Learn practical skills, build real-world projects, and advance your career
Updated 2 years ago
Identifying Flower Species using Deep Learning and PyTorch
Use the "Run" button to execute the code.
# Execute this to save new versions of the notebook
#jovian.commit(project="zero-to-gans-project")
Get some libraries
!pip install numpy matplotlib torch torchvision torchaudio --quiet
Pick a dataset from Kaggle:
- should not be a toy/standard dataset (MNIST, CFAR10, Titanic)
- should be >1000 images
- should have enough variety
- should be possible to build a good model
- should not be too large (< 5Gb)
- exception: work with samples
Dataset candidates
- https://www.kaggle.com/datasets/thedownhill/art-images-drawings-painting-sculpture-engraving
- https://www.kaggle.com/datasets/ma7555/cat-breeds-dataset
- https://www.kaggle.com/datasets/paultimothymooney/breast-histopathology-images
- https://www.kaggle.com/datasets/moltean/fruits (fruits 360)
- https://www.kaggle.com/datasets/olgabelitskaya/flower-color-images
- https://www.kaggle.com/datasets/ikarus777/best-artworks-of-all-time
- https://www.kaggle.com/c/jovian-pytorch-z2g Jovian Kaggle competition, highly recommended:
- sample notebookd show how to do:
- multi-label classification
- label imbalance resolution
- sample notebookd show how to do:
default or fallback:
- Flower recognition: https://www.kaggle.com/datasets/alxmamaev/flowers-recognition