Updated environment, minor improvements

• Update Docker image dokai:23.10 and Python requirements.
• Add doi to citation and readme.
• Readme text improvements.
• Fix flake8 and mypy errors.

Initial release

Solution

This repository contains the winning solution for the Sensorium 2023 Competition. The competition aimed to predict the activity of neurons in the primary visual cortex of mice in response to videos. The proposed solution includes a novel model architecture called DwiseNeuro and a training pipeline with a solid cross-validation strategy and knowledge distillation.

Key points:

• DwiseNeuro - novel model architecture for prediction neural activity in the mouse primary visual cortex.
• Solid cross-validation strategy with splitting folds by perceptual video hash.
• Training on all mice with an option to fill unlabeled samples via distillation.