Res2Net / Res2Net-PretrainedModels

(ImageNet pretrained models) The official pytorch implemention of the TPAMI paper "Res2Net: A New Multi-scale Backbone Architecture"
https://mmcheng.net/res2net/
1.07k stars 215 forks source link

维度不匹配 #63

Closed Lan2Kailee closed 2 years ago

Lan2Kailee commented 2 years ago

首先谢谢您的代码,我在运行过程中出现了这样的问题,请问是什么原因呢? PS D:\centernet\centernet> & D:/Anaconda3/envs/pytorch/python.exe d:/centernet/centernet/train.py Load weights model_data/res2net50_v1b_26w_4s-3cf99910.pth. Start Train Epoch 1/50: 0%| | 0/1034 [00:13<?, ?it/s<class 'dict'>] Traceback (most recent call last): File "d:/centernet/centernet/train.py", line 215, in epoch_step, epoch_step_val, gen, gen_val, end_epoch, Cuda, backbone) File "d:\centernet\centernet\utils\utils_fit.py", line 33, in fit_one_epoch hm, wh, offset = model_train(batch_images) File "D:\Anaconda3\envs\pytorch\lib\site-packages\torch\nn\modules\module.py", line 547, in call result = self.forward(*input, kwargs) File "d:\centernet\centernet\nets\centernet.py", line 31, in forward feat = self.backbone(x) File "D:\Anaconda3\envs\pytorch\lib\site-packages\torch\nn\modules\module.py", line 547, in call result = self.forward(*input, *kwargs) File "D:\Anaconda3\envs\pytorch\lib\site-packages\torch\nn\modules\container.py", line 92, in forward input = module(input) File "D:\Anaconda3\envs\pytorch\lib\site-packages\torch\nn\modules\module.py", line 547, in call result = self.forward(input, kwargs) File "D:\Anaconda3\envs\pytorch\lib\site-packages\torch\nn\modules\container.py", line 92, in forward input = module(input) File "D:\Anaconda3\envs\pytorch\lib\site-packages\torch\nn\modules\module.py", line 547, in call result = self.forward(*input, **kwargs) File "d:\centernet\centernet\nets\resnet50.py", line 81, in forward sp = sp + spx[i] RuntimeError: The size of tensor a (64) must match the size of tensor b (128) at non-singleton dimension 3

kirkzZ commented 2 years ago

你好,我也遇到了这个问题,你这个问题解决了吗

gasvn commented 2 years ago

https://github.com/Res2Net/Res2Net-PretrainedModels/issues/68