Roman Naeem

Conversion to a TFlite model for performing on-device inference

1

Hi, Do you think it would be possible to convert this model or the yolov3 rotated to ONNX-TensorFlow-TFlite and use it on an android device?

Loss cardinality does not make sense in the setting when using focal loss

1

Hi, I see that starting from deformable DETR, it doesn't make sense to have cardinality loss. In the original DETR paper, the class_embed head outputs num_classes + 1, to also...

Question about a line in the dino forward code

1

Hi, There is a line in the dino forward code [`hs[0] += self.label_enc.weight[0,0]*0.0`](https://github.com/IDEA-Research/DINO/blob/3ffda400b0f1d4a919fbe2a9cf567e79d210ffde/models/dino/dino.py#L272). The comment above it says "# In case num object=0" what does that mean and what is...

Loss cardinality does not make sense in the setting when using focal loss

1

Hi, I see that starting from deformable DETR, it doesn't make sense to have cardinality loss. In the original DETR paper, the class_embed head outputs num_classes + 1, to also...

Allow specifying a root position of the tree

7

Hi William, I am using Kimimaro, to generate centerline graphs for the Parse 2022 challenge dataset. It works well, however, the selected starting or root point is usually incorrect and...

Control how early or late we prefer the skeleton to fork at bifurcations

1

Hi, Is it possible to control (with a parameter) how early or late we want the skeleton to fork at a bifurcation point? There is the fix_branching flag, but having...