Any colab notebook for easy inference with pretrained model ?
SutirthaChakraborty opened this issue · 3 comments
Hi, I am glad that you are interested in my work. Since I am only a master's student and only master Python and a small part of TensorFlow language, I am not familiar with Jupyter Notebook and Google's Colab notebook framework.
If you just want to validate the test results, you can simply implement the test and inference process by running the following command with the data and environment deployed.
The test and inference instructions are: python main.py test
Due to the differences in the running environment, I would recommend that you retrain our model in your own environment to achieve more realistic results.
If you have any other questions, feel free to ask me.
Best wishes.
BTW, thank you for your advice. In the future, if I have time, I will learn and create a easier inference process in a way of the colab notebook.
I mean, just a infer infer.py that takes an audio input and use your checkpoints and save the output as a file.
Something like
python3 infer.py input.wav --checkpoint_path --output_filename
Thanks.