How to generate images using the checkpointed gpu model

[reachbp]

Hi,

How should we generate images using our trained model.
I am trying to generate images from a checkpoint state and I'm trying to use https://github.com/soumith/dcgan.torch/blob/master/generate.lua to generate images.
When I tried to convert the cuda tensor to a cpu-readable tensor using https://github.com/karpathy/neuraltalk2/blob/master/convert_checkpoint_gpu_to_cpu.lua , I get the following error:

/share/apps/torch/20160623/gnu/bin/luajit: convert_checkpoint_gpu_to_cpu.lua:79: bad argument #1 to 'pairs' (table expected, got nil)

Is there an easy way to generate images from the checkpointed model?

[reachbp]

Pls ignore my question. Just realized there is code under scripts/ to generate samples.