Ashwin Nanjappa comments

Results 26 comments of


                                            Ashwin Nanjappa

Getting a lower than expected accuracy with DLRMv2 using Nvidia code for v3.1

@viraatc Please look into this if you have spare time before the v4.1 submission deadline. Else we can investigate after the deadline.

Setting `min_query_count` for GPTJ

@nv-jinhosuh Please provide your feedback to Arjun/Anton.

Clarify allowed changes in the system scale for Inference

@psyhtest Bullet items in your diff are not showing up correctly in the rendered Markdown.

Using alternative sequence lengths for SQuAD-based models in the Open division

@rnaidu02 Let us discuss this in the first WGM in 2023.

Allow for larger scale submissions in Inference when moving from Preview to Available

Instead of an amendment asking for permissions before submission, it might be worth having a change in the main rules itself to permit an available submission if both these conditions...

DLRMv2 GPU Reference Implementation crashes with BusError

Thanks @nv-etcheng . Could you also share how much (smaller) disk space is needed to store the preprocessed data? I'm thinking that users could temporarily acquire the larger disk space,...