GitHub - RJT1990/mantra: A high-level, rapid development framework for machine l...

README.md

Mantra is a deep learning development kit that manages the various components in an deep learning project, and makes it much easier to do routine tasks like training in the cloud, model monitoring, model benchmarking and more. It works with your favourite deep learning libraries like TensorFlow, PyTorch and Keras.

You might like mantra if:

• You need a way to structure your deep learning projects: versioning, monitoring and storage of models and data.
• You need boring devops tasks like cloud integration and file syncing between cloud/local taken care for you.
• You need a way to easily compare and evaluate your model against benchmark tasks, e.g. accuracy on CIFAR-10.

This is an very early alpha release: we'd love your comments on the general concept and whether you find it useful. Issues and pull requests are also welcome. If the project gets enough love, we'll work towards a stable release!

You can read the docs here.

Get Started

? To launch your first Mantra project, execute the following to create a new project directory:

mantra launch my_project 

☁ Configure your cloud settings and API keys:

cd my_project; mantra cloud 

? Get the example datasets and models from here:

mantra import https://github.com/RJT1990/mantra-examples.git

? Here are the example models you can train:

mantra train relativistic_gan --dataset decks --cloud --dev --image-dim 256 256

mantra train log_reg --dataset epl_data --target home_win --features feature_1 feature_2 feature_3 

? During training, you can spin up the Mantra UI:

mantra ui

Installation

To install mantra, you can use pip:

pip install mantraml

You should also have TensorFlow or PyTorch installed depending on which framework you intend to use.

Mantra is tested on Python 3.5+. It is not currently supported on Windows, but we'll look to get support in the near future.

AWS Dependencies

You will need to install AWS CLI as a dependency.

1. Login to AWS through a browser, click your name in the menubar and click My Security Credentials.

2. Create a new Access Key and make a note of the Access Key ID and Secret Access Key.

3. From terminal enter the following:

johnsmith@computer:~$ pip install awscli
johnsmith@computer:~$ aws configure

Once prompted, enter your AWS details and your default region (e.g. us-east-1).

1. Now your credentials will be accessible by the boto3 AWS SDK library, which will allow Mantra to be used to provision cloud instances on your request.

2. Use mantra cloud from your mantra project root to configure your cloud settings.

You should also ensure you are happy with the default instance settings in mantra - you can check this in the settings.py file in your project root.

Have Fun

Please train responsibly.