ellamind/sui-1-24b
LoRA-style vs mistralai/Mistral-Small-3.2-24B-Instruct-2506. 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 40 of 40 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/131, MLP 120/192, Attention 160/256, other 0/4, lm_head 0/1
1 derivatives checked
- ellamind/sui-1-24b-fp8: Quantization