Pipeline mind map
Standard pipeline template for a images competition. Not yet specialised by an editor.
Data processing
- Inspect images and labels
- Resize / normalise
- Augmentation
Model design
- Pretrained CNN or ViT backbone
- Fine-tune the head first
Training
- Stratified folds
- Mixed precision
- Learning-rate schedule
Post-processing
- Test-time augmentation
- Threshold per metric
- Ensemble folds
Metric
Categorization Accuracy. Accuracy: share of rows predicted exactly right.
Public baselines
Most-voted public notebooks, refreshed daily from the Kaggle API (last 2026-10-10).
- Introduction to CNN Keras - 0.997 (top 6%) ▲16567
- Simple MNIST NN from scratch (numpy, no TF/Keras) ▲8757
- Convolutional Neural Network (CNN) Tutorial ▲5295
- Pytorch Tutorial for Deep Learning Lovers ▲5218
- Deep Neural Network Keras way ▲3852
- 25 Million Images! [0.99757] MNIST ▲3467
- Knowledge Graph & NLP Tutorial-(BERT,spaCy,NLTK) ▲2899
- Interactive Intro to Dimensionality Reduction ▲2825
- TensorFlow deep NN ▲2615
- A Beginner's Approach to Classification ▲2491
Discussion 0
No discussion yet. Ask the first question or post a team-up for this competition.
Sign in with PocketPlay
Sign-in opens soon. Until then the forum is read-only.
Same account as pocketplay.win (Google or username). Posting needs a Google-linked account.
Markdown: **bold**, *italic*, `code`, ``` code blocks, - lists, > quotes, [text](https://link). No HTML.