Model Releases
VQ-bench: a Composable Vector Quantization Framework
VQ-bench is an open‑source framework that lets researchers construct vector quantization schemes from a small set of reusable primitive operations, enabling easy comparison of different methods on the
VQ-bench is an open‑source framework that lets researchers construct vector quantization schemes from a small set of reusable primitive operations, enabling easy comparison of different methods on the same datasets and hardware. Created by Ashwin Padaki, Amir Ingber, and Edo Liberty (Sep 17 2026), it was designed to address the proliferation of heterogeneous quantizers in vector databases and LLMs by providing a standardized, recipe‑based interface for fair, reproducible benchmarking.
Related
- Revisiting RaBitQ and TurboQuant: A Symmetric Comparison of Methods, Theory, and Experiments
- TOPO-Bench: An Open-Source Topological Mapping Evaluation Framework with Quantifiable Perceptual Aliasing
- OSMa-Bench++: Toward Open-Ended Benchmarking of Semantic Mapping for Manipulation with Prompt-Generated Synthetic Scenes
- gym-invmgmt: An Open Benchmarking Framework for Inventory Management Methods
Source: Pinecone | 2026-09-17