Hastagaras/L3.2-JametMini-3B-MK.I

Full fine-tune vs meta-llama/Llama-3.2-3B-Instruct. Fine-tune: 98% of the weights changed, across 28 of 28 layers.

Tensors changed: Embeddings 1/1, Norms 15/57, MLP 82/84, Attention 109/112

3 derivatives checked

• mradermacher/L3.2-JametMini-3B-MK.I-GGUF: Same weights, converted
• SFBAI/multillama-3.2-merge: Full fine-tune
• bunnycore/Llama-3.2-3B-TitanFusion-v2: Full fine-tune

On the Hub · Downloads on Model Pulse