-1

I am new to Machine Learning, but I have basic knowledge in programming.

I have just completed fast.ai course and I know how to use Google Colab.

But I still have no idea on how to get this project work on Colab.

I have just created a new notebook, and set up the environment,

I run 

from google.colab import drive
drive.mount('/content/drive')
import keras
!git clone https://github.com/andabi/deep-voice-conversion

in my notebook, it copies all files to my google drive, but I do not know what is the next step.

eyllanesc
  • 235,170
  • 19
  • 170
  • 241
Grace
  • 9
  • 4
  • Mount your Google drive on Colab. https://towardsdatascience.com/downloading-datasets-into-google-drive-via-google-colab-bcb1b30b0166 – rayryeng Jun 02 '19 at 16:07
  • Thank you for answering me! Yes I have done `from google.colab import drive drive.mount('/content/drive')` so I have copied files from Github to `/content/drive/My Drive/Colab Notebooks/` – Grace Jun 02 '19 at 16:10
  • .... So what's the problem? – rayryeng Jun 02 '19 at 16:11
  • @rayryeng she said: next step?! :D next step is import libraries and your code for processing your mounted data probably – Mario Jun 02 '19 at 16:15
  • I have a bunch of .py in `/content/drive/My Drive/Colab Notebooks/` now. – Grace Jun 02 '19 at 16:16
  • `audio.py eval1.py materials requirements.txt utils.py convert.py eval2.py models.py scripts data_load.py hparam.py modules.py tensorpack_extension.py datasets hparams notes train1.py docker LICENSE README.md train2.py` Can you tell me how to use them? – Grace Jun 02 '19 at 16:16
  • @Grace Welcome to Stack Overflow, please read [How do I ask a good question](https://stackoverflow.com/help/how-to-ask)? and [How to create a Minimal, Reproducible Example](https://stackoverflow.com/help/minimal-reproducible-example). Please update your question and **minimal it to core problem**. – Mario Jun 02 '19 at 16:17
  • @Mario Thank you. can u give me any hint on how to import libraries? – Grace Jun 02 '19 at 16:19
  • @Mario Thank you Mario, I have just read it. and I have `import keras` already. What should I do next? Thank you. – Grace Jun 02 '19 at 16:24
  • Maybe you should figure out how to run this on your machine first, then get it going on Colab. The README on that repo is a good start. It tells you what to do after you clone the repo. – rayryeng Jun 02 '19 at 16:33

1 Answers1

0

Based on the Requirements in GitHub you should install and import following libraries not only Keras (actually it hasn;t mentioned you need Keras! it's mentioned Tensorflow clearly nevertheless you can use Keras if you modify the codes):

  • Tensorflow >= 1.1
  • Numpy >= 1.11.1
  • Librosa == 0.5.1

You can follow these videos tp learn how you install libraries in colab less than 2 mins:

Install TensorFlow 2.0 with GPU in Google CoLab

Getting Started with TensorFlow in Google Colaboratory (Coding TensorFlow)

# start import libraries and follow next steps
from google.colab import drive
drive.mount('/content/drive')
import numpy as np
import tensorflow as tf
import keras
import liborsa

...

Procedure:

  • Train phase: Net1 and Net2 should be trained sequentially.
    • Train1(training Net1)
      • Run train1.py to train and eval1.py to test.
    • Train2(training Net2)
      • Run train2.py to train and eval2.py to test.
        • Train2 should be trained after Train1 is done!

  • Convert phase: feed forward to Net2

    • Run convert.py to get result samples.
    • Check Tensorboard's audio tab to listen the samples.
    • Take a look at phoneme dist. visualization on Tensorboard's image tab.
      • x-axis represents phoneme classes and y-axis represents timesteps
      • the first class of x-axis means silence.

Note Don't forget to have a look to Readme.md file which gives you good vision.

Mario
  • 1,631
  • 2
  • 21
  • 51
  • Thank you very much Mario, the videos are very helpful. I will learn these steps first and come back later. ;) – Grace Jun 02 '19 at 16:53
  • The OP does not require installing Tensorflow 2.0 on Colab. Tensorflow 1.x should be fine. – rayryeng Jun 02 '19 at 16:55
  • @rayryeng version is not matter of OP's request, she needed to learn how she can install any libraries in Colab regardless of version! she needed at first to get familiar with those steps. She got a bit confused by copies of many files from GitHub I hope she's thankful by my hint. – Mario Jun 02 '19 at 18:48
  • @rayryeng since you're into **Deep Learning** , may I draw your attention to my [question](https://stackoverflow.com/questions/56344611/how-can-take-advantage-of-multiprocessing-and-multithreading-in-deep-learning-us?) with **bounty**. – Mario Jun 02 '19 at 18:55