ResearchBenchmarksVAE Video Reconstruction and Fine-tuning Evaluation on SimplerEnv WidowXFollow33.4PSNRCoWVLA32.67232.86133.0533.239Mar 3, 2026Evaluation ResultsMethodMethodLinksPSNRSSIMLPIPSTask Success: Stack BlockTask Success: Put CarrotTask Success: Put SpoonTask Success: Put EggplantAverage Task SuccessCoWVLATraining Stage=FinetuneTraining Stage=Finetune2026.0333.493.40.12362.566.779.295.876PretrainTraining Stage=PretrainTraining Stage=Pretrain2026.0332.792.30.12245.87579.291.772.9