Automatic Speech Recognition
Transformers
Safetensors
voxtral
feature-extraction
speech
speech-language-model
target-speaker-asr
multi-talker
speaker-diarization
meeting-transcription
Dixtral
Voxtral
DiCoW
BUT-FIT
custom_code
Instructions to use BUT-FIT/Dixtral with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use BUT-FIT/Dixtral with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="BUT-FIT/Dixtral", trust_remote_code=True)# Load model directly from transformers import AutoProcessor, AutoModel processor = AutoProcessor.from_pretrained("BUT-FIT/Dixtral", trust_remote_code=True) model = AutoModel.from_pretrained("BUT-FIT/Dixtral", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -20,9 +20,9 @@ datasets:
|
|
| 20 |
- edinburghcstr/ami
|
| 21 |
---
|
| 22 |
|
| 23 |
-
# 🧠
|
| 24 |
|
| 25 |
-
This repository hosts **
|
| 26 |
**Dixtral** couples the **Voxtral-Mini-3B** spoken-language model with the **DiCoW** diarization-conditioned encoder, giving the LLM target-speaker awareness in multi-talker audio.
|
| 27 |
|
| 28 |
This checkpoint is tuned for **target-speaker / multi-talker transcription (TS-ASR)** of conversational and meeting recordings. For spoken question answering, use [**Dixtral_QA**](https://huggingface.co/BUT-FIT/Dixtral_QA) instead.
|
|
|
|
| 20 |
- edinburghcstr/ami
|
| 21 |
---
|
| 22 |
|
| 23 |
+
# 🧠 Dixtral — BUT-FIT Diarization-Conditioned Voxtral for Target-Speaker ASR
|
| 24 |
|
| 25 |
+
This repository hosts **Dixtral**, developed by [BUT Speech@FIT](https://github.com/BUTSpeechFIT).
|
| 26 |
**Dixtral** couples the **Voxtral-Mini-3B** spoken-language model with the **DiCoW** diarization-conditioned encoder, giving the LLM target-speaker awareness in multi-talker audio.
|
| 27 |
|
| 28 |
This checkpoint is tuned for **target-speaker / multi-talker transcription (TS-ASR)** of conversational and meeting recordings. For spoken question answering, use [**Dixtral_QA**](https://huggingface.co/BUT-FIT/Dixtral_QA) instead.
|