Galaxea G0.5 — DROID CloseLaptop (10 epochs)

Task: "close laptop" — 100 episodes, 27,649 frames @ 30 fps (aarontung/droid-close-laptop)
Inputs/outputs: cameras left exterior / wrist + instruction; 8-D absolute joint state and action (7 Franka joints + gripper)
Base: OpenGalaxea/G05 g05-base (GalaxeaVLA)
Trained: Full fine-tune, vision LR x0.1, FSDP (R1Lite post-training recipe; data in R1Lite layout)
Optim: LR 4e-5, cosine, warmup 200, wd 0.03; GBS 16 (16x1 micro-batch x accumulation, 1 GPU); 10 epochs total: 5 in the DROID-models run + 5 more here; fp32 master weights, bf16 compute; 1x H200; wall time 2h23m (the 5 extra epochs)
Files: model.pt (final epoch), dataset_stats.json, action_tokenizer.pt.
Load: GalaxeaVLA with task config configs/galaxea/task/rc_droid.yaml
Code: RoboColosseum harness (scripts/train/g05.sh); logs in W&B project RoboColosseum
Continued from: RoboColosseum/DROID-models/g05/closelaptop (5 epochs). The 5 extra epochs restart the same schedule (same peak LR, warmup and decay) from those final weights with a fresh optimizer, at the same global batch size; logged to the same W&B run.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Collection including RoboColosseum/droid-close-laptop-g0.5