fo40225
fo40225
Cuda 11 didn't support compute capability 3.0.
I haven't got the RTX3090, but I think you can try those steps. 1. Download the latest CUDA toolkit, install the driver only. 2. Download the CUDA toolkit of the...
You should use the CPU version of tensorflow to confirm that your model and code worked. A misconfigured CUDA environment usually causes exceptions and exit.
所以你有一台機器上面安裝了三個世代的顯示卡，使用相同版本的驅動程式版本與CUDA函式庫與tf版本與原始碼跟模型 但只有安培顯卡得到錯誤結果 您可能真的遇到了舊版CUDA/cudnn在新顯卡上的bug 可以先試試將`%APPDATA%\NVIDIA\ComputeCache`清空，設定環境變數`CUDA_CACHE_MAXSIZE=4294967295`看能不能解決問題 要使用CUDA 11/cudnn 8建置原始的tf1.15，可能需要做非常多移植 修好NVIDIA版本的source code在windows上的建置問題應該比較簡單
方便說明一下您使用keras的範例重現問題的步驟嗎? 我想我應該能借到3090來做測試
Test result Windows AMD Ryzen 7 5800x gigabyte x570 aorus elite F30 4x ADATA DDR4-3200 32GB Crucial P5 1TB GIGABYTE RTX 3090 TURBO 24GB Windows 10 Pro 1903 NVIDIA Driver...
已修復nvidia的程式碼 修改如下 https://github.com/NVIDIA/tensorflow/pull/14 基於此PR建置的whl在 https://github.com/fo40225/tensorflow-windows-wheel/tree/master/1.15.4+nv20.12/ 建置環境 visual studio 2019 16.8 cuda 11.1.1 cudnn 8.0.5.39
Is the CPU+GPU version does not work?
If you use GPU, AVX or not doesn't matter. You can install the Intel Optimization for TensorFlow from conda on windows. https://software.intel.com/content/www/us/en/develop/articles/intel-optimization-for-tensorflow-installation-guide.html
`conda install tensorflow-mkl -c anaconda`