Gen Li

> Have you tested with this modification and does it resolve the mIoU problem discribed in [issue 120](https://github.com/GeorgeSeif/Semantic-Segmentation-Suite/issues/120) ? No, It's the same result, the test mean iou is still...

> Is the miou calculation correct (i read it and it seems to be) ? > Did you try to use **tf.metrics.mean_iou** like Deeplab do ? I think it is...

> @Reagan1311 I read the paper carefully, and I find that the author adds a GAP on the tail of Context Path. Then the author just combines the up-sampled output...

You can check [this function](https://github.com/Reagan1311/ASGNet/blob/cd898c55659fa0e14b5bddd8d5efaa009ce6557a/util/util.py#L67:5), which computes the area of intersection, and then the value will be updated in the intersection_meter. Normally, the intersection_meter.sum should be a numpy array, and...

Hi, as this paper focuses on 1/5-shot 1-way segmentation, so the model only predicts two classes (the desired foreground object and background).

Hi, SegmentationClassAug is from SBD augmented training data. You can refer to this [link](https://github.com/kazuto1011/deeplab-pytorch/blob/master/data/datasets/voc12/README.md)

@RozDavid Do you figure it out? I am also thinking about how to implement the text localization with BLIP2.

I found the segmentation annotation in this repository https://github.com/ivalab/affordanceNet_DA. It seems the number of images are different in these two links: [dropbox1](https://www.dropbox.com/s/9a3qoggh96qolj2/SimData_Gazebo.tar.gz?dl=0) [dropbox2](https://www.dropbox.com/s/52yo1hibsodspmx/data.tar.gz?dl=0) (66043 vs 37200). What is the difference?

@astrosaeed I have exactly the same error and warnings, did you figure it out?

> Have you figured out the reason for the drop in performance? No, I haven't. I gave up using this dataset.