search

ml5js / ml5-library

Friendly machine learning for the web! 🤖
https://ml5js.org
Other
6.46k stars 902 forks source link

Google Colab DCGAN: `!pip install -r requirements.txt` shows several errors #954

Closed Jars-of-jam-Scheduler closed 4 years ago

Jars-of-jam-Scheduler commented 4 years ago

I'm following https://github.com/ml5js/training-dcgan with 64x64 training images, using Google Colab. At the step !pip install -r requirements.txt, these errors are shown:

Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.6/dist-packages (from tensorboard<1.15.0,>=1.14.0->tensorflow==1.14.0->tensorflowjs==1.2.6->-r requirements.txt (line 1)) (1.0.1) Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.6/dist-packages (from tensorboard<1.15.0,>=1.14.0->tensorflow==1.14.0->tensorflowjs==1.2.6->-r requirements.txt (line 1)) (3.2.1) Building wheels for collected packages: chainer, absl-py Building wheel for chainer (setup.py) ... done Created wheel for chainer: filename=chainer-6.2.0-cp36-none-any.whl size=880596 sha256=fc47b6c96380bc78f0c93fd992ba6d49215464e6b691c2492ba0852ed0179db7 Stored in directory: /root/.cache/pip/wheels/2e/be/c5/6ee506abcaa4a53106f7d7671bbee8b4e5243bc562a9d32ad1 Building wheel for absl-py (setup.py) ... done Created wheel for absl-py: filename=absl_py-0.7.1-cp36-none-any.whl size=117848 sha256=6aeeaf140c33f1308eca2e8595b73e736e8b54af0b0d7e234889c743c476f751 Stored in directory: /root/.cache/pip/wheels/ee/98/38/46cbcc5a93cfea5492d19c38562691ddb23b940176c14f7b48 Successfully built chainer absl-py ERROR: umap-learn 0.4.2 has requirement numpy>=1.17, but you'll have numpy 1.16.4 which is incompatible. ERROR: google-colab 1.0.0 has requirement six~=1.12.0, but you'll have six 1.11.0 which is incompatible. ERROR: datascience 0.10.6 has requirement folium==0.2.1, but you'll have folium 0.8.3 which is incompatible. ERROR: albumentations 0.1.12 has requirement imgaug<0.2.7,>=0.2.5, but you'll have imgaug 0.2.9 which is incompatible. Installing collected packages: numpy, six, h5py, keras, protobuf, absl-py, tensorboard, tensorflow-estimator, tensorflow, tensorflow-hub, tensorflowjs, pillow, chainer, cupy-cuda100 Found existing installation: numpy 1.18.4 Uninstalling numpy-1.18.4: Successfully uninstalled numpy-1.18.4 Found existing installation: six 1.12.0 Uninstalling six-1.12.0: Successfully uninstalled six-1.12.0

These errors don't seem to interrupt the execution of the command (!pip install -r requirements.txt).

Additionnaly, the following warning is shown:

Found existing installation: chainer 6.5.0 Uninstalling chainer-6.5.0: Successfully uninstalled chainer-6.5.0 Successfully installed absl-py-0.7.1 chainer-6.2.0 cupy-cuda100-6.2.0 h5py-2.8.0 keras-2.2.4 numpy-1.16.4 pillow-6.1.0 protobuf-3.7.1 six-1.11.0 tensorboard-1.14.0 tensorflow-1.14.0 tensorflow-estimator-1.14.0 tensorflow-hub-0.5.0 tensorflowjs-1.2.6 WARNING: The following packages were previously imported in this runtime: [PIL,google,numpy,six] You must restart the runtime in order to use newly installed versions.

How to reproduce the problem?

Read the following (or just follow https://github.com/ml5js/training-dcgan):

  1. Create and configure your Google Colab: Activate GPU (GPU activation is optional)

  2. Execute:

    from google.colab import drive
drive.mount('/content/drive/', force_remount=True)

  3. Create your Google Drive directory in which you will store the GitHub repository "https://github.com/noisyneuron/training-dcgan"

  4. Execute:

    %cd '/content/drive/......................XYZ..........XYZ...'
!git clone https://github.com/noisyneuron/training-dcgan
%cd training-dcgan
!pip install -r requirements.txt
joeyklee commented 4 years ago

Thanks @Jars-of-jam-Scheduler for your documentation here. This issue would be better placed in the https://github.com/ml5js/training-dcgan repo.

The errors you have provided would indicate that there are issues with your python dependencies that can be a result of a many number of things which include conflicting packages and version numbers, platform issues and more. @noisyneuron might be able to speak more to these issues.

As mentioned in https://github.com/ml5js/ml5-library/issues/953 DCGAN training is not currently a major focal point of our development and with tensorflow changing so rapidly, we've not yet had time to adapt to the latest updates. I'm fine to keep this issue open, but likely we won't take further action on resolving this soon due to other priorities in the library. Sorry!

joeyklee commented 4 years ago

Unless there are further questions on this I will close this now. Feel free to comment as needed.