-
Notifications
You must be signed in to change notification settings - Fork 263
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Request - integrate code for all-in-one nets #118
Comments
Updates:
|
@lukeyeager may I close it? |
Actually, it doesn't look like we've pulled in BVLC#3863 after all. You can close this as will-not-fix if you like, but without those changes we still can't really do all-in-one nets. Of course, based on the lack of progress in BVLC#3864, it doesn't seem like there's much of a desire for this feature anyway. |
I'm new to caffe, but I don't understand why there isn't more enthusiasm about this feature. To me it seems unnecessarily convoluted and error prone to require maintaining two nigh identical prototxt files. Or am I missing some alternative way of keeping my train and deploy (or test) networks properly aligned? |
Keeping this in TODO list.. |
Done in v0.17.0 |
BVLC#1245 is nearing completion.
input
s BVLC/caffe#3211 adds support forInput
layersOnce all-in-one nets are fully supported, the DIGITS team would like to make use of these new features in order to specify networks all in a single file instead of having separate
train_val
anddeploy
networks.The text was updated successfully, but these errors were encountered: