# 16kHz Audio Processing

A language identification model fine-tuned from Facebook's Massively Multilingual Speech project, supporting audio classification for 126 languages

Wav2Vec2 is a self-supervised learning-based speech recognition model, pre-trained and fine-tuned on 960 hours of LibriSpeech audio data, supporting English speech transcription.

A large-scale Wav2Vec2-Conformer speech recognition model using relative position embedding, fine-tuned on 100 hours of Librispeech data

This is a speaker verification model based on the Wav2Vec2 architecture, specifically designed for the speaker verification task in the SUPERB benchmark.

A speech intent classification model fine-tuned on the SUPERB intent classification task, based on the Hubert-Base-LS960 pre-trained model

A speaker identification model fine-tuned on the VoxCeleb1 dataset based on the Wav2Vec2-base pre-trained model, designed for voice classification tasks

This is a speech emotion recognition model based on the Wav2Vec2 architecture, adapted from the S3PRL project, designed to identify emotional categories in speech.

This model is a keyword spotting model based on the Hubert architecture, designed to classify speech segments into predefined keyword sets.

This is a Turkish automatic speech recognition model based on the XLSR Wav2Vec2 architecture, fine-tuned on the Common Voice Turkish dataset.

This is an emotion recognition model based on the Wav2Vec2-Large model, specifically designed to identify emotion categories from speech.

SEW-D-mid is a speech pre-training model developed by ASAPP Research, focusing on automatic speech recognition tasks, achieving a good balance between performance and efficiency.

This model is a fine-tuned version of Facebook's Wav2Vec2-Large-Robust, specifically optimized for telephone speech recognition tasks, using 300 hours of Switchboard telephone speech corpus for fine-tuning.

Hubert-based speaker recognition model optimized for the SUPERB benchmark tasks

A Hindi speech recognition model fine-tuned from facebook/wav2vec2-large-xlsr-53, trained using OpenSLR and Common Voice Hindi datasets, supporting 16kHz sampling rate audio input.

Intent classification model based on Wav2Vec2-Large-LV60, fine-tuned on the SUPERB intent classification task for speech command intent recognition

SEW-D-tiny is an efficient speech recognition pre-trained model developed by ASAPP Research, focusing on the balance between performance and efficiency.

This model is an emotion recognition model based on the Hubert-Base architecture, trained on the SUPERB emotion recognition task for speech emotion classification

This model is an intent classification model based on Wav2Vec2-base, specifically designed for recognizing intents in voice commands, supporting the classification of speech segments into predefined intent categories.

SEW (Squeezed and Efficient Wav2vec) is a speech recognition pre-trained model developed by ASAPP Research, outperforming wav2vec 2.0 in both performance and efficiency.

SEW-D-mid-k127 is an efficient speech recognition pre-trained model developed by ASAPP Research, demonstrating significant improvements in performance and efficiency compared to wav2vec 2.0.

Large-scale German automatic speech recognition (ASR) model based on Facebook's Wav2Vec2 architecture, fine-tuned on the Common Voice German dataset

Speaker identification model based on the Wav2Vec2-Large architecture, trained on the VoxCeleb1 dataset for classifying speech by speaker identity

An emotion recognition model based on Hubert-Large pre-trained model for predicting emotion categories in speech

A Bengali automatic speech recognition model fine-tuned based on facebook/wav2vec2-large-xlsr-53, trained with 40,000 speech samples from the OpenSLR dataset

An automatic speech recognition model based on the UniSpeech-SAT base model, fine-tuned on 100 hours of LibriSpeech data