tiruppali/Falcon-3B-8k-data-new-classes

Partial fine-tune vs tiiuae/Falcon3-3B-Instruct. Partial change: 73% of the weights changed (attention, mlp, norms), in 22 of 22 layers.

Tensors changed: Embeddings 0/1, Norms 1/45, MLP 64/66, Attention 87/88, lm_head 0/1

1 derivatives checked

• DevQuasar-7/tiruppali.Falcon-3B-8k-data-new-classes-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse