I'm trying to train a custom dataset that has only one target per image. Currently, I am getting a class error of 0. Is this correct? I was not sure about how to set "num_queries" parameter for a single class (and single object) per image, so I have set it to 10 randomly for now.
I am using the command python src/train.py with infrared deformable multi_frame tracking output_dir=models/infrared_deformable_multi_frame
I have already set the custom data load function similar to mot17 and set my class_id as 1.Since the tracking_eval during training needs a custom dataloader or similar format to original mot17, I have set it as False for now.
Hello,
I'm trying to train a custom dataset that has only one target per image. Currently, I am getting a class error of 0. Is this correct? I was not sure about how to set "num_queries" parameter for a single class (and single object) per image, so I have set it to 10 randomly for now.
I am using the command python src/train.py with infrared deformable multi_frame tracking output_dir=models/infrared_deformable_multi_frame
I have already set the custom data load function similar to mot17 and set my class_id as 1.Since the tracking_eval during training needs a custom dataloader or similar format to original mot17, I have set it as False for now.