DeepVisionXplain

Model training environment

Project Description

Neural network training environment (including various MLOps tools) with training API service and model explainability tools.

Conda installation

# clone project
git clone https://github.com/lus105/DeepVisionXplain.git
# change directory
cd DeepVisionXplain
# update conda
conda update -n base conda
# create conda environment and install dependencies
conda env create -f environment.yaml -n DeepVisionXplain
# activate conda environment
conda activate DeepVisionXplain

Quickstart

Train model with default configuration (check if environment is properly set up):

# train on CPU (mnist dataset)
python src/train.py trainer=cpu
# train on GPU (mnist dataset)
python src/train.py trainer=gpu

Other sections

Training API service

CNN/ViT (ante-hoc) explainability

References

lus105/DeepTrainer