← Back to context Comment by Kuyawa 11 hours ago Duh, China? 1 comment Kuyawa Reply verdverm 11 hours ago We also see improvements like 7x between the current and last Google TPU generations, with different chips for training vs inference.
verdverm 11 hours ago We also see improvements like 7x between the current and last Google TPU generations, with different chips for training vs inference.
We also see improvements like 7x between the current and last Google TPU generations, with different chips for training vs inference.