mistralai/Ministral-3-3B-Instruct-2512-BF16
Partial fine-tune vs mistralai/Ministral-3-3B-Base-2512. Partial change: 89% of the weights changed (attention, embeddings, mlp, norms, other), in 26 of 26 layers.
Tensors changed: Embeddings 1/1, Norms 43/103, MLP 78/150, Attention 104/200, other 2/4
11 derivatives checked
- lmstudio-community/Ministral-3-3B-Instruct-2512-GGUF: Quantization (fewer layers)
- bartowski/mistralai_Ministral-3-3B-Instruct-2512-GGUF: Same weights, converted (fewer layers)
- StandardThinking/StandardOne-3B: LoRA-style
- mradermacher/Ministral-3-3B-Instruct-2512-BF16-i1-GGUF: Quantization (fewer layers)
- ggml-org/Ministral-3-3B-Instruct-2512-GGUF: Quantization (fewer layers)
- EnlistedGhost/Ministral-3-3B-Instruct-2512-GGUF: Same weights, converted (fewer layers)
- Aratako/Ministral-3-3B-Instruct-2512-BF16-TextOnly: Same weights, repacked (fewer layers)
- squ11z1/LeChatonFat: LoRA-style
- mradermacher/Ministral-3-3B-Instruct-2512-BF16-GGUF: Same weights, converted (fewer layers)
- vinci00/ministral-3-3b-gguf-q4: Quantization (fewer layers)
- OpenMed/Ministral-3B-MedVL: LoRA-style