bunnycore/Llama-3.2-3B-All-Mix
Full fine-tune vs Lyte/Llama-3.2-3B-Overthinker. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Embeddings 1/1, Norms 0/57, MLP 84/84, Attention 112/112
6 derivatives checked
- mradermacher/Llama-3.2-3B-All-Mix-GGUF: Same weights, converted
- bunnycore/Llama-3.2-3B-ProdigyPlusPlus: Partial fine-tune
- bunnycore/Llama-3.2-3B-RP-DeepThink: Full fine-tune
- bunnycore/Llama-3.2-3B-RRStock: Full fine-tune
- bunnycore/Llama-3.2-3B-Apex: Partial fine-tune
- bunnycore/Llama-3.2-3B-Creative: LoRA-style