Ashwin Nanjappa
Ashwin Nanjappa
@gyenesvi A new model named RetinaNet was added for 2.1.
LGTM. But we should not merge this to v2.1 branch because submitters are collecting submission logs already with that branch.
@arjunsuresh We are using the (slow) script for MLPerf Inference too. 😁 Since you seem to be familiar with it, would you like to contribute by adding support for the...
@saareliad It should be 600. Could you point out one example file where it is 60?
@TheKanter This has been discussed in both Training and Inference WGs. Do you have any further open questions?
We should get back to this issue after https://github.com/mlcommons/inference/pull/1677 is merged. That PR would simplify the submission checker a lot.
@mrmhodak NVIDIA members experienced with submission checker also agreed that we can drop the backwards compatibility and thus simplify the submission checker a lot. Arjun's PR is a good idea...
@tjablin Please review and approve if OK.
@tjablin Please merge.
@rnaidu02 Could you help?