Does OpenVLA-OFT run on ALOHA (stationary)?

Capabilities

Caveats

• Real-robot results use the OFT+ variant (OFT plus FiLM language conditioning) on the original ALOHA (two ViperX 300 S arms, 3 cameras), run at 25 Hz instead of ALOHA's native 50 Hz.

Evidence 2 sources; the cell shows the highest tier

1. T1 · upstream documentedpapersource 2025-02-27verified 2026-10-24 (1d ago)

   OpenVLA-OFT paper fine-tunes OpenVLA with OFT+ on four real ALOHA tasks (fold shorts, fold shirt, scoop X into bowl, put X into pot); ALOHA here is the original ALOHA of Zhao et al. 2023 (ref. [56], arXiv 2304.13705), not ALOHA 2.

   “The ALOHA platform comprises two ViperX 300 S arms, three camera viewpoints (one top-down, two wrist-mounted), and robot state inputs (14-dimensional joint angles). It operates at 25 Hz (reduced from the original 50 Hz to enable faster training while still maintaining smooth robotic control)”

   https://arxiv.org/abs/2502.19645
2. T1 · upstream documentedofficial-docsverified 2026-10-24 (1d ago)

   Upstream repo documents ALOHA data preprocessing, fine-tuning (vla-scripts/finetune.py) and server-client real-robot evaluation (run_aloha_eval.py + deploy.py).

   “# OpenVLA-OFT+ in Real-World ALOHA Robot Tasks”

   https://github.com/moojink/openvla-oft/blob/e4287e94541f459edc4feabc4e181f537cd569a8/ALOHA.md