← Leaderboard

RoboFlamingo

VLA
ByteDance Research · Nov 2, 2023

Early VLA from ByteDance Research with Tsinghua University and SJTU. It fine-tunes the open OpenFlamingo VLM with a separate policy head on CALVIN’s language-annotated data only. Its CALVIN number is self-reported.

Type
VLA
Parameters
—
Backbone
OpenFlamingo (MPT-3B variants) + LSTM policy head
Weights
Open
License
MIT
Commercial use
Allowed
Released
Nov 2, 2023
Robots
Franka Panda (sim)
SO-101
–
LeRobot
–
Inputs
RGB · L
VRAM
–

Benchmark results

1 of 13 benchmarks
CALVIN ABC→DNear saturationSimulation
#24 of 242.48 / 5, self-reported by the authors, not a leaderboard entry
Not reported on 12 benchmarks
Wrong or missing number? Report an error