A Pragmatic VLA Foundation Model

TL;DR

This paper introduces LingBot-VLA, a versatile vision-language-action foundation model trained on extensive real-world robotic data, demonstrating superior performance, efficiency, and generalizability across multiple robotic platforms and tasks.

Contribution

We present LingBot-VLA, a novel VLA foundation model trained on 20,000 hours of real-world data, with a highly efficient codebase and broad evaluation across platforms and tasks.

Findings

Outperforms competitors in robotic manipulation tasks

Achieves 261 samples/sec training throughput

Demonstrates strong generalization across platforms

Abstract

Offering great potential in robotic manipulation, a capable Vision-Language-Action (VLA) foundation model is expected to faithfully generalize across tasks and platforms while ensuring cost efficiency (e.g., data and GPU hours required for adaptation). To this end, we develop LingBot-VLA with around 20,000 hours of real-world data from 9 popular dual-arm robot configurations. Through a systematic assessment on 3 robotic platforms, each completing 100 tasks with 130 post-training episodes per task, our model achieves clear superiority over competitors, showcasing its strong performance and broad generalizability. We have also built an efficient codebase, which delivers a throughput of 261 samples per second with an 8-GPU training setup, representing a 1.5~2.8$\times$ (depending on the relied VLM base model) speedup over existing VLA-oriented codebases. The above features ensure that our model is well-suited for real-world deployment. To advance the field of robot learning, we provide open access to the code, base model, and benchmark data, with a focus on enabling more challenging tasks and promoting sound evaluation standards.