Era0-Preview
Evaluation-only VLA model for RoboDojo simulation, contributed by RobotEra Team. Supported environment and action types: arx_x5 and ee.
Evaluate
Install RoboDojo and use the policy/Era0_Preview adapter from XPolicyLab. From that policy directory:
bash install.sh
bash download_checkpoint.sh
export ERA0_CHECKPOINT_PATH="$(pwd)/checkpoints/era0-preview"
POLICY_ENV=.venv bash eval_three_seeds.sh \
--checkpoint "$ERA0_CHECKPOINT_PATH" --sim-env RoboDojo \
--execution-profile official --gpu-ids 0,1,2,3,4,5,6,7 \
--output "/path/to/eval_$(date +%F_%H%M%S)" --run
Replace RoboDojo with your simulator environment name or Python executable. Omit --run to preview. Keep all supplied model, configuration, processor/tokenizer and normalization files together. The matching adapter and asset hashes are documented in the policy repository.
Local results
Partial local evaluation from the official SDK summary: Score 38.36 ± 0.55, SR 33.11 ± 0.27%, 123/126 complete task–seed cells, 6,265/6,300 saved episodes. Incomplete cells are excluded. The remaining 35 episodes belong to make_toast_random (5/20/10 missing for seeds 0/1/2).
These are local test results, not official leaderboard scores.
- Downloads last month
- 83