cstr/Spaetzle-v63-7b

Full fine-tune vs DiscoResearch/DiscoLM_German_7b_v1. Fine-tune: 100% of the weights changed, across 32 of 32 layers.

Tensors changed: Norms 34/65, MLP 96/96, Attention 128/128

1 derivatives checked

On the Hub · Downloads on Model Pulse