htcr

Results 32 comments of htcr

Interested in testing the released trained checkpoints

Hi rbavery: We'll take care of this soon. Please stay tuned.

Interested in testing the released trained checkpoints

Have been quite busy recently. Will try to figure out how to upload to huggingface this week.

Interested in testing the released trained checkpoints

Our checkpoints have been uploaded to hugging face, see link in README. Closing, feel free to re-open if you encounter issues.

return error，When extract_graph_points doesn't extract useful points, it won't print out the result

Hi, thanks for flagging. What was the exact condition you encountered this issue? Were you using a custom dataset? If you have a fix, feel free to create a pull...

return error，When extract_graph_points doesn't extract useful points, it won't print out the result

Closing, feel free to reopen if need further discussion.

Value of "pixel_mean" and "pixel_std"

Thanks for your interest. These are the values used by SAM if I remembered correctly. I don't think they matter a lot as long as they are reasonable values for...

Value of "pixel_mean" and "pixel_std"

Closing, feel free to reopen if need further discussion.

from detectron2.modeling import ViT

https://github.com/facebookresearch/detectron2/blob/main/detectron2/modeling/backbone/vit.py notice this is not required for SAM-road, it's just an ablation study.

from detectron2.modeling import ViT

Closing, feel free to reopen if need further discussion.

Didn't find code about processing masks generated by Geometry Decoder for Topological Decoder

Hi Ruizhe, the logic you mentioned is in the inference.py. model.py contains mainly training code. As mentioned in the paper, during training, the model uses "teacher forcing", which adds noise...