search

etzinis / two_step_mask_learning

A two step optimization for sound source separation on the adaptive front-end domain
66 stars 16 forks source link

About readme #2

Open LittleFlyingSheep opened 4 years ago

LittleFlyingSheep commented 4 years ago

Excuse me,I want to study your code but I can't understand the correct run steps.Could you provide a run steps explanation on readme?

etzinis commented 4 years ago

Hey, thanks for reaching out. Sorry about the delay but I am kind of trying to catch a deadline, I will append some basic training steps for the recipe. However, if you wanna start playing with the code you can do the following (in a more abstract way to put it): 1) Download e.g. the ESC-50 dataset https://github.com/karolpiczak/ESC-50 2) Preprocess it in order to feed it to the models: https://github.com/etzinis/two_step_mask_learning/blob/master/two_step_mask_learning/utils/create_esc50_hierarchical.py 3) Train an adaptive front end using: https://github.com/etzinis/two_step_mask_learning/blob/master/two_step_mask_learning/dnn/experiments/run_adaptive_fes.py 4) Train the separation module: https://github.com/etzinis/two_step_mask_learning/blob/master/two_step_mask_learning/dnn/experiments/run_adaptive_fes.py

shulan-l commented 11 months ago

Hello, following the steps you provided, I still cannot run through the code correctly. Could you provide more detailed steps? Thank you.