sthenno-com/miscii-14b-1028
Full fine-tune vs Qwen/Qwen2.5-14B-Instruct. Fine-tune: 100% of the weights changed, across 48 of 48 layers.
Tensors changed: Embeddings 1/1, Norms 86/97, MLP 144/144, Attention 335/336, lm_head 1/1
3 derivatives checked
- bartowski/miscii-14b-1028-GGUF: Same weights, converted
- mradermacher/miscii-14b-1028-i1-GGUF: Quantization
- mradermacher/miscii-14b-1028-GGUF: Quantization