3 ms·
That's for inference though, the intensive part is training no? Then with a reasonable model (maybe distilled somehow from a larger one through more training) y
by mattnewton 9y ago
That's for inference though, the intensive part is training no? Then with a reasonable model (maybe distilled somehow from a larger one through more training) you could even do it off CPU or whatever.
- gcp 9y agofp16 is usable for training and inference, so the Vega would have been good for both. I care about inference performance too, anyway. CPU is only a (much slower) fallback if the OpenCL drivers are broken.