MBZUAI-Paris/Nile-Chat-12B
Full fine-tune (fewer layers) vs google/gemma-3-12b-pt. MBZUAI-Paris/Nile-Chat-12B matches part of google/gemma-3-12b-pt (626 of its 1065 tensors; the rest of google/gemma-3-12b-pt has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 48 of 48 layers.
Tensors changed: Embeddings 1/1, Norms 39/289, MLP 144/144, Attention 192/192
3 derivatives checked
- mradermacher/Nile-Chat-12B-GGUF: Quantization
- mradermacher/Nile-Chat-12B-i1-GGUF: Quantization
- DevQuasar/MBZUAI-Paris.Nile-Chat-12B-GGUF: Same weights, converted
1 other repos hold byte-for-byte identical weight files.