pytorch-image-models
pytorch-image-models copied to clipboard
MobileDets backbones
@xonobo I looked at this briefly, the mobiledet CPU, GPU and TPU backbone could be added to the efficientnet generator w/ an extra block type (Tucker) and some tweaks, but it doesn't look significantly different from mobilenetv2/v3/mnasnset so I'm not prioritizing it.
https://github.com/tensorflow/models/blob/420a7253e034a12ae2208e6ec94d3e4936177a53/research/object_detection/models/ssd_mobiledet_feature_extractor.py
Relation to blocks here, _fused_conv == EdgeTPU block, _inverterted_bottleneck == InvertedResidual , _inverted_bottleneck_no_expansion == DepthwiseSeparableConv , _tucker_conv is different
With minimal supervision I should be able to work on this. Please could I work on this?