3 ms·
I 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 compu
by slimsag 4y ago
I 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/