automl VS simple-faster-rcnn-pytorch

I'm using Google Brain's EfficientDet repo to train an object detector. What hyperparameters should I choose to increase accuracy for small objects.

EfficientDet-D0 has comparable accuracy as YOLOv3.

Code for https://arxiv.org/abs/2104.00298 found: https://github.com/google/automl/efficientnetv2

7 Min Read | Paper (CoAtNet) | Paper (EfficientNetV2) | Google blog | Code

Abstract: This paper introduces EfficientNetV2, a new family of convolutional networks that have faster training speed and better parameter efficiency than previous models. To develop this family of models, we use a combination of training-aware neural architecture search and scaling, to jointly optimize training speed and parameter efficiency. The models were searched from the search space enriched with new ops such as Fused-MBConv. Our experiments show that EfficientNetV2 models train much faster than state-of- the-art models while being up to 6.8x smaller. > Our training can be further sped up by progressively increasing the image size during training, but it often causes a drop in accuracy. To compensate for this accuracy drop, we propose to adaptively adjust regularization (e.g., dropout and data augmentation) as well, such that we can achieve both fast training and good accuracy. > With progressive learning, our EfficientNetV2 significantly outperforms previous models on ImageNet and CIFAR/Cars/Flowers datasets. By pretraining on the same ImageNet21k, our EfficientNetV2 achieves 87.3% top-1 accuracy on ImageNet ILSVRC2012, outperforming the recent ViT by 2.0% accuracy while training 5x-11x faster using the same computing resources. Code will be available at this https URL.

Hi all, as the title suggests I'm looking for a GitHub repo where I can edit a Faster RCNN implementation rather easily. I'm basically looking to test an idea where I have multiple branches with feature map and bounding boxes as inputs. I've modified the built-in torchvision implementation once before, but I think it's a little more complicated than I like, and I'd rather not release the entire torchvision package as part of my own work in the future. I have looked briefly into this repo https://github.com/chenyuntc/simple-faster-rcnn-pytorch/blob/master/trainer.py but it only supports a batch size of 1, and I'm not sure what it'd take to expand that capability. Is there anything better out there?