New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[FEA] Reduce small-batch size inference overhead #744

Closed

3 tasks done

karlhigley opened this issue Apr 15, 2021 · 0 comments

Closed

3 tasks done

[FEA] Reduce small-batch size inference overhead #744

karlhigley opened this issue Apr 15, 2021 · 0 comments

Assignees

Labels

Inference P0 Perf

Milestone

Contributor

karlhigley commented Apr 15, 2021 •

edited by benfred

Loading

Update operator API's to indicate what data representations they support, and allow configuring for use with inference (Add 'supports' and 'inference_initialize' for operators. #918)
Update triton python model to use alternative 'dict' representations (Update triton model.py to use dict array representations #921)
Provide optimized c++ versions of fill + categorify operators (Use C++ operators for categorify/fill during inference #926)

karlhigley added the Inference label

EvenOldridge added the Perf label

viswa-nvidia added this to the NVTabular v0.6 milestone

benfred added the P0 label

benfred self-assigned this

benfred closed this as completed

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment