Subject: TensorFlow GPU Failure on RTX 5090 Laptop GPU in WSL2/Docker with Latest NVIDIA Drivers (576.xx)

same issue here, I am using GeoForce RTX 5090 Laptop GPU in WSL, I directly install tensorflow[and-cuda] by pip,
tf.config.list_physical_devices() can see GPU, but when running a real python code, it show CUDA_ERROR_INVALID_HANDLE

1 Like