Insertion on Simulation
98.2Insertion Success RateViTaS
Evaluation results
| Method | Links | |
|---|---|---|
| ViTaSTask type=Reinforcement Learning (PPO), Training timesteps=3 × 10^6, Evaluation=Average of 5 random seeds2026.02 | 98.2 | |
| MAE-DiffusionViewpoint=Both2026.02 | 92 | |
| MAE-SelectViewpoint=Adaptive (Selecting from Top/Left)2026.02 | 88 | |
| Diffusion PolicyViewpoint=Both2026.02 | 84 | |
| MAE-DiffusionViewpoint=Top2026.02 | 84 | |
| CVTTask type=Reinforcement Learning (PPO), Training timesteps=3 × 10^6, Evaluation=Average of 5 random seeds2026.02 | 83.9 | |
| Diffusion PolicyViewpoint=Top2026.02 | 80 | |
| RLDCSeparate Head=false2026.06 | 79 | |
| VTTTask type=Reinforcement Learning (PPO), Training timesteps=3 × 10^6, Evaluation=Average of 5 random seeds2026.02 | 78.6 | |
| MAE-DiffusionViewpoint=Left2026.02 | 78 | |
| RLDC2026.06 | 74 | |
| M3LTask type=Reinforcement Learning (PPO), Training timesteps=3 × 10^6, Evaluation=Average of 5 random seeds2026.02 | 72.1 | |
| RLDCState Pre-training=false2026.06 | 68 | |
| MVTTask type=Reinforcement Learning (PPO), Training timesteps=3 × 10^6, Evaluation=Average of 5 random seeds2026.02 | 64.3 | |
| Diffusion PolicyViewpoint=Left2026.02 | 64 | |
| iDP32026.06 | 47 | |
| DP32026.06 | 38 | |
| ConcatTask type=Reinforcement Learning (PPO), Training timesteps=3 × 10^6, Evaluation=Average of 5 random seeds2026.02 | 19.3 | |
| PoETask type=Reinforcement Learning (PPO), Training timesteps=3 × 10^6, Evaluation=Average of 5 random seeds2026.02 | 11.4 |