-
Client was built from scratch! Details about this project, including scope, stack, etc. can be found on the details page.
-
Math details can be found on the math page.
-
Model details can be found on the models page.
-
Finally, browse some results on the gallery page.
The client is a react app built with yarn:
cd client
yarn install && yarn dev
By default, the client runs on port 5173
.
To run the server, create a python virtual environment and install the dependencies in requirements.txt
. Then, in the root directory, run
python3 run.py
By default, the server runs on port 3000
. The only function of the server is to parse requests from the client and submit to our model endpoint (see below).
The client and flask server is deployed on Render.
All models are deployed on ModelBit, and all model inference requires around 4GB RAM. The deployment script is very simple and lives in modelbit.ipynb
. One thing to note when deploying is that the python package image-to-image-translation-server
must be self-contained, i.e., it will not have access to a .env
. This means that server/config.py
must be modified to include actual values rather than what is included right now.
- finish the remainder of the models
- update the gallery with actual results