pi05_telearms_v5_exp_random_1000

Final checkpoint: step 30,000.

π₀.₅ (lerobot/pi05_base) full fine-tune on mliu48/telearms-yam-cube-v5-exp-random-1000 (revision 94fef4f3): 10,000 simulated bimanual YAM demonstrations of 10 tasks, 9,488,515 frames at 30 Hz, three 224×224 cameras, no domain randomization.

Selection. 1000 demonstrations per task drawn at random from all successful demonstrations of the task in the full dataset that have a demo score and a rated player ELO. The Random sets are nested: Random-50 is contained in Random-100, and so on up to Random-1500.

One of eight v5 experiment runs (Random-50/100/500/1000/1500, Elo-1500, Score-1500, DemInf-1500) trained with the same recipe and the same 15.36M-sample budget so they can be compared.

• Code: helen9975/RoboEval-telearms branch PI05_training @ 9dc9e9f plus local changes (NaN guards, PyAV logging fix for data-loading speed, configurable minimum episodes per task), LeRobot 0.6.1 @ e40b58a8 with the repo's patch (tools/install_lerobot_patch.py).
• Full fine-tune (LLM + vision + action expert), bf16, gradient checkpointing; per-task prompt variants (meta/task_prompt_variants.json); no episodes held out, so there is no validation loss.
• 8 × H100 (1 node), batch 64/GPU (global 512), 30,000 steps = 15.36M samples; this checkpoint has seen 15.36M of them.
• AdamW (β = 0.9, 0.95, weight decay 0.01), peak LR 1e-4 (2.5e-5 × sqrt(512/32)), 1,000 warmup steps, cosine decay to 1e-5 over 30,000 steps; 50-step action chunks; relative actions with absolute grippers.
• Recent training loss: 0.023.
• These files are the checkpoint's pretrained_model/.
• Evaluation: simulation, training/evaluate_pi.py suite on the 10 trained tasks, 20 episodes per task (fixed seeds 900000001+, the same scenes for all eight experiment models), up to 5,400 steps, 25-step action chunks, success = the task succeeded at any point. 175/200 = 87.5% (95% Wilson interval 82-91%).

Task	Success
Cube Patch Placement	19/20
Front Edge Cube Row	17/20
Front To Back Cube Line	18/20
L Cube Arrangement	16/20
Letter Face Alignment	14/20
Red Cube In Front Of Blue Cube	20/20
Red Left Of Blue Placement	20/20
Semantic Cube On Red Patch A Up	13/20
Sequential Push Block to Target	19/20
Three Cube Wall Row	19/20