AXCXEPT/EZO-Llama-3.2-3B-Instruct-dpoE
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 15/57, MLP 84/84, Attention 112/112
5 derivatives checked
- mradermacher/EZO-Llama-3.2-3B-Instruct-dpoE-i1-GGUF: Quantization
- mradermacher/EZO-Llama-3.2-3B-Instruct-dpoE-GGUF: Same weights, converted
- MaziyarPanahi/EZO-Llama-3.2-3B-Instruct-dpoE-GGUF: Same weights, converted
- mergekit-community/CodeMix-JPID-3B-Llama3.2: Full fine-tune
- Sakalti/Llama3.2-3B-Uranus-1: Same weights, converted