Shanice

Shanice

@swillcodes.bsky.social

Twitch: https://www.twitch.tv/swillcodes?sr=a

41 followers43 following45 posts

No streams so far this week while I work through module 3. There is far to many keys that I don't want to risk sharing 😅 Should be done tomorrow, so maybe Friday. Will post here when confirmed.

Another module in the books! This time experimenting with serverless deployment using AWS Lambda. Scikit-learn, Keras and Pytorch models, all implemented with Lambda functions, pushed as container images to AWS ECR, and created within AWS Lambda #MLzoomcamp

Been quietly working on this, limited time with my PC out of commission for a week, but we got there. Midterm Project - containerised and deployed! Predicting Type II Diabetes in Women who with history of Gestational Diabetes. #mlzoomcamp @DataTalksClub

Another module in the books in #mlzoomcamp. This module centered on decision trees and ensemble methods such as Random Forests and XGBoost! It is so easy to over fit a decision tree!

View full thread

2️⃣ FastAPI + Pydantic gotchas My first time really diving into FastAPI and Pydantic together. I now see why they go together so well. That validation told be everything I needed to know 😅

3️⃣ Dockerfiles and missing ports Again getting to grips with Dockerfiles and digging through some documentation came in clutch! I'm quite glad I organised the project the way I did. Without it I would have missed out on these learning opportunities! 💪

1️⃣ Import pain & Python packaging quirks Ran into `ModuleNotFoundError: No module named 'src'` errors when running the project with nested directories. The fix? Run modules with `python -m src.pipeline.predict` Adding __init__.py files alone did not help☠️

Module 4 homework? Completed ✅ Accuracy is not all it seems 👀 So good to cement my understanding of binary classification evaluation metrics! #mlzoomcamp #DataTalksClub

Another homework in the books! #mlzoomcamp has got me in curiosity mode. I've tried to consciously question the choices we make when selecting models or metrics for examining feature importance. Looking forward to carrying this into model evaluation week! #machinelearning #DataTalksClub

You are up to date.