InternImage

InternImage copied to clipboard

How to install InternImage/detection with A100

2

I already installed python 3.9 pytorch 2.0.1 torchvision 0.15.2 torchaudio 2.0.2 cuda 11.8 and I command pip install -U openmim mim install mmcv-full==1.5.0 then, I faced this situation Installing mmcv-full...

Deokhwan-Kyeong

xgddyrq

sampling_locations = (ref + grid * offset_scale).repeat(N_, 1, 1, 1, 1).flatten(3, 4) + \ RuntimeError: The size of tensor a (16) must match the size of tensor b (32) at non-singleton dimension 2

4

tensor at non-singleton dimension 2

Result1

a bug when converting the dino detector to tensorrt

2

Does anyone successfully export DINO from pytorch to tensorrt? Please share your experience. There is the bug (and lots of warnings) in my experience. An error `RuntimeError: Expected all tensors...

goldentimecoolk

failed to convert DINO det model

1

When I convert dino_4scale_internimage_l_1x_coco_0.1x_backbone_lr.pth to onnx, I got KeyError: 'batch_input_shape'. any suggestions?  many thanks, Amber

j0987834204

请问在DCNv3_pytorch代码中： output = (sampling_input_ * mask).sum(-1).view(N_, group*group_channels, H_out*W_out)这一步对应公式里边m_gk * x_g, 而x = self.output_proj(x)这一步对应于w_g * (m_gk * x_g)这一步吗？

lp-094

Do you have a tentative date to release InternImage-H pretrained weights for cityscapes or mapillary or mapillary2cityscapes? Please let me know if it has already been published elsewhere?

aravind3134

I used internimage base segmentation model, however GPU memory of inference is more than training (batch size 1). I am very confused.

ker95qq

Latency too high with DCNv3_pytorch op on cpu

2

Hi guys: when use DCNv3_pytorch for inference, the letency is too high for cpu only device I test with both cuda and cpu infer for cuda device , per image...

xbkaishui

Thank you for this awesome implementation! I've been using the InternImage as a backbone of my Detectron2 Faster-R-CNN implementation. My trained models need to run on CPU, so I tried...

jonassoebro