Search results

Proposal: URML (substrate-neutral robot intent) as the pre-flight typed safety check above OpenVLA action output

Status: Open.

#338 In openvla/openvla;

· idoco2003 opened on May 28, 2026

Using multiple GPUs doesn't accelerate the training

Status: Open.

#336 In openvla/openvla;

· XiaoweiLinXL opened on May 20, 2026

Low success rate on Libero-object

Status: Open.

#335 In openvla/openvla;

· HIT-DLX opened on May 5, 2026

LoRA fine-tune on LIBERO-Spatial: 0% success, mode collapse — missing clip_grad_norm_ in finetune.py?

Status: Open.

#333 In openvla/openvla;

· shant0602 opened on Apr 27, 2026

How to finetune on LIBERO dataset via lora?

Status: Open.

#332 In openvla/openvla;

· HIT-DLX opened on Apr 25, 2026

机械硬盘对推理的影响

Status: Open.

#330 In openvla/openvla;

· qi-zhang777 opened on Apr 8, 2026

A Question about ActionTokenizer

Status: Open.

#327 In openvla/openvla;

· XLINYIN opened on Mar 24, 2026

Where to download Franka-Tabletop dataset?

Status: Open.

#326 In openvla/openvla;

· omsrisagar opened on Mar 23, 2026

📋 Documentation Enhancement Suggestion

Status: Open.

#325 In openvla/openvla;

· croviatrust opened on Mar 15, 2026

Post-Training Quantization (FP8/NVFP4) on OpenVLA with LIBERO Benchmark

Status: Open.

#324 In openvla/openvla;

· jtao34 opened on Mar 4, 2026

Is the dataset actually correctly sharded in DDP finetuning training?

Status: Open.

#323 In openvla/openvla;

· David0tt opened on Feb 6, 2026

cmu_play_fusion dataset action space

Status: Open.

#322 In openvla/openvla;

· D222097 opened on Feb 2, 2026