How sifat192/bean-leaf-transfer-learning-pytorch works
This repository is a runnable notebook project comparing a custom CNN with two GoogLeNet transfer-learning strategies on bean-leaf images. The user runs the notebook, which obtains the dataset, preprocesses and batches images, trains the models, evaluates validation performance, and presents plots and comparisons. The README documents these stages and capabilities; the notebook itself was not excerpted, so internal call-level relationships cannot be verified.
Main entry points: bean_leaf_transfer_learning.ipynb Experiment notebook
Components 11
Data pipeline
- Bean leaf dataset
- Image preprocessing
notebooks/bean_leaf_transfer_learning.ipynb - ImageFolder loaders
notebooks/bean_leaf_transfer_learning.ipynb
Model strategies
- Custom CNN
notebooks/bean_leaf_transfer_learning.ipynb - GoogLeNet variants
notebooks/bean_leaf_transfer_learning.ipynb
Training and evaluation
- Experiment notebook
notebooks/bean_leaf_transfer_learning.ipynb - Model training
notebooks/bean_leaf_transfer_learning.ipynb - Validation metrics
notebooks/bean_leaf_transfer_learning.ipynb - Learning curves
notebooks/bean_leaf_transfer_learning.ipynb - Model comparison
notebooks/bean_leaf_transfer_learning.ipynb
Other components
- Notebook user
Connections 10
- Notebook user → Experiment notebook: runs
README.md - Experiment notebook → Bean leaf dataset: downloads
README.md - Experiment notebook → Image preprocessing: applies
README.md - Experiment notebook → ImageFolder loaders: loads images
README.md - Experiment notebook → Custom CNN: defines baseline
README.md - Experiment notebook → GoogLeNet variants: configures variants
README.md - Experiment notebook → Model training: trains models
README.md - Experiment notebook → Validation metrics: evaluates
README.md - Experiment notebook → Learning curves: plots metrics
README.md - Experiment notebook → Model comparison: compares results
README.md