beomi/KoAlpaca-Polyglot-12.8B

Full fine-tune vs EleutherAI/polyglot-ko-12.8b. Fine-tune: 99% of the weights changed, across 40 of 40 layers.

Tensors changed: Attention 160/280, Norms 162/162, MLP 160/160

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse