Automatic Speech Recognition
Transformers
Safetensors
MLX
voxtral
mistral
quantized
8bit
llm
language-model
8-bit precision
Instructions to use VincentGOURBIN/voxtral-small-8bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use VincentGOURBIN/voxtral-small-8bit with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="VincentGOURBIN/voxtral-small-8bit")# pip install -U transformers accelerate # Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("VincentGOURBIN/voxtral-small-8bit") model = AutoModelForMultimodalLM.from_pretrained("VincentGOURBIN/voxtral-small-8bit", device_map="auto") - MLX
How to use VincentGOURBIN/voxtral-small-8bit with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download VincentGOURBIN/voxtral-small-8bit --local-dir voxtral-small-8bit
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Download params.json from VincentGOURBIN/voxtral-small-8bit: direct link, hf CLI and curl.
- Browser
- Download file 732 Bytes
-
https://huggingface.co/VincentGOURBIN/voxtral-small-8bit/resolve/main/params.json
- Command line
-
hf download hf://VincentGOURBIN/voxtral-small-8bit/params.json
-
curl -L -o params.json https://huggingface.co/VincentGOURBIN/voxtral-small-8bit/resolve/main/params.json
732 Bytes
| { | |
| "dim": 5120, | |
| "n_layers": 40, | |
| "head_dim": 128, | |
| "hidden_dim": 32768, | |
| "n_heads": 32, | |
| "n_kv_heads": 8, | |
| "rope_theta": 100000000.0, | |
| "norm_eps": 1e-05, | |
| "vocab_size": 131072, | |
| "max_position_embeddings": 32768, | |
| "multimodal": { | |
| "whisper_model_args": { | |
| "encoder_args": { | |
| "dim": 1280, | |
| "n_layers": 32, | |
| "head_dim": 64, | |
| "hidden_dim": 5120, | |
| "n_heads": 20, | |
| "vocab_size": 51866, | |
| "max_source_positions": 1500, | |
| "audio_encoding_args": { | |
| "sampling_rate": 16000, | |
| "num_mel_bins": 128, | |
| "hop_length": 160, | |
| "window_size": 400 | |
| } | |
| }, | |
| "downsample_args": { | |
| "downsample_factor": 4 | |
| } | |
| } | |
| } | |
| } |