpack solves the annoying gpu memory allocation problem https://devtalk.nvidia.com/default/topic/1013464/jetson-tx2/gpu-out-of-memory-when-the-total-ram-usage-is-2-8g/post/5191096/#5191096
Here is a repo you can use until it is fixed for anyone that finds this...
https://github.com/Goddard/installTensorFlowTX2
Should be fixed with commit: 10b022462ecbf10c534964cdee5530c3a182907a Thank you for your fix.
I've had one problem with Jetpack 3.1, solved it with this small change
the latest Jetpack solves the annoying gpu memory allocation problem https://devtalk.nvidia.com/default/topic/1013464/jetson-tx2/gpu-out-of-memory-when-the-total-ram-usage-is-2-8g/post/5191096/#5191096