List your company
ResearchFormerly WizWand

VLM Reasoning on NumberLine Out-of-Distribution

18.6Success RateGFlowVLM w/ DB
-0.7444.2789.314.322Mar 9, 2025
Updated 8mo ago

Evaluation results

MethodLinks
GFlowVLM w/ DBTrain Data=Off-Policy, Assump.=Non-Markovian, SFT Init.=Yes2025.0318.6
GFlowVLM w/ Var-TBTrain Data=Off-Policy, Assump.=Non-Markovian, SFT Init.=Yes2025.0317.3
GFlowVLM w/ SubTBTrain Data=Off-Policy, Assump.=Non-Markovian, SFT Init.=Yes2025.0316.7
GFlowVLM w/ DBTrain Data=On-Policy, Assump.=Non-Markovian, SFT Init.=Yes2025.039.1
GFlowVLM w/ SubTBTrain Data=On-Policy, Assump.=Non-Markovian, SFT Init.=Yes2025.037
GFlowVLM w/ Var-TBTrain Data=On-Policy, Assump.=Non-Markovian, SFT Init.=Yes2025.036.2
RL4VLM+Train Data=On-Policy, Assump.=Non-Markovian, SFT Init.=Yes2025.034.4
RL4VLMTrain Data=On-Policy, Assump.=Markovian, SFT Init.=Yes2025.033.1
RL4VLM*Train Data=On-Policy, Assump.=Markovian, SFT Init.=Yes, Reward function=Same as ours2025.031.9
SFT-w/o- [DONE]Train Data=Off-Policy, Assump.=N/A, SFT Init.=N/A2025.030
SFT-w/- [DONE]Train Data=Off-Policy, Assump.=N/A, SFT Init.=N/A2025.030
GFlowVLM w/ SubTBTrain Data=On-Policy, Assump.=Non-Markovian, SFT Init.=No2025.030
GFlowVLM w/ DBTrain Data=On-Policy, Assump.=Non-Markovian, SFT Init.=No2025.030
GFlowVLM w/ SubTBTrain Data=Off-Policy, Assump.=Non-Markovian, SFT Init.=No2025.030
GFlowVLM w/ DBTrain Data=Off-Policy, Assump.=Non-Markovian, SFT Init.=No2025.030