zademn/mnist-mlops-learning

In this project I played with mlflow, streamlit and fastapi to create a training and prediction app on digits

Hello there, I'm Dacian Stroia, also known as Zademn in the online world. I'm an applied researcher in ML and cryptography. I like GNNs, FHE and Rust.

Projects 📦

• facto-rs - Small factorization library in Rust
• A ML monitoring PoC
• Small Rust projects: Game of life in Rust, Ray tracing in one weekend, AoC2021, CryptoPals - partially
• Small Zig projects: Ray tracing in one weekend
• Detoxitweet - An extension for classifying the toxicity of the tweets (team of 3 people, hackaton).

Self taught notes ✒️

• General cryptography
• Network science and GNN labs - I currently teach them at my university
• My crust of rust implementations

University projects and notes 🏫

• PUGCN paper reimplemented in Pytorch + some extra models I tried out.
• KMeans in Rust
• FBNet reimplemented in Pytorch Geometric as practice.
• Server and Client - phone app that allows you to track friend locations in real time (team of 2 people).
• Healthcheck app - An app that allows you to do healthchecks for different devices (team of 2 people).
• AtomC-compiler - A compiler for a C subset.
• omnet++ schedulers - Some schedulers implemented in omnet++.

mnist-mlops-learning's Issues

Fix for docker-compose sqlite volume mount path

Hi, nice project and kudos for sharing it !
I wanted to play with your code using the docker-compose file, however it was not working for freshly created project.

Here is what I changed to make it work (for both mlflow and fastapi services):

    volumes:
      - ./db:/db

instead of path to the database file.