Instructions to use facebook/mms-1b-all with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use facebook/mms-1b-all with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="facebook/mms-1b-all")# Load model directly from transformers import AutoProcessor, AutoModelForCTC processor = AutoProcessor.from_pretrained("facebook/mms-1b-all") model = AutoModelForCTC.from_pretrained("facebook/mms-1b-all") - Notebooks
- Google Colab
- Kaggle
Wanted argument value for the target language
#13
by airsheysr - opened
The code snippet for the Inference API only requires the audio data as input, without knowing the target language. Is there a way to pass this argument (target language) to the query? e.g. output = query("sample1.flac", language)
'Coz based on some of my inferences, the input audio is transcribed in random language. I mean, it can't accurately identify the target language (esp. for low resource language), so the resulting transcription is incorrect.
Please help ๐
airsheysr changed discussion title from Wanted argument value for the source language to Wanted argument value for the target language