Easy to use small framework for faster model development and visualization
- clone the repo
- create conda environment (conda env create -f env_specs.yml)
- choose or define a model (
models/your_model.py) - choose or define a dataset_loader
- modify
constants.py - run model
- run tensorboard (tensorboard --logdir=runs) to see the results
- Refactor and improve backbone functionality
- Add examples (MixNet, EfficientNet, gans etc) and improve documentation
- Add multiple models and layers
- Add more image enhancement techniques
- Add or facilitate more complex image visualization methods (matplotlib ... facets)
- Add python "artisan" for easier use