Wav2vec2 Large Baltic Voxpopuli V2
Facebook's Wav2Vec2 large model, pre-trained on 27.5 hours of unlabeled data from the Baltic language subset of the VoxPopuli corpus.
Downloads 25
Release Time : 3/2/2022
Model Overview
This model is a speech processing model based on the Wav2Vec2 architecture, specifically pre-trained for Baltic languages and suitable for speech recognition tasks.
Model Features
Baltic Language Pre-training
Specifically pre-trained on 27.5 hours of unlabeled data for Baltic languages, making it suitable for speech recognition tasks in these languages.
16kHz Audio Sampling
The model uses a 16kHz sampling rate for speech audio during pre-training. Ensure input speech data is also sampled at 16kHz.
Unsupervised Pre-training
The model is pre-trained on unlabeled data, making it suitable for semi-supervised learning and representation learning tasks.
Model Capabilities
Automatic Speech Recognition
Speech Representation Learning
Use Cases
Speech Recognition
Baltic Language Speech-to-Text
Convert speech audio in Baltic languages to text
Speech Research
Speech Representation Learning
Used for research on representation learning of speech signals
Featured Recommended AI Models
Š 2025AIbase