ValiantLabs/Ministral-3-3B-Reasoning-2512-Esper3.1
LoRA-style vs mistralai/Ministral-3-3B-Reasoning-2512. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of k_proj, o_proj, q_proj, v_proj changed, in 26 of 26 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/103, MLP 0/150, Attention 104/200, other 0/4
2 derivatives checked
- mradermacher/Ministral-3-3B-Reasoning-2512-Esper3.1-i1-GGUF: Quantization (fewer layers)
- mradermacher/Ministral-3-3B-Reasoning-2512-Esper3.1-GGUF: Same weights, converted (fewer layers)