mistralai/Voxtral-4B-TTS-2603
Full fine-tune vs mistralai/Ministral-3-3B-Base-2512. Fine-tune: 100% of the weights changed, across 26 of 26 layers.
Tensors changed: Norms 53/53, MLP 78/78, Attention 104/104
6 derivatives checked
- mlx-community/Voxtral-4B-TTS-2603-mlx-4bit: Quantization
- mlx-community/Voxtral-4B-TTS-2603-mlx-bf16: Same weights, repacked
- mlx-community/Voxtral-4B-TTS-2603-mlx-6bit: Quantization
- cstr/voxtral-4b-tts-GGUF: Same weights, converted
- loom-ai-org/voxtral-4b-tts-2603-loom: Not comparable yet
- TrevorJS/voxtral-tts-q4-gguf: Quantization
12 other repos hold byte-for-byte identical weight files.