Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Steps for Ubuntu 18.04 based OS #10

Open
macramole opened this issue Dec 15, 2020 · 1 comment
Open

Steps for Ubuntu 18.04 based OS #10

macramole opened this issue Dec 15, 2020 · 1 comment

Comments

@macramole
Copy link

I'm currently training a model and these are some steps I needed for making it work. Using Linux Mint based on Ubuntu 18.04.

I don't know if this is the correct place to post this but I hope it will be helpful

  • Created a conda enviroment with Python version 3.6 (README.md doesn't specify python version)
  • I had to change some versions on requirements.txt
numpy==1.14.5 #this is the same as you provide, some changes below were because I didn't want to change it
tensorflow-gpu==1.14.0 #you need this if you are using cuda 10.2
librosa==0.6 #this version of librosa will work just fine
pillow
h5py
sklearn
scikit-image
pyyaml
-e .
-e tfnntools
matplotlib
jupyter
moviepy==0.2.3.5 #newer version will not work
requests
numba
ltfatpy
  • sudo apt install cmake llvm-8
  • After doing these changes, run pip install like this: LLVM_CONFIG=/usr/bin/llvm-config-8 pip install -r requirements.txt
  • For training on a new dataset, keep in mind files should be 16000hz , mono , 16bits . only the first second of each audio file will be used for training
  • Bonus, this script will convert your files:
cd <where_your_wavs_are>
mkdir 16000_mono_16bit
for file in `ls *wav`; do filename=$(basename -- "$file"); sox $file -r 16000 -c 1 -b 16 16000_mono_16bit/$filename ; done
@macramole
Copy link
Author

( these steps will probably work for 20.04 as well )

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant