Official Repository of paper VideoGPT+: Integrating Image and Video Encoders for Enhanced Video Understanding
Dual Cross Encoder for Dense Retrieval
Efficient interpolation-based ranking on CPUs
Understand and build embedding models, focusing on word and sentence embeddings, dual encoder architectures. Learn to train embedding models using contrastive loss, implement them in semantic search and RAG systems.
This repository provides the code for "Improving Query-by-Vocal Imitation with Contrastive Learning and Audio Pretraining", presented at DCASE 2024. The paper addresses the challenge of audio retrieval using vocal imitations as queries, proposing a dual encoder architecture that leverages pretrained CNNs and an adapted NT-Xent loss for fine-tuning.
undergraduate thesis-based project
Policy recommendation using Two Tower Neural Network and Configurable business heuristics for SBI Hackathon (Rank 3).
Dual-Encoder in Tensorflow
Revamped Source code for Single-step Retrosynthesis Prediction by Leveraging Commonly Preserved Substructures
Trained an embedding model based on a causal language model (HuggingFaceTB/SmolLM2-135M) using contrastive loss with in-batch and hard negatives.
AI Player Finder is an intelligent sports scouting tool that matches natural-language descriptions to real player statistics. Using a dual-tower (text + numeric) AI model, it understands phrases like "creative midfielder" or "efficient 3-point shooter" and finds the best player fit.
Add a description, image, and links to the dual-encoder topic page so that developers can more easily learn about it.
To associate your repository with the dual-encoder topic, visit your repo's landing page and select "manage topics."