Cannot use GPU

[hjj-lmx]

pip install rembg[gpu]
call
from rembg import remove
from PIL import Image
input_path = 'input.png'
output_path = 'output.png'
input = Image.open(input_path)
output = remove(input)
output.save(output_path)
GPU is not running

[dicebattle]

The same goes for me. I'll write down a little more detailed environment.
OS: Ubuntu 20.04
CUDA: 12.2
ONNX-Runtime: 1.18.1
rembg: (latest)

If I launch the program in server mode and check the GPU utilisation with the nvidia-smi command, the GPU is not running at all. On the other hand, if I check the CPU utilisation with the htop command, it is using all 12 cores.

[FuzzyCatZ]

Same with me

[gongminmin]

My solution is

1. Download CUDNN v9.2, extract, put its bin into the PATH.
2. Download zlib (zlibwapi.dll on Windows), also put its path into PATH.
3. pip uninstall onnxruntime_gpu
4. pip install onnxruntime_gpu

Then it runs on GPU.

[lhh-pi]

providers = ['CUDAExecutionProvider', 'CPUExecutionProvider']
session = new_session(providers=providers)
f = remove(image, session=session)

You can try this. It worked for me.

[github-actions]

This issue is stale because it has been open for 30 days with no activity.

[github-actions]

This issue was closed because it has been inactive for 14 days since being marked as stale.