-
Hi Author,
First thanks for your great work to improve the convergence speed of DETR with such a large margin. When reading the paper, I get a little bit confused on how do you exactly draw the at…
-
### Problem
Hi,
this is related to https://github.com/sofa-framework/sofa/issues/3201.
Same as before, I want to cut a deformable object.
I wanted to check, if it is easier in total to replace…
-
I can't understand the way of the reference point generation. In the `DeformableDetrTransformer` class, the `get_reference_points` is used to calculate the initial reference point in the encode stage,…
-
Error compiling objects for extension in CUDA11.1, pytorch1.7, RTX3090
ninja: build stopped: subcommand failed.
Traceback (most recent call last):
File "D:\anaconda3\envs\pytorch\lib\site-packa…
-
I used many other network models to replace the original Mobile net, such as ghostnet, Octmobile net, and so on. But their losses fall slowly and remain constant for a long time. Is there any way to …
-
my environment:
```
# packages in environment at /fastdeploy/my_soft/anaconda3/envs/tformer:
#
# Name Version Build Channel
_libgcc_mutex 0.1 …
-
**Checklist**
1. I have searched related issues but cannot get the expected help.
2. I have read the [FAQ documentation](https://mmdetection.readthedocs.io/en/latest/faq.html) but cannot get the e…
-
Hi, Thank you for your good work!
Recently i am studying DETR-like, and i have questions as following:
1.there is no implement about reference point iterative refinement , but many other works d…
-
### Duplicates
- [X] I have searched the existing issues
### Steps to reproduce 🕹
_No response_
### Current behavior 😯
When using Chinese text, the length increases after encoding, which may caus…
-
Q: I trained a model with a dataset in coco format that I made myself, and now have two problems: 1) The score is high on the validation set, but low for other datasets of the same type. This doesn't …