RLRS

Installation

• python=3.10
• pip install -r requirements.txt

Install Pytorch

• Follow instruction from the website.
• In my case: conda install pytorch==1.11.0 cudatoolkit=11.3 -c pytorch.

Finally, install the package: pip install -e .

Development

Unit Test

make test

Formatting

make fmt

Loss Visualization

** To run any training, we need to start mlflow first. **

1. Start mlflow server:

mlflow server --host 127.0.0.1 --port 8080

Open 127.0.0.1:8080 to view MLflow, choose Experiments, select the run, click on Model Metrics tab to see all losses, metrics.

Debugging Training

MovieLen training

make mock-train-movie

Ayampp training

make mock-train-food

Model Inference

Start the server

make server

Go to http://127.0.0.1:8000/docs to check how to call API(s).

Evaluation

python scripts/eval.py <config-file> [--verbose]