When I tried the tutorials dataioBasics.ipynb and pretraining.ipynb, I met the same problem, it can not find the module “parse_data”, and I can not use pip to install it, because pip also can not find it.
the code is here:
from parse_data import parse_to_json
parse_to_json("./LibriSpeech/dev-clean-2")
And I uninstall speechbrain and run the pip installation commands again, and start jupyter notebook under my speechbrain conda environment, it is no change, same problem .
@titouan.parcollet I dont see parse_data as python module or built_in functionality. seems like colab notebook is bit outdated? since parse_data is not being used in any recipes/templates, instead create_csv/json is being used. And even in notebook, the error generating code line is parse_to_json("./LibriSpeech/dev-clean-2") # how to build this function will be shown later
the comment probably indicates that you need to build this function in any case and it never “shown later” in the notebook
Hi, you are right. However, it works for me on the colab notebook. I’m looking for an answer from the author of this notebook. Nevertheless, it just is a JSON reader, which is something SpeechBrain provides for your recipe.