New ask Hacker News story: How is TPU v3-8 different from v5e-8

How is TPU v3-8 different from v5e-8
2 by sr5434 | 0 comments on Hacker News.
Kaggle announced that they are replacing their TPU v3-8s with v5e-8s, but for some reason I get an OOM when running my code on v5e-8 and not when running it on v3-8. Does anybody know why this might be happening? For reference, I am training a 1.5b GPT model using Torch XLA

Comments

Popular posts from this blog

How can Utilize Call Center Outsourcing for Increase your Business Income well?

New ask Hacker News story: EVM-UI – visual tool to interact with EVM-based smart contracts

New ask Hacker News story: Ask HN: Should I quit my startup journey for now?