- Status
- Finished and delivered. It does what it set out to do and no more work is planned.
- Kind
- Machine Learning
- Started
- Jan 2024

About it
The problem
Recognising emotions from facial images is hard to generalise: models overfit small datasets and perform poorly on faces they have not seen.
The approach
I fine-tuned pretrained models from the TIMM library in PyTorch and used strong augmentation to improve generalisation.
- Pretrained TIMM backbones as a high-quality starting point.
- Augmentation with random resizing, flipping, colour jitter, CutMix and MixUp.
- Careful training and validation splits for honest evaluation.
The result
The classifier generalises better than a model trained from scratch, and the pipeline is a reusable template for image classification with transfer learning.
What it does3
Pretrained Models
Strong Augmentation
Careful Evaluation
More work

Rone Arena API
A public FastAPI for Mobile Legends hero data, analytics and player tools, with a web playground and SDKs.

BeliMadu.com
An online honey shop on Next.js 16 and Cloudflare Workers, with a WhatsApp-checkout cart and a Vitest suite.

ridwaanhall.com
This site: a portfolio and blog on Next.js 16 with an admin, Markdown versions of every page and a CV generated from the About page.
Comments0
Signed-in readers can comment.