license: openrail
language:
- en
metrics:
- f1
- recall
- accuracy
library_name: speechbrain
pipeline_tag: audio-classification
Model Card for Model ID
We build a CTC-based ASR model using wav2vec 2.0 (W2V2) for children under 4-year-old. We use two-level fine-tuning to gradually reduce age mismatch between adult ASR to child ASR.
We first fine-tune W2V2-LibriSpeech960h using My Science Tutor corpus (consists of conversational speech of students between the third and fifth grades with a virtual tutor) on character level. Then we fine-tune W2V2-MyST using Providence corpus (consists of longititude audio of 6 English-speaking children aged from 1-4 years interacting with their mothers at home) on phoneme sequences or consonant/vowel sequences.
We show W2V2-Providence is helpful for improving children's vocalization classification task on two corpus, including Rapid-ABC and BabbleCor.
Model Sources
For more information regarding this model, please checkout our paper
- Paper: Coming soon
Model Description
Folder contains the best checkpoint of the following setting
- W2V2-MyST by fine-tuning on Librispeech 960h: save_960h/wav2vec2.ckpt
- W2V2-Pro trained on phone sequence: save_MyST_Providence_ep45_filtered/wav2vec2.ckpt
- W2V2-Pro trained on consonant/vowel sequence: save_MyST_Providence_ep45_filtered_cv_only/wav2vec2.ckpt
Uses
We develop our complete fine-tuning recipe using SpeechBrain toolkit available at
- https://github.com/jialuli3/speechbrain/tree/infant-voc-classification/recipes/RABC (used for Rapid-ABC corpus)
- https://github.com/jialuli3/speechbrain/tree/infant-voc-classification/recipes/BabbleCor (used for BabbleCor corpus)
Paper/BibTex Citation
If you found this model helpful to you, please cite us as
Coming soon
Model Card Contact
Jialu Li (she, her, hers)
Ph.D candidate @ Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign
E-mail: jialuli3@illinois.edu
Homepage: https://sites.google.com/view/jialuli/