Downloads · 30 days
0
voxtral-4b-realtime-2602-mlx
NeoRoth/voxtral-4b-realtime-2602-mlx
voxtral-4b-realtime-2602-mlx is a automatic speech recognition model from NeoRoth. Use it when you need speech turned into text. It is set up for mlx. The card lists the license as apache-2.0.
MLX quantized weights for Voxtral-Mini-4B-Realtime-2602 on Apple Silicon.
Access
Public
Updated Mar 10, 2026
Repo size
3.6 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors3.6 GB · 100%
At a glance
- Task
- Automatic Speech Recognition
- Library
- mlx
- License
- apache-2.0
- Access
- Public
- Created
- Mar 10, 2026
- Updated
- Mar 10, 2026
- SHA
- 435c3b64
Base models
Model card
From the Hugging Face model README
Voxtral Mini 4B Realtime 2602 — MLX
MLX quantized weights for Voxtral-Mini-4B-Realtime-2602 on Apple Silicon.
Architecture
Voxtral Realtime differs from the standard Voxtral Mini 3B. It uses additive fusion (audio embeddings + token embeddings) instead of sequence concatenation, and a causal transformer encoder with sliding window attention (window=750) instead of a bidirectional Whisper-style encoder. This makes it suitable for streaming / low-latency inference.
Variants
| Folder | Quantization | Description |
|---|---|---|
mlx-mxfp4-mixed/ | MXFP4 mixed precision | Text decoder: MXFP4 4-bit (group_size=32). Audio encoder/projector: 8-bit affine (group_size=64). Embeddings: full precision. |
License
This model is distributed under the Apache 2.0 license, following the upstream Voxtral-Mini-4B-Realtime-2602 license.
Requirements
- Apple Silicon (M1+, M5+ recommended for native MXFP4)
- MLX >= 0.30.0
- Python 3.11+
Source
Converted from mistralai/Voxtral-Mini-4B-Realtime-2602 using oriloq-mlx.
- Task
- Automatic Speech Recognition
- Library
- mlx
- License
- apache-2.0
- Languages
- fr, en, de, es, it, pt
- Created
- Mar 10, 2026
- Updated
- Mar 10, 2026