ColdRice

ML & DL framework using dask for its parallel computing, numpy, eagerpy & jax for its backend and numba for its JIT and CUDA support.

Development Setup

• Create an environment: conda create -n coldrice
• Create the environment using the coldrice.yml file: conda env create -f coldrice.yml
• Activate coldrice environment: conda activate coldrice
• Update the environment using the coldrice.yml file: conda env update -f coldrice.yml --prune
• Export your active environment to coldrice.yml file: conda env export | grep -v "^prefix: " > coldrice.yml

⚠️ You can't create an environment if the environment was exported on a different platform than the target machine.

ℹ️ coldrice.yml was exported on Linux.

Production Usage

$ pip install coldrice

TODOs

• Computational Graph for Gradient Descent
• Add ML Toolkit
• Visualization
• Architecture Search
• CUDA Support
• Complete Production Usage Guide(Docker Setup)