List your company
ResearchFormerly WizWand

Robotic Manipulation on LIBERO Object

99.4Success RateForgeVLA
-1.68824.55650.877.044May 8, 2026May 17, 2026May 27, 2026Jun 6, 2026Jun 16, 2026Jun 26, 2026Jul 6, 2026
Updated 3mo ago

Evaluation results

MethodLinks
ForgeVLAE=100, P=1, # Params (M)=3882.724, # Trainable Params (M)=128.1012026.0599.4100
Centralized# Params (M)=3882.72, # Trainable Params (M)=128.102026.0598.8100
CentralizedE=100, P=1, # Params (M)=3882.724, # Trainable Params (M)=128.1012026.0598.8100
CentralizedE=10, P=10, # Params (M)=3882.724, # Trainable Params (M)=128.1012026.0598.8100
ForgeVLA# Params (M)=3882.72, # Trainable Params (M)=128.102026.0598.6100
FedAvgE=100, P=1, # Params (M)=3882.724, # Trainable Params (M)=128.1012026.0598.2100
FedAvg# Params (M)=3882.72, # Trainable Params (M)=128.102026.0597.6100
MolmoAct22026.0797—
π0.52026.0796—
ForgeVLAE=10, P=10, # Params (M)=3882.724, # Trainable Params (M)=128.1012026.0595.6100
GaP2026.0795—
A2A-ExplicitVisual conditioning=A2A-GroundingModel mask rendered as an RGB highlight, Action head=Diffusion Policy (DP) [10]2026.0694.4—
FedAvgE=10, P=10, # Params (M)=3882.724, # Trainable Params (M)=128.1012026.0594100
DP-RGBVisual conditioning=raw RGB observations, Action head=Diffusion Policy (DP) [10]2026.0687.8—
π0.5 w/GaPNavigation=GaP, Policy=π0.52026.0785—
WAM-RLDescription=Ours2026.0682—
πRLDescription=actor-only reinforcement learning2026.0678—
A2A-ImplicitVisual conditioning=intermediate features from A2A-GroundingModel, Action head=Diffusion Policy (DP) [10]2026.0675.8—
UAD-DPVisual conditioning=fine-tuned UAD [51] affordance heatmap, Action head=Diffusion Policy (DP) [10]2026.0674.4—
MolmoAct2 w/GaPNavigation=GaP, Policy=MolmoAct22026.0770—
BaseDescription=pretrained WA model without RL2026.0668—
TipTop2026.0722—
FedVLA* + CLIP# Params (M)=519.82, # Trainable Params (M)=92.902026.0518.230
FedVLA*# Params (M)=79.23, # Trainable Params (M)=79.232026.052.210