This benchmark evaluates typical workloads on cloud services for vector databases. It is a fork of qdrant/vector-db-benchmark, and accommodates cloud services such as MyScale, Pinecone, Weaviate, Qdrant, and Zilliz.
A summary of the benchmark results can be found in our blog post. For setup, datasets and detailed results, visit: https://myscale.github.io/benchmark.
First, install the necessary libraries on the client used for the benchmark.
pip install -r requirements.txt
python3 run.py --datasets gist-960-euclidean-l2
Afterwards, follow the Step-by-Step Guide for Benchmark to execute the benchmark for each cloud service. You can refer to Results Visualize for visualizing the test results.