Search results

Inference Extremely Slow (~10–20s per action) in sample_client.py

Status: Open.

#30 In NVlabs/vla0;

· Pratyush-1234 opened on Dec 27, 2025

[Reproducibility] Discrepancy between reported LIBERO benchmark results (94.7) and reproduced results (~92.5)

Status: Open.

#29 In NVlabs/vla0;

· JiaH-w opened on Dec 25, 2025

[RELEASE] Community reimplementation using TRL SFTTrainer

Status: Open.

#28 In NVlabs/vla0;

· MilkClouds opened on Dec 11, 2025

[INFO] Cause of Severe Multi-GPU Training Slowdown (Transformers v4.57.3)

Status: Open.

#27 In NVlabs/vla0;

· MilkClouds opened on Dec 10, 2025

Why are only images and prompts used as input during training, not states?

Status: Open.

#24 In NVlabs/vla0;

· MrFleix opened on Dec 1, 2025

How to train use 4 GPU with 48GB?

Status: Open.

#22 In NVlabs/vla0;

· zmf2022 opened on Nov 29, 2025

Multi-gpu training

Status: Open.

#19 In NVlabs/vla0;

· findcodematrix opened on Nov 22, 2025

Batch Inference and Flash Attention

Status: Open.

#18 In NVlabs/vla0;

· shivakanthsujit opened on Nov 21, 2025

Please avoid running shell commands (git/pip/conda) inside <code>setup.py</code>

Status: Open.

#17 In NVlabs/vla0;

· MilkClouds opened on Nov 21, 2025

Release VLA-0 on Hugging Face

Status: Open.

#1 In NVlabs/vla0;

· NielsRogge opened on Oct 16, 2025