algorithmiahq

algorithmiahq / qa_parallel_pipelining / 0.1.1

README.md

This orchestrator algorithm is used to test the ability of the platform to handle parallel pipelined or chained requests to an image processing algorithm which has a number of large dependencies and pipelines of it's own.

inputs

This algorithm supports two input modes:

Example QA Mode

250

In this mode, a single integer value is passed as input, which gets interpreted as a batch size . The algorithm will then create a batch consisting of a simple image for processing.

Example Real Mode

["www.pathtoimage.com/image.png", "..", ...]

In this mode, we provide a list of web hosted URLs or data api URIs to image files we plan to use for processing. The algorithm will use this url list as it's batch for processing.

Example Output

{
   "maximum_time":25.4,
   "minimum_time":4.65,
   "average_time":5.45
}