TF installation, WSL2, and Visual Studio Code

[kdlin1]

I followed the instructions on tensorflow.org to install tensorflow on WSL2. It seems to work. This is by running the GPU verification code on the website. However, when I run the same code in Visual Studio Code. It couldn’t find the GPU. On VSC, I did pick the same kernel.

Let me show step by step. On WSL2, the verification code does show GPU:

(tf) dlin@SCL-DLIN:~$ python3 -c “import tensorflow as tf; print(tf.config.list_physical_devices(‘GPU’))”
2024-04-04 20:21:11.437247: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable TF_ENABLE_ONEDNN_OPTS=0.
2024-04-04 20:21:11.463715: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2024-04-04 20:21:11.843862: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
2024-04-04 20:21:12.267983: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:984] could not open file to read NUMA node: /sys/bus/pci/devices/0000:01:00.0/numa_node
Your kernel may have been built without NUMA support.
2024-04-04 20:21:12.292401: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:984] could not open file to read NUMA node: /sys/bus/pci/devices/0000:01:00.0/numa_node
Your kernel may have been built without NUMA support.
2024-04-04 20:21:12.292469: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:984] could not open file to read NUMA node: /sys/bus/pci/devices/0000:01:00.0/numa_node
Your kernel may have been built without NUMA support.
[PhysicalDevice(name=‘/physical_device:GPU:0’, device_type=‘GPU’)]

However, on VSC, it shows the following messages:

What do I do to make the verification code run on VSC? Thanks.

[kdlin1]

I later found that the Tensorflow GPU test scripts work in a .py file. But it does not work in Jupyter notebook. This is when use with VS Code. Somehow my VS Code configuration does not enable Jupyter notebooks to use GPU. Please help.

[Mc_Heights]

Setting up TensorFlow on Windows Subsystem for Linux 2 (WSL2) involves enabling WSL2, installing TensorFlow within WSL2, and integrating it with Visual Studio Code (VS Code) using the Remote - WSL extension for seamless development.
Visit Our Website: MC Height

[kdlin1]

I am able to run TensorFlow on WSL2 on VS Code if I run .py files. My issue is when I use Jupyter notebook under the same conda environment, it cannot find the GPU. The problem exists in both Windows native and WSL2.

What is the URL of MC Height?

[kdlin1]

As shown below, if I run TF .py code, it works fine.

However, if I put the same code in Jupyter notebook, it cannot find the GPU.

.

This problem exists in only one of my 4 computers. I set these 4 computers up the same way. The other 3 work perfectly. Just this new laptop is giving me problems and I have spent 3 days trying to debug. Just odd…