pi0.5 LIBERO full-SFT (comparison baseline)
Full-parameter supervised fine-tune of π0.5 (pi05) on the LIBERO benchmark, trained with the openpi PyTorch backend. This is a full SFT checkpoint (not LoRA), intended as a comparison baseline.
Training
- Base:
pi05_base(paligemmagemma_2b+ action expertgemma_300m), action horizon 10. - Config:
pi05_libero, full fine-tune (no LoRA, no freeze). - 60,000 steps, global batch 128 (32/GPU), 4× RTX A6000 DDP, bf16 + gradient checkpointing.
- Data:
physical-intelligence/libero(LeRobot), 2 cameras (image + wrist_image). - Final training loss ≈ 0.0036.
LIBERO eval (partial)
Evaluated on the 4 standard suites (openpi client/server, MuJoCo EGL). Numbers below are from a partial run (early-stopped), not the full 50 trials/task:
| Suite | Success rate (partial) |
|---|---|
| libero_spatial | 100% (84/84) |
| libero_object | 96.7% (58/60) |
| libero_goal | 96.3% (79/82) |
| libero_10 | 100% (37/37) |
Reference (openpi pi0.5 @ 30k): spatial 98.8 / object 98.2 / goal 98.0 / libero_10 92.4 (avg 96.85).
Files
model.safetensors— bf16 model weights (inference).optimizer.pt— AdamW optimizer state (for resuming training).metadata.pt— training metadata (global_step).assets/physical-intelligence/libero/norm_stats.json— normalization stats (required for inference).
Usage
Serve with openpi (auto-detects the PyTorch checkpoint via model.safetensors):
python scripts/serve_policy.py --env LIBERO \
policy:checkpoint --policy.config pi05_libero --policy.dir <path-to-this-checkpoint>