All issues

Search results

Verify evals on Papers with Code

Status: Open.

#161 In moojink/openvla-oft;

· NielsRogge opened on Sep 14, 2026

Action head gradients are not synchronized in multi-GPU training because DDP forward is bypassed

Status: Open.

#160 In moojink/openvla-oft;

· TJ12342 opened on Jul 29, 2026

Evaluation aborts

Status: Open.

#158 In moojink/openvla-oft;

· invoidstar opened on Jun 15, 2026

Query regarding the coupling of unnorm stats and the fine-tuned check point of OpenVLA-oft

Status: Open.

#156 In moojink/openvla-oft;

· praneetnayak opened on Jun 4, 2026

Modified libero HDF5

Status: Open.

#153 In moojink/openvla-oft;

· reinaldoyang opened on May 6, 2026

Question about fine-tuning OpenVLA‑OFT with non‑RLDS (GR00T / LeRobot v2 style) datasets

Status: Open.

#152 In moojink/openvla-oft;

· fsiken opened on Mar 19, 2026

Re-merging LORA weights resulted in the success rate dropping to zero.

Status: Open.

#151 In moojink/openvla-oft;

· prilgrim-whtie opened on Mar 13, 2026

success rate of libero-spatial is lower than the paper

Status: Open.

#150 In moojink/openvla-oft;

· atticus77777 opened on Mar 10, 2026

[Question] Adapting OpenVLA-OFT for 7-DOF Robot Arm (Joint Angle Action Space)

Status: Open.

#149 In moojink/openvla-oft;

· lozoRUST opened on Feb 28, 2026

ValueError: Unknown split 'val'. Should be one of ['train'].

Status: Open.

#148 In moojink/openvla-oft;

· suixin1424 opened on Feb 2, 2026

Question regarding training with chunk=1 and loss curves

Status: Open.

#147 In moojink/openvla-oft;

· maplecatyql opened on Jan 31, 2026

The issues in building new scenarios for testing models.

Status: Open.

#146 In moojink/openvla-oft;

· Xj1567589354 opened on Jan 30, 2026