analysing the output after plot is generated for the databases

erikbern / ann-benchmarks

Benchmarks of approximate nearest neighbor libraries in Python

http://ann-benchmarks.com

MIT License

4.88k stars 735 forks source link

analysing the output after plot is generated for the databases #504

Open dipshirajput opened 6 months ago

dipshirajput commented 6 months ago

Hi all,

I was working with benchmarking Milvus and qdrant . i am able to generate the plots for both the db. please let me know how can we analyse the graph and conclude their performance comparison ?

how can we analyse the following plots ? please help me with this.
also which plot should we go through in order to conclude the performance ?

fashion-mnist-784-euclidean_10_euclidean glove-100-angular_10_angular

Thanks in advance.

GreateFang commented 4 months ago

HI, @dipshirajput ,could you pls tell me how you successfully run the milvus test? It keeps reminding missing docker images while idk how to find this image.Didn't dound in internet too. Thanks a lot for your help!