Quadra: Effortless and reproducible deep learning workflows with configuration files.

Actually Robust Training - Tool Inspired by Andrej Karpathy "Recipe for training neural networks". It allows you to decompose your Deep Learning pipeline into modular and insightful "Steps". Additionally it has many features for testing and debugging neural nets.

End-to-End Speech Recognition Training with Conformer CTC using PyTorch Lightning⚡

A ⚡️ Lightning.ai ⚡️ component for train and test data drift detection

A computer vision pipeline built with PyTorch for advanced NBA analytics. The system uses YOLO for player/ball detection, multi-object tracking, and a zero-shot classifier for team affiliation, and court key point detection to create a tactical top-down view and calculate real-world metrics like speed, distance, and passes.

Leverage ModernGLiNER's capabilities using LitServe.

Leverage ModernBERT's capabilities using LitServe.

Leverage TangoFlux's text-to-audio capabilities using LitServe.

Leverage Reader-LM's capabilities using LitServe.

Leverage GOT-OCR2's optical character recognition capabilities using LitServe.

Leverage Gemma 3's capabilities using LitServe.

Leverage VideoLLaMA 3's capabilities using LitServe.

Leverage Aya Vision's capabilities using LitServe.

Leverage SANA's capabilities using LitServe.

Leverage PaliGemma 2's DOCCI fine-tuned variant capabilities using LitServe.

Leverage SigLIP 2's capabilities using LitServe.

Leverage PaliGemma 2 mix model variant capabilities using LitServe.

FungiScan is a mobile mushroom identification app that uses a cloud-hosted Vision Transformer (ViT) model for real-time camera or gallery identification, with results and species data stored locally for offline use. The app also features an interactive OSM layer integrated with the iNaturalist API, allowing users to explore nearby forageables.

Leverage Kokoro's TTS capabilities using LitServe.

Leverage Dragonfly-Med's capabilities using LitServe.