jahyungu/Falcon3-7B-Instruct_openbookqa

Full fine-tune vs tiiuae/Falcon3-7B-Instruct. Fine-tune: 95% of the weights changed, across 28 of 28 layers.

Tensors changed: Embeddings 0/1, Norms 25/57, MLP 84/84, Attention 112/112, lm_head 1/1

1 derivatives checked

• mradermacher/Falcon3-7B-Instruct_openbookqa-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse