shisa-ai/shisa-v2.1-llama3.2-3b
Full fine-tune vs meta-llama/Llama-3.2-3B-Instruct. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Embeddings 1/1, Norms 57/57, MLP 84/84, Attention 112/112
4 derivatives checked
- mradermacher/shisa-v2.1-llama3.2-3b-GGUF: Same weights, converted
- dahara1/shisa-v2.1-llama3.2-3b-UD-japanese-imatrix: Quantization
- mradermacher/shisa-v2.1-llama3.2-3b-i1-GGUF: Quantization
- XpressAI/shisa-v2.1-llama3.2-3b-GGUF: Same weights, converted