semantic-segmentation-pytorch icon indicating copy to clipboard operation
semantic-segmentation-pytorch copied to clipboard

Error while inferencing binary classification trained model (hrnet+c1)

Open Akash-Bhiwgade opened this issue 4 years ago • 1 comments

image

Hi Team,

Greetings..!!

This is in regards to an inference error. I have trained the model as a binary classifier (object and non-object). I have made changes to unm_class to 2 in the yaml config file. The annotations were made keeping 1 as object and 2 as non-object.

The model is trained on hrnetv2 and c1, however I am getting the above error. (please find attached screenshot)

Request you to assist me or guide me to the right direction.

Regards, Akash Bhiwgade

Akash-Bhiwgade avatar Jul 11 '20 04:07 Akash-Bhiwgade

Hello, I am trying to do something similar with HRNet. But I'm getting an error during training on my own dataset (I have just one class: road and the rest is background). Can you please tell me where do I need to change the labels and number of classes? (I have changed the num_class in hernetv2 and in the rest of the scripts.)

Param-Raval avatar Jul 14 '20 12:07 Param-Raval