Delta-Vector/Hamanasu-4B-PT

Full fine-tune vs IntervitensInc/Llama-3.1-Minitron-4B-Width-Base-chatml. Fine-tune: 91% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 0/1, Norms 18/65, MLP 96/96, Attention 128/128, lm_head 1/1

3 derivatives checked

• mradermacher/Hamanasu-4B-PT-i1-GGUF: Quantization
• mradermacher/Hamanasu-4B-PT-GGUF: Same weights, converted
• Delta-Vector/Hamanasu-4B-Instruct: Full fine-tune

On the Hub · Downloads on Model Pulse