This is a heavily interactive web application, and JavaScript is required. Simple HTML interfaces are possible, but that is not what this is.
Post
Fern
fernbear.bsky.social
did:plc:3rf7l25x3gxezm4cruiq63ni
New NanoGPT training speed record: 3.28 FineWeb val loss in 4.66 minutes
Previous record: 5.03 minutes
Changelog:
- FlexAttention blocksize warmup
- hyperparameter tweaks
2024-11-25T01:53:01.653Z