How singlezombie/dl-demos works
DL-Demos is a collection of runnable deep-learning demonstrations, not one end-user application. Its principal workflow is per-demo: prepare data, instantiate a model, train/evaluate it, then produce predictions, generated samples, or visualizations. The graph preserves the README’s major learning and generative-model branches. README and source-index references support many component links; unexcerpted demos are represented by their exact directory paths, and their internal wiring is left unasserted where it was not shown.
Main entry points: AdvancedOptimizer/main.py Optimizer variants · ddpm/main.py DDPM · pixelcnn/main.py PixelCNN · ddim/main.py DDIM
Components 25
Learning demos
- Logistic regression
dldemos/LogisticRegression/main.py - Shallow network
dldemos/ShallowNetwork - Deep network
dldemos/DeepNetwork - Initialization
dldemos/Initialization/main.py - Regularization
dldemos/Regularization/main.py - Optimizer variants
dldemos/AdvancedOptimizer/main.py - Multiclass classification
dldemos/MulticlassClassification - Distributed training
dldemos/PyTorchDistributed/main.py
Vision demos
- Basic CNN
dldemos/BasicCNN - ResNet
dldemos/ResNet/tf_main.py - NMS and boxes
dldemos/nms/nms.py - YOLO data demo
dldemos/MyYOLO/load_coco.py
Generative models
Sequence models
- Letter language model
dldemos/BasicRNN/main.py - Sentiment analysis
dldemos/SentimentAnalysis/main.py - Attention translation
dldemos/attention/main.py - Transformer translation
dldemos/Transformer
Shared runtime
- Activation functions
dldemos/utils/function.py - LMDB loader
dldemos/lmdb_loader.py
Other components
- Demo user
Connections 34
- Demo user → Logistic regression: runs demo
- Demo user → Shallow network: runs demo
- Demo user → Deep network: runs demo
- Demo user → Initialization: runs demo
- Demo user → Regularization: runs demo
- Demo user → Optimizer variants: runs demo
dldemos/AdvancedOptimizer/main.py - Optimizer variants → Deep network: loads data
dldemos/AdvancedOptimizer/main.py - Optimizer variants → Activation functions: uses activations
- Initialization → Activation functions: uses activations
dldemos/Initialization/main.py - Regularization → Activation functions: uses activations
dldemos/Regularization/main.py - Demo user → Multiclass classification: runs demo
- Demo user → Basic CNN: runs demo
- Demo user → ResNet: runs demo
- Demo user → NMS and boxes: runs demo
- Demo user → YOLO data demo: runs demo
- YOLO data demo → NMS and boxes: uses boxes
dldemos/MyYOLO/load_coco.py - Demo user → VAE: runs demo
- Demo user → DDPM: runs demo
dldemos/ddpm/main.py - DDPM → DDPM: samples images
dldemos/ddpm/main.py - Demo user → PixelCNN: runs demo
dldemos/pixelcnn/main.py - PixelCNN → PixelCNN: samples images
dldemos/pixelcnn/main.py - Demo user → VQVAE: runs demo
dldemos/VQVAE/main.py - VQVAE → LMDB loader: loads data
- Demo user → DDIM: runs demo
dldemos/ddim/main.py - DDIM → DDIM: samples images
dldemos/ddim/main.py - Demo user → Style transfer: runs demo
- Demo user → Letter language model: runs demo
dldemos/BasicRNN/main.py - Letter language model → Letter language model: trains model
dldemos/BasicRNN/main.py - Demo user → Sentiment analysis: runs demo
dldemos/SentimentAnalysis/main.py - Sentiment analysis → Sentiment analysis: trains model
dldemos/SentimentAnalysis/main.py - Demo user → Attention translation: runs demo
dldemos/attention/main.py - Attention translation → Attention translation: evaluates model
dldemos/attention/main.py - Demo user → Transformer translation: runs demo
- Demo user → Distributed training: runs demo