minnesotanlp/Finch-8B-KTO
Full fine-tune vs minnesotanlp/Finch-8B. Fine-tune: 92% of the weights changed, across 36 of 36 layers.
Tensors changed: lm_head 1/1, Embeddings 0/1, Norms 83/145, MLP 108/108, Attention 144/144
2 derivatives checked
- mradermacher/Finch-8B-KTO-i1-GGUF: Quantization
- mradermacher/Finch-8B-KTO-GGUF: Same weights, converted