3 ms·
That's assuming Apple isn't training their models on Nvidia hardware. Inference is on Apples own silicon though.
by shdh 2y ago
That's assuming Apple isn't training their models on Nvidia hardware.
Inference is on Apples own silicon though.
- mupuff1234 2y agoLooks liked they used Google TPUs https://www.reuters.com/technology/artificial-intelligence/how-apple-used-googles-help-train-its-ai-models-2024-06-11/ https://www.reuters.com/technology/artificial-intelligence/h...