OpenLLM-France/Luciole-8B-Instruct-1.1

Full fine-tune vs OpenLLM-France/Luciole-8B-Base. Fine-tune: 100% of the weights changed, across 52 of 52 layers.

Tensors changed: Embeddings 1/1, other 143/168, Norms 66/77, MLP 48/48, Attention 16/16, lm_head 1/1

2 derivatives checked

• OpenLLM-France/Luciole-8B-Instruct-1.1-GGUF: Same weights, converted
• OpenLLM-France/Luciole-8B-Audio-1.0: LoRA-style (fewer layers)

On the Hub · Downloads on Model Pulse