3 ms·
How about running (Py)Torch?
by amelius 4y ago
How about running (Py)Torch?
- slimsag 4y agoI think WebDNN[0] is working on enabling this sort of thing using WebGPU compute as a backend, but no experience with ML myself sorry. I do think WebGPU's compute API would be sufficient for ML inference, but can't say for sure. [0] https://mil-tokyo.github.io/webdnn/ https://mil-tokyo.github.io/webdnn/