Vignesh Kothapalli

kvignesh1420

·

https://kvignesh1420.github.io/

AI & ML interests

Deep learning

Recent Activity

liked a model 6 days ago

nvidia/Kumo-Relational

upvoted a paper 6 days ago

UniEvo-VL: An On-policy Self-Distillation Training Recipe for Multimodal Model Self-improvement

upvoted an article 7 days ago

NVIDIA Kumo Tabular Sets a New Accuracy-Efficiency Frontier for Tabular Prediction

View all activity

Organizations

New activity in stanford-star/rt-plurel 4 months ago

RelBench leaderboard checkpoints + README (flat layout) Checkpoints behind the PluRel rows on https://relbench.stanford.edu/leaderboard/: - cntd-pretrain_<db>_<task>.pt: synthetic+real continued pretraining (leave-one-DB-out per RelBench database incl rel-event; per-task best; val-NMAE selection for regression) - finetune_<db>_<task>.pt: per-task fine-tune from the matching cntd-pretrain checkpoint - README section documenting both. Full official test splits.

#5 opened 4 months ago by

New activity in stanford-star/rt-plurel 7 months ago

Add pipeline tag and link to original Relational Transformer paper

#1 opened 7 months ago by