/MXNet-EfficientNet

A Gluon Implement of EfficientNet

Primary LanguagePythonApache License 2.0Apache-2.0

MXNet-EfficientNet

A Gluon Implement of EfficientNet

Sorry guys, I think this implement is bugg now. my training efficient-b0 model size is about 40MB, but it should be 20MB with float32, I'm seeking this bug. If you know where it is, tell me, please. Thank you all.

Tmp Usage

I will complete the usage in some days

used in insightface

Reference

Efficient Net