kaust-academy / tensorflow-gpu-data-science-project Goto Github PK View Code? Open in Web Editor NEW 10.0 7.0 5.0 173 KB Template repository for a Python 3-based (data) science project with GPU acceleration using the TensorFlow ecosystem. License: BSD 3-Clause "New" or "Revised" License Dockerfile 32.62% Shell 67.38% template-project data-science scientific-computing scientific-workflows scientific-research python-3 gpu-acceleration gpu tensorflow deep-learning Introduction ยท People ยท Discuss
JupyterLab-Tensorboard integration Would be nice to have JupyterLab-TensorBoard integration. There are a couple of related projects for this. https://github.com/chaoleili/jupyterlab_tensorboard https://github.com/lspvic/jupyter_tensorboard Looks like both are needed for the integration to work properly.
Add Horovod dependencies for easy support for distributed training Need to install Horovod and all necessary dependencies. https://github.com/horovod/horovod/blob/master/docs/gpus.rst Looks like you can install everything except horovod itself using conda.
Hyper-parameter optimization Consider installing keras-tuner so make it easier to perform hyper-parameter optimization for tf.keras models. https://github.com/keras-team/keras-tuner
Check to see whether TensorRT is included in the docker image (if not add it!) Not sure whether the TensorRT library is included in the base image provided by NVIDIA. If it is not included then it makes sense to add it. Unfortunately the library does not seem to be available from pip or conda.
Experiment with one-cycle learning rate scheduling See relevant sections of Chapter 11 notebook... https://github.com/ageron/handson-ml2/blob/master/11_training_deep_neural_networks.ipynb
Consider adding TFT dependency Tensorflow Transform is a preprocessing library that looks interesting... https://www.tensorflow.org/tfx/guide ...need to better understand who TFT relates to Apache Beam. https://beam.apache.org/
Vue.js ๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
javascript JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Machine learning Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Facebook We are working to build community through open source technology. NB: members must have two-factor auth.