Skip to main content

No project description provided

Project description

tf2_aip_base

This repository is designed as a research framework for supervised machine learning. It aims to reduce your work on the train-loop, validation, saving, optimizer, multi-gpu and provides lot more features, which can be configured via command line.

Setup

see: Setup

Usage

To setup your own scenario see: Scenario setup

Running the tutorial scenario:

The default tutorial scenario is 'fashion-mnist'. Run your first training with:

tfaip-train tutorial --trainer_params checkpoint_dir=models/fashion_default

tfaip-train refers to tfaip/scripts/train.py. You can switch to mnist data with --data_params dataset=mnist.

You can evaluate the model on the validation dataset with: tfaip-lav --export_dir models/fashion_default/export

Most hyper parameter can be configured via command line see tfaip-train -h and tfaip-train tutorial -h. Checkout the Wiki for further explanations.

Contributions are welcome, and they are greatly appreciated!

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

tfaip-1.0.1.tar.gz (117.0 kB view hashes)

Uploaded Source

Built Distribution

tfaip-1.0.1-py3-none-any.whl (270.9 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page