How can I infer my single img?

[ShuxunoO]

I want to test the model with my demo img， how can I do this？

[ZhiJingWu]

Hello everyone, I've encountered a similar issue. Has anyone found a solution for it?

[meisa233]

modify the .yml file in options/test. For example, in HAT_GAN_Real_SRx4.yml

line 15
    dataroot_lq: input_dir

please modify it to a real folder path including your images.

This is my example:

line 15
    dataroot_lq: ./datasets/input_dir

I make a folder named input_dir in ./datasets/ and move a .png file to it. Then

python hat/test.py -opt options/test/HAT_GAN_Real_SRx4.yml

[Nan1107]

hello，Infer whether a single image was calculated using a GPU? I feel that my computing speed is a bit too slow, so I suspect that I haven't used GPU calculations. How should I check this

[meisa233]

hello，Infer whether a single image was calculated using a GPU? I feel that my computing speed is a bit too slow, so I suspect that I haven't used GPU calculations. How should I check this

I haven't tested this code in a long time, but I probably remember it took a few seconds to infer an image. when you run your script, you can use nvidia-smi to see the Memory-Usage of your GPU. On the other hand, you can use torch.cuda.is_available() to check your cuda.

[Nan1107]

hello，Infer whether a single image was calculated using a GPU? I feel that my computing speed is a bit too slow, so I suspect that I haven't used GPU calculations. How should I check this

I haven't tested this code in a long time, but I probably remember it took a few seconds to infer an image. when you run your script, you can use nvidia-smi to see the Memory-Usage of your GPU. On the other hand, you can use torch.cuda.is_available() to check your cuda.

Thank you for your reply. It's the reason for my computer's insufficient performance