EfficientDet.Pytorch issues

Results 101 EfficientDet.Pytorch issues

Sort by recently updated

Incorrect feature map extraction

One of the large differences between this and the original implementation is this implementation follows the diagram in the original paper and extracts feature levels 3, 4, 5, 6 and...

rmcavoy

Training results in detections at all anchorboxes

After training on the VOC2007 dataset with default options except epochs=5, and efficientdet-d2, I am getting detections of all classes at all anchorboxes when using both eval.py and demo.py. Changing...

aaronrmm

inputs_clone used error

in bifpn.py, the 196th line, as following: 1. w2[2, i] * inputs_clone[i + 1])/(w2[0, i] + w2[1, i] + w2[2, i] + self.eps) inputs_clone should be replace for pathtd. the...

eeric

The official Tensorflow implementation is now available.

https://github.com/google/automl/blob/master/efficientdet/efficientdet_arch.py

rmcavoy

Thank you for uploading. Have you trained on the coco dataset? What is the map of EfficientDet-D0 on the coco dataset?

limengfei3675

It does'nt work for dectecting 1600 objects for the funtion "torch.clamp(classification, min=1e-4, max=1.0 - 1e-4)" in focol_loss

Recently, I found that efficientDet-do didn't work for 1600 objects because cls_loss still didn't downgrade in training. So , i modified the funtion **"torch.clamp(classification, min=1e-4, max=1.0 - 1e-4)" into "torch.clamp(classification,...

y78h11b09