Test results are not correct

[sanahtech]

I'm testing the code by using pre-trained model. i have executed following command python3 main.py --task pre-trained --is_training 0 But result's images do not make any sense. here is screenshot

[ceciliavision]

Can you make sure you successfully 1) loaded the model, 2) read in the images correctly 3) and write out the images correctly?

[yilun123]

I'm testing the code by using pre-trained model. i have executed following command python3 main.py --task pre-trained --is_training 0 But result's images do not make any sense. here is screenshot

I'm also testing the code by using pre-trained model, and executing the testing command. I get same problem as you. May I ask you how to solved this problem?

[ceciliavision]

I'm testing the code by using pre-trained model. i have executed following command python3 main.py --task pre-trained --is_training 0 But result's images do not make any sense. here is screenshot

I'm also testing the code by using pre-trained model, and executing the testing command. I get same problem as you. May I ask you how to solved this problem?

Could you you please make sure both the VGG model and our pre-trained model are correctly loaded?

[Tangyy12]

I'm testing the code by using pre-trained model. i have executed following command python3 main.py --task pre-trained --is_training 0 But result's images do not make any sense. here is screenshot

I'm also testing the code by using pre-trained model, and executing the testing command. I get same problem as you. May I ask you how to solved this problem?

I get same problem as you Have you solved this problem?

[Tangyy12]

I'm testing the code by using pre-trained model. i have executed following command python3 main.py --task pre-trained --is_training 0 But result's images do not make any sense. here is screenshot

I'm also testing the code by using pre-trained model, and executing the testing command. I get same problem as you. May I ask you how to solved this problem?

I get same problem as you Have you solved this problem?

I solved the above problem by changing the version of cudatoolkit to 11.1.1, and my GPU version is NVIDA RTX3070. I hope this method can help others .

[tiger-tangjj]

I get the same problem with cudatoolkit 10.0, tensorflow1.13.1 and NVIDA RTX3080Ti. Has your problem been solved? If so, I sincerely hope to get your help~

[ceciliavision]

Sorry about a delayed reply! It seems to be a package issue with slim. I now use tf_slim instead of contrib.slim and also with tf2 (tf2.10). I have updated the code and instructions. Hopefully this resolves the issue here.