Comment on Next Grok model training with 10T parameter modelComments−ramshankerOP5moThis is the best publically posted model size, ever since top AI labs started treating model size as a trade secret. This should also guide next generation of inference ASICs.
Comments
This is the best publically posted model size, ever since top AI labs started treating model size as a trade secret. This should also guide next generation of inference ASICs.