DaiShiResearch/TransNeXt

Issues

• About Grand Cam

  #18 opened 5 months ago by Are1ana
  1

• Which part of the code uses the test set to evaluate the model results?

  #28 opened 2 months ago by learnmat
  0

• Robustness test

  #27 opened 2 months ago by LMMMEng
  0

• Issue encountered when utilizing 'nn.Conv2d' for training my own dataset on MMSegmentation.

  #26 opened 3 months ago by ZhoumingZhang
  0

• 加载预训练模型unexpected key in source state_dict:

  #25 opened 3 months ago by llyyccccc
  0

• Why is there no shortcut connection in the forward of ConvolutionalGLU code?

  #23 opened 4 months ago by codequan
  1

• A Warning:/usr/bin/ld: skipping incompatible /usr/lib/libcuda.so when searching for -lcuda

  #24 opened 3 months ago by LeoRainly
  0

• Thanks for u works, i have a small question as below.

  #9 opened 7 months ago by Fortuneteller6
  5

• Visualization using overlapping patches in TransNeXt might conflict with DINO’s code assumptions

  #22 opened 4 months ago by JuergenZhang
  0

• About the data

  #10 opened 8 months ago by huaxiaochun
  2

• OOD validation code

  #21 opened 4 months ago by XavierHeart
  0

• GPU memory on training the segmentation model

  #19 opened 5 months ago by KKIverson
  0

• Test the TransNext model

  #20 opened 5 months ago by JuergenZhang
  0

• RuntimeError: The size of tensor a (196) must match the size of tensor b (3136) at non-singleton dimension 2

  #8 opened 8 months ago by HKLCXQ
  1

• About foveal_peripheral_vision.jpg

  #17 opened 6 months ago by AKAxiaobai-123
  0

• Problems with the building the swattention !

  #16 opened 6 months ago by xyb1314
  0

• Uncontrollable increase in GPU graphics memory

  #15 opened 6 months ago by chenller
  0

• ONNX file conversion

  #14 opened 6 months ago by shiomi326
  2

• About attention_cuda.py and attention_native.py files

  #13 opened 7 months ago by logicvanlyf
  1

• About the FLOPs of detection and segmenation

  #11 opened 7 months ago by yuweihao
  6

• Error when using non-standard image sizes (e.g., 138x138) with TransNeXt Tiny model

  #12 opened 7 months ago by Coolog
  1

• File "D:\code\TransNeXt\transnext_native.py", line 234, in forward attn_pool = q_norm_scaled @ c.transpose(-2, -1) + pool_bias RuntimeError: Expected size for first two dimensions of batch2 tensor to be: [147, 24] but got: [147, 49].

  #5 opened 8 months ago by aoe1223
  1

• Questions about the theory in the article

  #6 opened 8 months ago by 895318
  2

• CUDA layout

  #7 opened 8 months ago by Liuhp133
  3

• A bug?

  #1 opened 8 months ago by luckydog11
  4

• ImageNet Pretrain

  #2 opened 8 months ago by kkkkkk123-ops
  1

• pretrained weight

  #3 opened 8 months ago by HC-Zhou
  1

• Parameter setting problem

  #4 opened 8 months ago by hero-White
  1