mistralai/Devstral-Small-2507
Full fine-tune (fewer layers) vs mistralai/Mistral-Small-3.1-24B-Instruct-2503. mistralai/Devstral-Small-2507 matches part of mistralai/Mistral-Small-3.1-24B-Instruct-2503 (363 of its 585 tensors; the rest of mistralai/Mistral-Small-3.1-24B-Instruct-2503 has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 40 of 40 layers.
Tensors changed: Embeddings 1/1, Norms 7/81, MLP 120/120, Attention 160/160, lm_head 1/1
23 derivatives checked
- lmstudio-community/Devstral-Small-2507-MLX-8bit: Quantization
- lmstudio-community/Devstral-Small-2507-MLX-4bit: Quantization
- lmstudio-community/Devstral-Small-2507-MLX-6bit: Quantization
- lmstudio-community/Devstral-Small-2507-MLX-bf16: Same weights, repacked
- unsloth/Devstral-Small-2507-GGUF: Same weights, converted
- mistralai/Devstral-Small-2507_gguf: Same weights, converted
- bartowski/mistralai_Devstral-Small-2507-GGUF: Same weights, converted
- cyankiwi/Devstral-Small-2507-AWQ-4bit: Quantization
- lmstudio-community/Devstral-Small-2507-GGUF: Quantization
- gabriellarson/Devstral-Small-2507-GGUF: Quantization
- brittlewis12/Devstral-Small-2507-GGUF: Same weights, converted
- unsloth/Devstral-Small-2507-unsloth-bnb-4bit: Quantization
- apolloparty/Devstral-Small-2507-NVFP4A16: Quantization
- RedHatAI/Devstral-Small-2507-FP8-Dynamic: Quantization
- jkim96/Devstral-Small-2507-DASHQ-Q2-GGUF: Quantization
- stelterlab/Devstral-Small-2507-FP8: LoRA-style (fewer layers)
- DevQuasar/mistralai.Devstral-Small-2507-GGUF: Quantization
- RedHatAI/Devstral-Small-2507-quantized.w8a8: Quantization
- btbtyler09/Devstral-Small-2507-AWQ: Quantization
- mlx-community/Devstral-Samll-2507-bf16: Same weights, repacked
- Tesslate/WEBGEN-Devstral-24B: LoRA-style
- QuixiAI/Devstral-Vision-Small-2507: Partial fine-tune (fewer layers)
- kmouratidis/Devstral-Small-2507-Rebased-Vision: Same weights, converted (fewer layers)