A hands-on repository dedicated to building modern deep learning layers, models and tasks from scratch using PyTorch.
.. toctree:: :caption: Pytorch Basics :maxdepth: 1 Tensors <pytorch-basics/Tensors> Automatic Differentiation <pytorch-basics/AutoGrad> Modules <pytorch-basics/Modules> Dataset and Dataloader <pytorch-basics/DataLoading> Save and Load <pytorch-basics/SaveAndLoad>
.. toctree:: :caption: Pytorch Advanced :maxdepth: 1 Distributed Training <pytorch-advanced/Distributed> TensorBoard <pytorch-advanced/TensorBoard> AutoMixPrecision <pytorch-advanced/AutoMixPrecision>
.. toctree:: :caption: ConvNets :maxdepth: 1 VGG <models/convnets/VGG> GoogLeNet <models/convnets/GoogLeNet> ResNet <models/convnets/ResNet> MobileNet <models/convnets/MobileNet> ConvNeXt <models/convnets/ConvNeXt> MLPMixer <models/convnets/MLPMixer> ConvMixer <models/convnets/ConvMixer>
.. toctree:: :caption: Vision Trasnformer :maxdepth: 1 ViT：VisionTrasnformer <models/vision-transformers/ViT.ipynb> MAE: Masked AutoEncoder <models/vision-transformers/MAE.ipynb> SwinTrasnformer <models/vision-transformers/Swin.ipynb> CvT: Convolutional vision Transformer <models/vision-transformers/CvT.ipynb> DiNAT: Dilated Neigborhood Attention Transformer <models/vision-transformers/DiNAT.ipynb> MobileViT <models/vision-transformers/MobileViT.ipynb> DETR: DEtection TRansformer <models/vision-transformers/DETR.ipynb> MaskFormer: Mask Classification-based Segmentation Transformer <models/vision-transformers/MaskFormer.ipynb> OneFormer: Universal Architecture For Segmentation Tasks <models/vision-transformers/OneFormer.ipynb>
.. toctree:: :caption: Compute Vision Tasks :maxdepth: 1 COCO Dataset <cv-tasks/datasets/coco> PascalVOC Dataset <cv-tasks/datasets/pascal_voc> YOLO <cv-tasks/object_detection/yolo> Transfer Learning ViT Image Classifiction <cv-tasks/classification/transfer-learning-image-classification> LoRA Fine-tuning Vision Transformer <cv-tasks/classification/LoRA-Image-Classification> Fine-Tuning Multilable Image Classification <cv-tasks/classification/fine-tuning-multilabel-image-classification> Knowledge Disstillation on ViT Image Classification <cv-tasks/KnowledgeDistillation>
.. toctree:: :caption: OpenAI Triton Kernels :maxdepth: 1 Vector Add <triton/vector-add> Fused Softmax <triton/fused-softmax> Matrix Multiplication <triton/matrix-multiplication> Low Memory Dropout <triton/low-memory-dropout> LayerNorm <triton/layer-norm>