Trainging Step 1 do not have sampling_training ？

[kongyuzhuo]

Hi,

Thank you for your great work, I have some questions regarding the first phase of training.

1. Do you need to specify --checkpoint as None for the first stage of training?
2. What does sampling_training mean? Do I need to specify it? If I don't specify it, the training will go directly to out_net = model(d, quality = list_quality) (which in line 86 in step.py), but an error will be reported because quality is not specified.

Looking forward to your reply :)

Best,
Sonia

[cream-sketch]

Hi,

Thank you for your great work, I have some questions regarding the first phase of training.

1. Do you need to specify --checkpoint as None for the first stage of training?
2. What does sampling_training mean? Do I need to specify it? If I don't specify it, the training will go directly to out_net = model(d, quality = list_quality) (which in line 86 in step.py), but an error will be reported because quality is not specified.

Looking forward to your reply :)

Best, Sonia
Hi Sonia,

Did you solve the problem? During the process of starting the first training, I encountered some boolean type errors in the train.py file, as well as an undefined error for list_quality. Have you experienced something similar?

Best regards,
KK

[kongyuzhuo]

Hi,
Thank you for your great work, I have some questions regarding the first phase of training.

1. Do you need to specify --checkpoint as None for the first stage of training?
2. What does sampling_training mean? Do I need to specify it? If I don't specify it, the training will go directly to out_net = model(d, quality = list_quality) (which in line 86 in step.py), but an error will be reported because quality is not specified.

Looking forward to your reply :)
Best, Sonia
Hi Sonia,

Did you solve the problem? During the process of starting the first training, I encountered some boolean type errors in the train.py file, as well as an undefined error for list_quality. Have you experienced something similar?

Best regards, KK

Hi KK,

I tried it myself, sampling_training was set to false, quality was set to [0,10], you can refer to line 132 of train.py.
Hope that will help you.

Best,
Sonia

[cream-sketch]

Hi,
Thank you for your great work, I have some questions regarding the first phase of training.

1. Do you need to specify --checkpoint as None for the first stage of training?
2. What does sampling_training mean? Do I need to specify it? If I don't specify it, the training will go directly to out_net = model(d, quality = list_quality) (which in line 86 in step.py), but an error will be reported because quality is not specified.

Looking forward to your reply :)
Best, Sonia
Hi Sonia,

Did you solve the problem? During the process of starting the first training, I encountered some boolean type errors in the train.py file, as well as an undefined error for list_quality. Have you experienced something similar?
Best regards, KK

Hi KK,

I tried it myself, sampling_training was set to false, quality was set to [0,10], you can refer to line 132 of train.py. Hope that will help you.

Best, Sonia

Thanks a lot for your reply!