search

keras-team / keras-cv

Industry-strength Computer Vision workflows with Keras
Other
976 stars 319 forks source link

Migrate DenseNet to the new Backbone model style #1624

Closed LukeWood closed 1 year ago

LukeWood commented 1 year ago

To align KerasCV with the KerasNLP API, we're migrating our backbone models to subclass keras_cv.models.Backbone. This allows us to have a more structured, modular API for high-level task models, including the new ImageClassifier model (see #1458).

A great step-by-step guide to this migration can be found in the description of #1508, which performed this migration for ResNetV1.

If you'd like to take this issue up, please leave a comment and the KerasCV team can assign the issue to you. When opening a PR, please tag this issue as fixed by your PR, and tag our tracking bug #1627 for bookkeeping.

soma2000-lang commented 1 year ago

@LukeWood I want to take this up .Please assign this to me.

NiharJani2002 commented 1 year ago

@LukeWood, Can I Take this issue? Assuming this issue is still open and now solved yet.

aseemk98 commented 1 year ago

Hello @LukeWood , I've been working on this issue referring to #1508, I'm confused with step 4 of creating preset files for densenet backbone. I can see the weight file URL in weights.py but what about the rest of the information for the no weights preset?

ID6109 commented 1 year ago

Considering this is a high priority, I think closing this soon is best. I'll make a ready-for-review PR by the weekend.

jbischof commented 1 year ago

Thanks @ID6109!