FlowCast: Trajectory Forecasting for Scalable Zero-Cost Speculative Flow Matching

Divya Jyoti Bajpai, Shubham Agarwal, Apoorv Saxena, Kuldeep Kulkarni, Subrata Mitra, Manjesh K. Hanawal, [0.5em] Indian Institute of Technology Bombay · 2026

Read FlowCast: Trajectory Forecasting for Scalable Zero-Cost Speculative Flow Matching by Divya Jyoti Bajpai, Shubham Agarwal, Apoorv Saxena, Kuldeep Kulkarni, Subrata Mitra, Manjesh K. Hanawal, [0.5em] Indian Institute of Technology Bombay (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-01329.txt.

Description

Divya Jyoti Bajpai and co-authors build FlowCast, a no-train speed plan for flow picture makers. Flow nets make great pictures but crawl through many denoise steps, blocking live use, while distill, cut, or consistency tricks dent quality, cost retrains, or travel poorly. FlowCast bets on constant speed learned by flow nets, guessing future speed from current speed at zero extra cost inside a mean-square error cap. Steady zones thus skip redundant steps fast while tricky zones stay precise, with speedups beyond 2.5x, no quality loss.

Catalog record

CREATORDivya Jyoti Bajpai, Shubham Agarwal, Apoorv Saxena, Kuldeep Kulkarni, Subrata Mitra, Manjesh K. Hanawal, [0.5em] Indian Institute of Technology Bombay
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE49086
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602.01329

Browse the full library · About the Tower of Babel