-
Reference to the paper's decription:
'Bias parameters of the linear projection are initialized to make A(mlqk) = 1/LK'
But in the code implementation:
https://github.com/fundamentalvision/Defor…
-
Just putting some suggestions out there. Maybe we could organize these into [projects](https://github.com/dlib-users/darknet/projects).
- [x] support scaled yolov4
- [x] support yolov5 models
-…
-
Hi guys, it is a wonderful work.
Regarding the latency, I have two confusions:
1) When paired with HGNetv2-L and HGNetv2-X, the RT-DETRs achieve 114 and 74 FPS, respectively. However, the latency…
-
## 🐞Describing the bug
I get the error `Unable to map torch_upsample_nearest_neighbor to core upsample` when I try to convert the DETR PyTorch model. I tried to go deep into the package to see that t…
-
![yolo_v5_tb](https://user-images.githubusercontent.com/30591790/87405610-a50fe6…
-
Is there anyone who encountered the same issue as me?
When I was fine-tuning deformable-detr with a custom dataset, the cost of graphics memory gradually increased from 11G to 29G during training.
I…
ghost updated
2 years ago
-
Hi, Happy Spring Festival, thx for your great work!
I did a performance analysis on the inference of the torch code and it seems that the `reshape` operation in `FrozenBatchNorm2d` (`src/nn/backbon…
-
model = dict(
type='FCOS',
pretrained=None,
backbone=dict(
type='SwinTransformer',
embed_dim=96,
depths=[2, 2, 6, 2],
num_heads=[3, 6, 12, 24],
…
-
使用的配置文件是co_dino_5scale_vit_large_coco.py这个配置，test_cfg的配置我把score_thr设置为0.9仍然没什么用，推理出来的框有1000个，就是设置的max_per_img=1000，好像score_thr和nms没有发挥作用？
```
test_cfg=[
dict(
max_per_img=1000,…
-
Hello I am facing following error when I tried to train the model using custom dataset that has 5 classes. I know the error below would majorly occur because of difference in num_classes, but not sure…