001
Breast Cancer Cell MoA Classification
ResNet50 transfer learning model trained on the BBBC021 dataset to classify 5 mechanism-of-action classes in cancer cell microscopy images. Includes Grad-CAM visual explainability and is fully deployed on Streamlit Cloud with auto model loading from Hugging Face.
PyTorch
ResNet50
Grad-CAM
Hugging Face
Streamlit
Transfer Learning
87.34%Accuracy
0.88Macro F1
5MoA classes
002
Sentiment Analysis Dashboard
NLP dashboard powered by Cardiff NLP's twitter-roberta-base-sentiment-latest model. Supports single text and batch CSV analysis with real-time confidence scores, interactive Plotly visualisations, and a polished modern UI.
Transformers
RoBERTa
Plotly
Streamlit
HuggingFace
NLP
RoBERTaModel
BatchCSV support
LiveDeployed
003
Object Detection App
Real-time object detection web app using YOLOv8n capable of detecting 80 different object classes from uploaded images. Features bounding boxes with confidence scores, NMS support, an interactive results table, and a Plotly bar chart — all wrapped in a dark modern UI.
YOLOv8
PyTorch
OpenCV
Streamlit
Plotly
COCO Dataset
80Object classes
YOLOv8nModel
NMSDedup support
LiveDeployed