MBZUAI-Paris/Nile-Chat-4B
Full fine-tune (fewer layers) vs google/gemma-3-4b-pt. MBZUAI-Paris/Nile-Chat-4B matches part of google/gemma-3-4b-pt (444 of its 883 tensors; the rest of google/gemma-3-4b-pt has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 34 of 34 layers.
Tensors changed: Embeddings 1/1, Norms 27/205, MLP 102/102, Attention 136/136
3 derivatives checked
- mradermacher/Nile-Chat-4B-GGUF: Same weights, converted
- mradermacher/Nile-Chat-4B-i1-GGUF: Quantization
- DevQuasar/MBZUAI-Paris.Nile-Chat-4B-GGUF: Same weights, converted