thesysdev/OUI-1
LoRA-style vs google/diffusiongemma-26B-A4B-it. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 30 of 30 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/4, MLP 90/234, Norms 0/434, other 0/152, Attention 115/223
6 derivatives checked
- Abiray/OUI-1-GGUF: Not comparable yet
- dicksondickson/OUI-1-gguf: Not comparable yet
- sakamakismile/OUI-1-NVFP4: Quantization
- dicksondickson/OUI-1-8bit-MLX: Quantization
- dicksondickson/OUI-1-4bit-mixed-4-8-MLX: Quantization
- dicksondickson/OUI-1-6bit-MLX: Quantization