A fast, easy-to-use, production-ready inference server for computer vision supporting deployment of many popular model architectures and fine-tuned models.
This PR adds possibility to explicitly sub-sample video frame rate for each individual video source in InferencePipeline.
The functionality collides with max_fps setting which is now not satisfactory for clients. How it works now:
it is applicable globally for all video sources and regulates the maximum pace at wich consumer will operate
in default settings, for video stream it modulates the stream FPS, for video file - it slows down the processing which seems not to be the desired end
In this PR we are introducing new meaning for the max_fps option that can be optionally enabled via env flag and
is announced to be the default behaviour end of Q4 2024.
This solution has various benefits:
we do not introduce breaking change now
we are announcing change of API in advance letting people to migrate
we are letting users struggling with current implementation move forward
Type of change
Please delete options that are not relevant.
[x] Bug fix (non-breaking change which fixes an issue)
[x] New feature (non-breaking change which adds functionality)
[x] This change requires a documentation update
How has this change been tested, please provide a testcase or example of how you tested the change?
YOUR_ANSWER
Any specific deployment considerations
For example, documentation changes, usability, usage/costs, secrets, etc.
Description
This PR adds possibility to explicitly sub-sample video frame rate for each individual video source in InferencePipeline.
The functionality collides with
max_fpssetting which is now not satisfactory for clients. How it works now:In this PR we are introducing new meaning for the
max_fpsoption that can be optionally enabled via env flag and is announced to be the default behaviour end of Q4 2024.This solution has various benefits:
Type of change
Please delete options that are not relevant.
How has this change been tested, please provide a testcase or example of how you tested the change?
YOUR_ANSWER
Any specific deployment considerations
For example, documentation changes, usability, usage/costs, secrets, etc.
Docs