search

askforalfred / alfred

ALFRED - A Benchmark for Interpreting Grounded Instructions for Everyday Tasks
MIT License
352 stars 77 forks source link

Question about Resnet-18 to encode vision info in Alfred benchmark #129

Closed xyzhu123 closed 1 year ago

xyzhu123 commented 1 year ago

Hello, may I ask what is the training set used for the resnet-18 used to encode visual scenes in the Alfred dataset? image

MohitShridhar commented 1 year ago

@WhyTheMoon, it's just a standard ResNet-18 from PyTorch pre-trained on ImageNet. See this script for details.