Python scripts for performing stereo depth estimation using the HITNET Tensorflow model.

Pytorch implementation of Paper by Google Research - Representation Learning for Information Extraction from Form-like Documents.

Python implementation of the video stabilizing algorithm used in YouTube

First open-source implementation of Google TurboQuant (ICLR 2026) -- near-optimal KV cache compression for LLM inference. 5x compression with near-zero quality loss.

This is a working demo of the pegasus summarization model trained on cnn_dailymail

Generate fixed dimensional embeddings for multi-dimensional vectors in python based on Muvera from Google.

Multi-encoder segmentation for contrail detection in satellite imagery | Google Researc

Pytorch Lightning implementation of Vision Transformer with support for loading checkpoints saved in official Flax implementation.

Your compass for the JAX multiverse. 🌌 A curated, up-to-date map of the best libraries, state-of-the-art research, and high-performance tools

A TensorFlow implementation of NVIDIA's PilotNet Neural Network

A Pytorch implementation of the MixMatch algorithm developed by google-research.

Secure FedRAG framework for distributed health data search and knowledge exchange

Visual animated walkthroughs of the DeepMind "Titans: Learning to Memorize at Test Time" paper using Manim, aimed at making complex ML concepts accessible.

Application of imitation learning to football (soccer) video game.

End-to-end medical AI platform with MedGemma specialist models, Claude orchestration, and explainable AI reports.

Derived from Medical Literature Development: Injecting domain expertise into T5 via precision vocabulary-guided masking. 源自医学文献开发：通过精准词库引导遮蔽，为 T5 注入领域专业知识。医学文献開発に基づく実装：高精度な語彙ガイド付きマスキングにより、T5モデルにドメイン専門知識を注入。(機械翻訳)

🌟 Enhance T5 model performance with domain-specific word masking for improved text processing and data refinement.

First open-source LoRA fine-tuning of BanglaSenti on XLM-RoBERTa-base for Bengali sentiment analysis, trained with Google Cloud TPUs. Includes code, configs, and reproducible results.

Node.js/TypeScript reimplementation of Google Research's TimesFM 2.5 — a decoder-only foundation model for zero-shot time-series forecasting. Built on ONNX Runtime.

Multi-language AI-powered flight price monitoring dashboard with TimesFM forecasting, real-time data, and global airport coverage (432 airports, 182 countries, 6 languages)