|
--- |
|
license: mit |
|
language: |
|
- en |
|
|
|
--- |
|
### Model Description: |
|
|
|
This is the wav2vec2-base model being pre-trained on the My Science Tutor (MyST 470h) dataset (from [LDC](https://catalog.ldc.upenn.edu/LDC2021S05)). |
|
|
|
The pertaining is done by using [fairseq](https://github.com/facebookresearch/fairseq/blob/main/examples/wav2vec/README.md) (wav2vec2_base_librispeech config). |
|
|
|
The converge checkpoint is converted from PyTorch model to Hugging Face model by using a modified version of [convertor script](https://github.com/huggingface/transformers/blob/main/src/transformers/models/wav2vec2/convert_wav2vec2_original_pytorch_checkpoint_to_pytorch.py) offered by Huggingface |
