how to inference with gpu in ai-mask-model? #1621
-
|
how to inference with gpu in ai-mask-model?i find the code mean that use |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 2 replies
-
|
nms is set always cpu because it's usually not computation heavy (as we don't process images there). but the cuda should be enabled for model inference if you install 
|
Beta Was this translation helpful? Give feedback.
-
|
It's working fine, and I couldn't agree with you more right now! it's usually not computation heavy 😁😁😁👍👍👍 |
Beta Was this translation helpful? Give feedback.
-
|
great. thanks for confirming on your end fast. |
Beta Was this translation helpful? Give feedback.
nms is set always cpu because it's usually not computation heavy (as we don't process images there).
but the cuda should be enabled for model inference if you install
pip install onnxruntime-gpu.https://github.com/wkentaro/osam/blob/4d92ff168aaa851f0d00fb9831e10a15ec450712/osam/types/_model.py#L34