snap-research/GRID
GRID: Generative Recommendation with Semantic IDs
Implements a three-stage pipeline converting item text into LLM embeddings, then learning hierarchical semantic IDs via residual quantization (RQ-KMeans, RQ-VAE, RVQ), and finally generating recommendations using transformer-based sequence modeling. Built on PyTorch Lightning with Hydra configuration management, it enables end-to-end generative recommendation without explicit item IDs.
609 stars.
Stars
609
Forks
108
Language
Python
License
—
Category
Last pushed
Oct 15, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/embeddings/snap-research/GRID"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related tools
eugeneyan/semantic-ids-llm
Semantic IDs: How to train an LLM-Recommender Hybrid with steerability and reasoning on recommendations.
HKUDS/XRec
[EMNLP'2024] "XRec: Large Language Models for Explainable Recommendation"
HKUDS/RLMRec
[WWW'2024] "RLMRec: Representation Learning with Large Language Models for Recommendation"
HKUDS/EasyRec
[EMNLP'2025] "EasyRec: Simple yet Effective Language Model for Recommendation"
taherfattahi/recommendation-systems-by-llms
Enhancing Recommendation Systems with Large Language Models (RAG - LangChain - OpenAI)