Model zoo with Pretrained VGG16/19 #21
Closed
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
|
Also, this needed the |
|
I think that we should remove json files from github. |
- The `mode` now defaults to `cross`, which is what all other packages
use and thus it allows loading their weights without needing to fiddle
with mirroring/flipping them correctly.
- Adds an alternative `border` parameter for these two reasons:
1. So that we have a flexible shortcut for 'same'
2. So that Theano can use some tricks for 'valid' and 'full'
The old interface is still available and used since `border=None` by default.
This change is fully backwards compatible.
(I need it for ILSVRC-pretrained VGG)
|
Yeah, I felt uncomfortable adding them, but for some reason didn't think of putting them into a github gist and using the I've also undone the change to the |
This includes implementing a slightly fancier download functionality which caches the data in a `~/.cache/beacon8` folder. For the example code, this also necessitated writing `imread` and `imresize` functions which use either `cv2` or `PIL` depending on what's available.
|
Interface is still different from Torch7. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
Adds a model "zoo" à la Caffe. Currently, this contains VGG16 and VGG19, it allows to just create the model (and then train yourself) or download the pretrained ILSVRC weights from the internet and create a pretrained version of the net. The latter is done in the example.
Again, I made the commits so that it's best to read the commits individually, the commit message contains a lot of information.
The two large
jsonfiles will move from the example to thedatasubmodule, whose PR will follow in a few days.Looking forward to your feedback; especially, I think you may not like the change of parameters to
SpatialConvolutionCUDNN?