dphn/Dolphin3.0-Mistral-24B
Full fine-tune vs mistralai/Mistral-Small-24B-Base-2501. Fine-tune: 100% of the weights changed, across 40 of 40 layers.
Tensors changed: Norms 74/81, MLP 120/120, Attention 160/160
13 derivatives checked
- bartowski/cognitivecomputations_Dolphin3.0-Mistral-24B-GGUF: Quantization
- eaddario/Dolphin3.0-Mistral-24B-GGUF: Same weights, converted
- mradermacher/Dolphin3.0-Mistral-24B-GGUF: Quantization
- mradermacher/Dolphin3.0-Mistral-24B-i1-GGUF: Quantization
- DevQuasar-7/cognitivecomputations.Dolphin3.0-Mistral-24B-GGUF: Same weights, converted
- Valdemardi/Dolphin3.0-Mistral-24B-AWQ: Quantization
- weehal/Dolphin3.0-Mistral-24B-mlx-4Bit: Quantization
- tensorblock/Dolphin3.0-Mistral-24B-GGUF: Quantization
- Saxanth/Heretic-Doliphin3-24b-Base-GGUF: Quantization
- moot20/Dolphin3.0-Mistral-24B-MLX-8bits: Quantization
- Cran-May/SCE-24B: Full fine-tune
- mergekit-community/24B-MS-PRO-V0.01: Full fine-tune
- Vortex5/LuckyRP-24B: Full fine-tune
1 other repos hold byte-for-byte identical weight files.