generate new images from checkpoints

[inbalaha]

Hi, we trained the model with our data and got good samples. Now we are trying to generate new images by using our checkpoint, with the following command "python generate.py checkpoint/train_step-4.model" .
we get this result:

the images from the samples lookes like this :

it seems like the generated image is a combination of the samples and a random style.
is there a way to generate the image without the style?
thanks in advance

[rosinality]

You need to set --size argument to the generate.py that matches with the resolutions model trained with.

[inbalaha]

thank you, it works!