How syedmuhammadubaidurrehman22/crop_disease_identification works
This repository is a research workflow for multi-dataset crop-disease classification with transfer-learning CNNs. The principal path is to prepare image datasets, train a selected model, evaluate its checkpoint, and inspect performance reports or sample predictions. Dataset cleaning, merging, and splitting are separate scripts; the excerpts do not show them orchestrated together. Evaluation and analysis edges reflect shown calls or file reads, while README-level capabilities are retained without assuming unsampled implementation details.
Main entry points: evaluate.py Checkpoint evaluation · train.py Model training · generate_sample_predictions.py Sample predictions · preprocessing.py Image preprocessing
Components 17
Dataset preparation
- Image cleaning
data_scripts/cleaning.py - Dataset merging
data_scripts/merge_datasets.py - Dataset splitting
data_scripts/split_datasets.py - Image preprocessing
data_scripts/preprocessing.py - Plant disease datasets
Model learning
- Model training
models/train.py - CNN model factory
models/model_factory.py - Model checkpoint
Evaluation
- Checkpoint evaluation
models/evaluate.py - Metrics and plots
models/metrics.py - Sample predictions
models/generate_sample_predictions.py - Evaluation outputs
Result analysis
- Class report analysis
models/analyze_class_reports.py - Result collection
models/collect_results.py - Model comparison plot
models/generate_model_comparison_plot.py - Comparison outputs
outputs
Other components
- Researcher
Connections 24
- Researcher → Image cleaning: runs cleaning
data_scripts/cleaning.py - Researcher → Dataset merging: runs merge
data_scripts/merge_datasets.py - Researcher → Dataset splitting: runs split
data_scripts/split_datasets.py - Researcher → Model training: starts training
models/train.py - Researcher → Checkpoint evaluation: starts evaluation
models/evaluate.py - Researcher → Sample predictions: generates samples
models/generate_sample_predictions.py - Plant disease datasets → Image cleaning: supplies images
data_scripts/cleaning.py - Plant disease datasets → Dataset merging: supplies sources
data_scripts/merge_datasets.py - Plant disease datasets → Dataset splitting: supplies classes
data_scripts/split_datasets.py - Plant disease datasets → Image preprocessing: provides splits
data_scripts/preprocessing.py - Model training → Image preprocessing: loads datasets
models/train.py - Model training → CNN model factory: creates model
models/train.py - Model training → Metrics and plots: plots history
models/train.py - Model training → Model checkpoint: saves checkpoint
models/train.py - Checkpoint evaluation → Image preprocessing: loads datasets
models/evaluate.py - Checkpoint evaluation → CNN model factory: creates model
models/evaluate.py - Checkpoint evaluation → Model checkpoint: loads checkpoint
models/evaluate.py - Checkpoint evaluation → Metrics and plots: requests metrics
models/evaluate.py - Checkpoint evaluation → Evaluation outputs: writes evaluation
models/evaluate.py - Sample predictions → CNN model factory: creates model
models/generate_sample_predictions.py - Sample predictions → Model checkpoint: loads checkpoint
models/generate_sample_predictions.py - Evaluation outputs → Class report analysis: reads reports
models/analyze_class_reports.py - Evaluation outputs → Result collection: collects results
- Comparison outputs → Model comparison plot: provides comparison CSV
models/generate_model_comparison_plot.py