Search results

[feature] Make DataLoader throughput tunable from config

Status: Open.

#332 In starVLA/starVLA;

· Arrebol2020 opened on May 15, 2026

Qwen3-VL-Bridge-RT-1 checkpoint issues

Status: Open.

#324 In starVLA/starVLA;

· Xiaoxi-Sun opened on May 9, 2026

[Question] Reproducing StarVLA-GR00T on RoboTwin 2.0 gives much lower success than the paper

Status: Open.

#322 In starVLA/starVLA;

· Shark-Overlord opened on May 8, 2026

[Bug] Missing Commit b7f9047 in starVLA_dev After PR #308 Merge — Branch Reset or Force-Push Suspected

Status: Open.

#310 In starVLA/starVLA;

· luocfprime opened on May 6, 2026

Inquiries regarding the reproduction of Table-5 (Mixed Benchmark Training)

Status: Open.

#296 In starVLA/starVLA;

· zhihelu-zero opened on Apr 28, 2026

Adaptive LayerNorm may cause loss explosion in QwenPI

Status: Open.

#237 In starVLA/starVLA;

· Yysrc opened on Apr 3, 2026

LoRA / Parameter-Efficient Fine-Tuning

Status: Open.

#229 In starVLA/starVLA;

· da2853 opened on Mar 30, 2026

KV Caching for Inference

Status: Open.

#224 In starVLA/starVLA;

· da2853 opened on Mar 29, 2026

Cross-embodiment training

Status: Open.

#215 In starVLA/starVLA;

· markyyds opened on Mar 25, 2026

Robocasa (Panda Omron) evaluation results

Status: Open.

#103 In starVLA/starVLA;

· youthHan opened on Jan 7, 2026

🚧 Daily Development Log: Optimizations & New Features

🚧 Daily FeaturesWe use this thread to document **performance optimizations** and **new features**.We use this thread to document **performance optimizations** and **new features**.

Status: Open.

Feature

#64 In starVLA/starVLA;

· JinhuiYE opened on Dec 14, 2025