Fix Densenet module keys having dot #474
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
fmassa
approved these changes
Apr 21, 2018
|
@fmassa Yes, some models might be broken. But those are only the ones that use |
|
I'm new to pyTorch , sorry if my question is redundant. Has the solution already been provided for the issue ? I'm using the newly released pyTorch 0.4 , |
|
Yes, just reinstall torchvision. |
|
Hi @lanpa I am getting the above error when I use model.load_local_state_dict . And also when I use the previously used load_state_dict . |
|
@kish7 You should load state dict before |
|
@lanpa And still getting an error |
|
load_state_dict, not load_local_state_dict
…On Fri, Apr 27, 2018 at 3:09 PM kish7 ***@***.***> wrote:
@lanpa <https://github.com/lanpa>
I changed the order of execution to
modelCheckpoint = torch.load(pathModel)
model.load_local_state_dict(modelCheckpoint['state_dict'])
model = torch.nn.DataParallel(model).cuda()
And still getting an error
File "C:\Users\Kishore\Anaconda3\lib\site-packages\torch\nn\
modules\module.py", line 532, in __getattr__
type(self).__name__, name))
AttributeError: 'DenseNet121' object has no attribute 'load_local_state_dict'
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#474 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AFaWZaABCp-zwYg7H0ytUnydOnZlBAX0ks5tssQxgaJpZM4Teiq8>
.
|
|
@ssnl Takes me back to the same error as before . |
|
It looks like you are using a different model that is not provided by |
|
Hi @fmassa , This is how DensenetModels.py looks like . You can see that model is being imported from torchvision. Thank You. |
|
@kish7 sorry for the late reply. densenet121 = torchvision.models.densenet121(pretrained=True)If it still gives the same error, and if you are using the latest torchvision, I'd say it's a problem with our Windows version, but I don't have any other ideas yet |
|
hi @kish7 Eyal |
|
Hi Eyal and @kish7 , i believe that you should update to torchvision 0.2.1, which includes this patch that should fix the issue. For loading checkpoints before 0.4/0.2.1, I think it unfortunately won’t work directly. You can apply similar changes to the state dict as I did in the patch.
… On Jun 11, 2018, at 05:07, EyalRozenberg1 ***@***.***> wrote:
hi @kish7
Did you manage to solve the issue?
Eyal
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub, or mute the thread.
|
varunagrawal
pushed a commit
to varunagrawal/vision
that referenced
this pull request
Jul 23, 2018
`Traceback (most recent call last):
|
|
@IlSLY the solution is to rename the |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
After pytorch/pytorch#6639, '.'s are no longer allowed in module keys.
cc @lanpa