# Renet-runtime and training execusion fails

**URL:** <https://discuss.pytorch.org/t/renet-runtime-and-training-execusion-fails/116065>\
**Category:** glow\
**Created:** [March 26, 2021, 10:58am UTC](https://discuss.pytorch.org/t/renet-runtime-and-training-execusion-fails/116065 "2021-03-26T10:58:19Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![jr.r](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/j/9fc348/32.png) [@jr.r](https://discuss.pytorch.org/u/jr.r)\
**Post date:** [March 26, 2021, 10:58am UTC](https://discuss.pytorch.org/t/renet-runtime-and-training-execusion-fails/116065/1 "2021-03-26T10:58:19Z")

</div>

Hi, I am trying to execute resnet-runtime example provided in the glow. Here I am getting the following error:  
**$./bin/resnet-runtime /test/imagenet/n04548280/ -backend=CPU -max-images=1**

glow/lib/Base/Tensor.h:422: glow::Tensor glow::Tensor::getUnowned(llvm::ArrayRef\<dim\_t\>, llvm::ArrayRef\<dim\_t\>) const: Assertion `actualSize() == unownedTensor.actualSize() && "The size of the unowned tensor " "should be the same as the size of " “the original tensor”’ failed.  
Aborted (core dumped)

and I got another error while executing the resenet-training example as below

**$./bin/resnet-training /test/image\_path\_list.txt -backend=CPU**  
Loading resnet50 model.  
WARNING: Logging before InitGoogleLogging() is written to STDERR  
I0327 14:22:28.656437 11843 Graph.cpp:5912] Layout requirements checking is enabled  
Preparing for training.  
Differentiating graph.  
Compiling backend.  
I0327 14:22:28.959044 11843 Partitioner.cpp:482] The model is too small for applying partition.  
Model size : 103058848  
Backend Name : CPU  
Device memory: 2000000000  
I0327 14:22:33.969435 11843 Partitioner.cpp:88] The number of partitions is : 2  
I0327 14:22:33.969450 11843 PartitionerUtils.cpp:549] Partition 0:  
Name : resnet50  
BackendKind : CPU  
context count : 1  
total Memory : 103062848  
input size: 103042720  
input count : 268  
input only from peers count : 267  
output size: 4000  
constant size: 16128  
I0327 14:22:33.969477 11843 PartitionerUtils.cpp:570] LogicalDeviceIDs : 0  
I0327 14:22:33.969482 11843 PartitionerUtils.cpp:549] Partition 1:  
Name : resnet50\_grad  
BackendKind : CPU  
context count : 1  
total Memory : 409959712  
input size: 205270848  
input count : 429  
input only from peers count : 0  
output size: 204672736  
constant size: 16128  
I0327 14:22:33.969525 11843 PartitionerUtils.cpp:570] LogicalDeviceIDs : 0  
F0327 14:23:07.816742 11843 Image.cpp:638] Check failed: std::equal(localCopy.dims().begin(), localCopy.dims().end(), inputImageData.dims().begin() + 1) All images must have the same dimensions  
\*\*\* Check failure stack trace: \*\*\*  
Aborted (core dumped)

All the images are png format with resolution 300x300. I coun’t get the issue here.

Thank you

---

<div class="post-metadata">

**Author:** ![jfix](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/jfix/32/21510_2.png) [@jfix](https://discuss.pytorch.org/u/jfix)\
**Post date:** [April 2, 2021, 6:03am UTC](https://discuss.pytorch.org/t/renet-runtime-and-training-execusion-fails/116065/2 "2021-04-02T06:03:04Z")

</div>

I think this might have to do with the image resolution, or something else about the input images. Can you try with the images located at `glow/tests/images/imagenet/` to see if that works?