https://github.com/stofte/tf-playground
Random tensorflow
https://github.com/stofte/tf-playground
jupyter-notebook machine-learning python3 tensorflow windows
Last synced: about 1 month ago
JSON representation
Random tensorflow
- Host: GitHub
- URL: https://github.com/stofte/tf-playground
- Owner: stofte
- Created: 2017-09-12T12:29:12.000Z (over 8 years ago)
- Default Branch: master
- Last Pushed: 2018-06-06T20:37:51.000Z (about 8 years ago)
- Last Synced: 2025-02-01T02:18:43.538Z (over 1 year ago)
- Topics: jupyter-notebook, machine-learning, python3, tensorflow, windows
- Language: Jupyter Notebook
- Size: 11.4 MB
- Stars: 0
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: readme.md
Awesome Lists containing this project
README
ML/Tensorflow playground (Windows/python36)
===============================
Recommended to use Anaconda for managing installation of components. Repo tested with Python 3. Currently [WSL does not support exposing GPU](https://github.com/Microsoft/WSL/issues/1788), limiting what use can be done usefully in the WSL Bash console on Windows. Windows/x64 specific Tensorflow [wheels](https://www.python.org/dev/peps/pep-0427/) for [Python can be found here](https://github.com/fo40225/tensorflow-windows-wheel).
MNIST dataset
- [Softmax regression](src/mnist_softmax.ipynb)
- [Convolutional neural net](src/mnist_convnn.ipynb)
- [Convnet using Keras](src/mnist_convnet_keras.ipynb)
[OpenAI Gym](https://gym.openai.com/)
- [CartPole](src/gym_cartpole.ipynb)
Software/library versions:
- Windows 10 (v1709)
- CUDA 9.1.85
- python 3.6.4
- anaconda 4.4.10
- jupyter 1.0
- tensorflow 1.7
- gym 0.10.5
- [GraphViz 2.38](https://graphviz.gitlab.io/_pages/Download/Download_windows.html) (Keras visualization)
- pyglet 1.2.4 (for `gym`)
Notes:
- Measure perf on Windows: `powershell -Command "Measure-Command { python script.py | Out-Default}"`