run with error

[hktxt]

totally use your code, does not change any thing. but error comes out....
here is error info
C:\Users\Max\Anaconda3\envs\Pytorch\lib\site-packages\torch\nn\modules\upsampling.py:122: UserWarning: nn.Upsampling is deprecated. Use nn.functional.interpolate instead.
warnings.warn("nn.Upsampling is deprecated. Use nn.functional.interpolate instead.")

RuntimeError Traceback (most recent call last)
in ()
1 model = Darknet("cfg/yolov3.cfg")
2 inp = get_test_input()
----> 3 pred = model(inp, torch.cuda.is_available())
4 print (pred)

~\Anaconda3\envs\Pytorch\lib\site-packages\torch\nn\modules\module.py in call(self, *input, **kwargs)
475 result = self._slow_forward(*input, **kwargs)
476 else:
--> 477 result = self.forward(*input, **kwargs)
478 for hook in self._forward_hooks.values():
479 hook_result = hook(self, input, result)

in forward(self, x, CUDA)
216 #Transform
217 x = x.data
--> 218 x = predict_transform(x, inp_dim, anchors, num_classes, CUDA)
219 if not write: #if no collector has been intialised.
220 detections = x

F:\condaDev\util.ipynb in predict_transform(prediction, inp_dim, anchors, num_classes, CUDA)

RuntimeError: invalid argument 2: size '[1 x 255 x 3025]' is invalid for input with 689520 elements at ..\aten\src\TH\THStorage.cpp:84

[cumtcsys]

I have encountered the same problem as you,after changed the third line "grid_size = inp_dim // stride" to "grid_size = prediction.size(2)" in function predict_transform,the problem fixed.

[ghostPath]

I have encountered the same problem as you,after changed the third line "grid_size = inp_dim // stride" to "grid_size = prediction.size(2)" in function predict_transform,the problem fixed.

but the result I got is different with the blog

[ghostPath]

I have encountered the same problem as you,after changed the third line "grid_size = inp_dim // stride" to "grid_size = prediction.size(2)" in function predict_transform,the problem fixed.

I konw why. Thx

[allenwu5]

I have encountered the same problem as you,after changed the third line "grid_size = inp_dim // stride" to "grid_size = prediction.size(2)" in function predict_transform,the problem fixed.

I konw why. Thx

Hi @ghostPath , I got the different result as well.
Would you mind to share your insight ?

[ghostPath]

I have encountered the same problem as you,after changed the third line "grid_size = inp_dim // stride" to "grid_size = prediction.size(2)" in function predict_transform,the problem fixed.

I konw why. Thx

Hi @ghostPath , I got the different result as well.
Would you mind to share your insight ?

我觉得是因为权重是随机初始化的？

[allenwu5]

I have encountered the same problem as you,after changed the third line "grid_size = inp_dim // stride" to "grid_size = prediction.size(2)" in function predict_transform,the problem fixed.

I konw why. Thx

Hi @ghostPath , I got the different result as well.
Would you mind to share your insight ?

我觉得是因为权重是随机初始化的？

Thank you @ghostPath .
I think you're right.
I just found related paragraph:

At this point, our network has random weights, and will not produce the correct output. We need to load a weight file in our network. We'll be making use of the official weight file for this purpose.

[pypeaday]

@ayooshkathuria can you please update the blog and close this issue? The code base and tutorial both have grid_size = inp_dim//stride which leads to the error mentioned in this issue.

[feng3245]

I have encountered the same problem as you,after changed the third line "grid_size = inp_dim // stride" to "grid_size = prediction.size(2)" in function predict_transform,the problem fixed.

but the result I got is different with the blog

It's just random weights. It's expected to be random and different