NAS-FPN

open-mmlab / mmdetection

Last updated on Feb 23, 2021
NAS-FPN (R-50-FPN, 50e)
Memory (M) 12900.0
inference time (s/im) 0.04367
File Size 217.31 MB
Training Data MS COCO
Training Resources 8x NVIDIA V100 GPUs
Training Time
Architecture ResNet, FPN, NAS-FPN
lr sched 50e
Memory (M) 12900.0
Backbone Layers 50
inference time (s/im) 0.04367
SHOW MORE
SHOW LESS
NAS-FPN (R-50-NASFPN, 50e)
Memory (M) 13200.0
inference time (s/im) 0.04348
File Size 301.35 MB
Training Data MS COCO
Training Resources 8x NVIDIA V100 GPUs
Training Time
Architecture ResNet, FPN, NAS-FPN
lr sched 50e
Memory (M) 13200.0
Backbone Layers 50
inference time (s/im) 0.04348
SHOW MORE
SHOW LESS
README.md

NAS-FPN: Learning Scalable Feature Pyramid Architecture for Object Detection

Introduction

[ALGORITHM]

@inproceedings{ghiasi2019fpn,
  title={Nas-fpn: Learning scalable feature pyramid architecture for object detection},
  author={Ghiasi, Golnaz and Lin, Tsung-Yi and Le, Quoc V},
  booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
  pages={7036--7045},
  year={2019}
}

Results and Models

We benchmark the new training schedule (crop training, large batch, unfrozen BN, 50 epochs) introduced in NAS-FPN. RetinaNet is used in the paper.

Backbone Lr schd Mem (GB) Inf time (fps) box AP Config Download
R-50-FPN 50e 12.9 22.9 37.9 config model | log
R-50-NASFPN 50e 13.2 23.0 40.5 config model | log

Note: We find that it is unstable to train NAS-FPN and there is a small chance that results can be 3% mAP lower.

Results

Object Detection on COCO minival

Object Detection
BENCHMARK MODEL METRIC NAME METRIC VALUE GLOBAL RANK
COCO minival NAS-FPN (R-50-NASFPN, 50e) box AP 40.5 # 69
COCO minival NAS-FPN (R-50-FPN, 50e) box AP 37.9 # 95
Contact us on: hello@paperswithcode.com . Papers With Code is a free resource with all data licensed under CC-BY-SA.
Terms Data policy Cookies policy