How adhammostafaa/ae-mega-project-team-2 works
The visible runtime is a command-line generator for synthetic object-detection images and YOLO labels. Its workflow creates backgrounds and labeled shapes, composes randomized images, then writes split image/label files and dataset metadata. The README also describes a separate Wikipedia-to-text-generation Transformer pipeline, including training, generation, and attention visualization, but no corresponding source file appears in the supplied tree; that branch is therefore shown only at the documented level. The RAR archive’s contents are unknown.
Main entry points: generate_odlc_dataset.py Dataset CLI
Components 14
ODLC Dataset Generator
- Dataset CLI
generate_odlc_dataset.py - Background maker
generate_odlc_dataset.py - Object composer
generate_odlc_dataset.py - Shape drawer
generate_odlc_dataset.py - Image splits
- YOLO labels
- Dataset metadata
Text Generation Pipeline
- Wikipedia corpus
- Corpus preprocessing
- Causal Transformer
- Model training
- Text generation
- Attention heatmap
Other components
- Dataset user
Connections 12
- Dataset user → Dataset CLI: runs
generate_odlc_dataset.py - Dataset CLI → Background maker: creates
generate_odlc_dataset.py - Dataset CLI → Object composer: creates
generate_odlc_dataset.py - Object composer → Shape drawer: draws
generate_odlc_dataset.py - Dataset CLI → Image splits: writes
generate_odlc_dataset.py - Dataset CLI → YOLO labels: writes
generate_odlc_dataset.py - Dataset CLI → Dataset metadata: writes
generate_odlc_dataset.py - Wikipedia corpus → Corpus preprocessing: provides text
README.md - Corpus preprocessing → Causal Transformer: provides blocks
README.md - Causal Transformer → Model training: trains
README.md - Model training → Text generation: produces model
README.md - Causal Transformer → Attention heatmap: provides attention
README.md