Skip to content

Releases: smpanaro/apple-silicon-4bit-quant

gpt2 family fisher information

06 Mar 04:48
Compare
Choose a tag to compare

Pre-computed Fisher information sensitivities ("weighting" modification from the blog post).

gpt2-large and gpt2-xl are too large for GitHub releases. To re-assemble them download all parts and run:
cat gpt2-large-grads.safetensors.tar.gz.* | tar -xzvf - or cat gpt2-xl-grads.safetensors.tar.gz.* | tar -xzvf -