Miao Zheng

might check this arg https://github.com/open-mmlab/mmsegmentation/blob/5c113d98ecf362d9a215fbb3016ea7e76850ed98/mmseg/apis/test.py#L41 or just remove these lines https://github.com/open-mmlab/mmsegmentation/blob/5c113d98ecf362d9a215fbb3016ea7e76850ed98/mmseg/apis/test.py#L122-L124

VOC dataset didn't support saving the predicted results, and we will support it asap, or if you have interests, could you please create a nice pr for this function, which...

it's much appreciated for it. We are looking forward to this feature

if you do not mind, could you share the `thresh` value you set?

Many thanks for your kind reminder and the figure is very clear for us. We are working on figuring out it and if we have any processes or conclusions, we...

Please follow this documentation to set the `samples_per_gpu=1` for the validation loop, as mmseg don't support samples_per_gpu>1 during validation. https://github.com/open-mmlab/mmsegmentation/blob/master/docs/en/tutorials/customize_datasets.md#data-configuration

Would you mind providing the full-error log and following the issue template to tell us where raise this error?

during validation, `sample_per_gpu` must be 1, and mmseg haven't supported batch inference, as the shape of image is different. By default, mmseg set validation batch_size=1, did you modify it?

we will fix it asap

Thank you for supporting new mmseg, and we are working on it, and you will see it next week.