Downloads · 30 days
15
6% of all-time downloads
mimi_throat_microphone
Cnam-LMSSC/mimi_throat_microphone
mimi_throat_microphone is a feature extraction model from Cnam-LMSSC. Use it when you need embeddings to search or compare text. It is set up for transformers.
For streaming usage, please refer to this script
All-time downloads
267
Public
Parameters
96.2M
1.9 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors1.2 GB · 100%
At a glance
- Task
- Feature Extraction
- Library
- transformers
- Model type
- mimi
- Access
- Public
- Created
- Jun 3, 2025
- Updated
- Jun 16, 2025
- SHA
- 2544c0df
Model card
From the Hugging Face model README
Inference script :
import torch, torchaudio
from datasets import load_dataset
from moshi.models import loaders
weight_path = loaders.hf_hub_download("Cnam-LMSSC/mimi_throat_microphone", "kyutai_implementation.safetensors")
model = loaders.get_mimi(weight_path).eval()
model.set_num_codebooks(model.total_codebooks) # use all codebooks
test_dataset = load_dataset("Cnam-LMSSC/vibravox", "speech_clean", split="test", streaming=True)
audio_48kHz = torch.Tensor(next(iter(test_dataset))["audio.throat_microphone"]["array"])
audio_24kHz = torchaudio.functional.resample(audio_48kHz, orig_freq=48_000, new_freq=24_000)
enhanced_audio_24kHz = model.decode(model.encode(audio_24kHz[None, None, :]))
For streaming usage, please refer to this script
- Task
- Feature Extraction
- Library
- transformers
- Type
- mimi
- Created
- Jun 3, 2025
- Updated
- Jun 16, 2025