lkevinzc/Llama-3.2-3B-NuminaQA
Full fine-tune vs HuggingFaceTB/FineMath-Llama-3B. 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
3 derivatives checked
- mradermacher/Llama-3.2-3B-NuminaQA-i1-GGUF: Quantization
- mradermacher/Llama-3.2-3B-NuminaQA-GGUF: Same weights, converted
- sail/Llama-3.2-3B-Oat-Zero: Full fine-tune