PAA

open-mmlab / mmdetection

Last updated on Feb 23, 2021
PAA (R-101-FPN, 12e, score voting=False)
Memory (M) 6200.0
Backbone Layers 101
Training Data MS COCO
Training Resources 8x NVIDIA V100 GPUs
Training Time
Architecture ResNet, FPN, Probabilistic Anchor Assignment
lr sched 12e
Memory (M) 6200.0
Backbone Layers 101
SHOW MORE
SHOW LESS
PAA (R-101-FPN, 12e, score voting=True)
Memory (M) 6200.0
Backbone Layers 101
File Size 196.16 MB
Training Data MS COCO
Training Resources 8x NVIDIA V100 GPUs
Training Time
Architecture ResNet, FPN, Probabilistic Anchor Assignment
lr sched 12e
Memory (M) 6200.0
Backbone Layers 101
SHOW MORE
SHOW LESS
PAA (R-101-FPN, 24e, score voting=True)
Memory (M) 6200.0
Backbone Layers 101
File Size 196.16 MB
Training Data MS COCO
Training Resources 8x NVIDIA V100 GPUs
Training Time
Architecture ResNet, FPN, Probabilistic Anchor Assignment
lr sched 24e
Memory (M) 6200.0
Backbone Layers 101
SHOW MORE
SHOW LESS
PAA (R-101-FPN, 36e, score voting=True)
Memory (M) 6200.0
Backbone Layers 101
File Size 196.16 MB
Training Data MS COCO
Training Resources 8x NVIDIA V100 GPUs
Training Time
Architecture ResNet, FPN, Probabilistic Anchor Assignment
lr sched 36e
Memory (M) 6200.0
Backbone Layers 101
SHOW MORE
SHOW LESS
PAA (R-50-FPN, 12e, score voting=False)
Memory (M) 3700.0
Backbone Layers 50
Training Data MS COCO
Training Resources 8x NVIDIA V100 GPUs
Training Time
Architecture ResNet, FPN, Probabilistic Anchor Assignment
lr sched 12e
Memory (M) 3700.0
Backbone Layers 50
SHOW MORE
SHOW LESS
PAA (R-50-FPN, 12e, score voting=True)
Memory (M) 3700.0
Backbone Layers 50
File Size 123.47 MB
Training Data MS COCO
Training Resources 8x NVIDIA V100 GPUs
Training Time
Architecture ResNet, FPN, Probabilistic Anchor Assignment
lr sched 12e
Memory (M) 3700.0
Backbone Layers 50
SHOW MORE
SHOW LESS
PAA (R-50-FPN, 18e, score voting=False)
Memory (M) 3700.0
Backbone Layers 50
Training Data MS COCO
Training Resources 8x NVIDIA V100 GPUs
Training Time
Architecture ResNet, FPN, Probabilistic Anchor Assignment
lr sched 18e
Memory (M) 3700.0
Backbone Layers 50
SHOW MORE
SHOW LESS
PAA (R-50-FPN, 18e, score voting=True)
Memory (M) 3700.0
Backbone Layers 50
File Size 123.47 MB
Training Data MS COCO
Training Resources 8x NVIDIA V100 GPUs
Training Time
Architecture ResNet, FPN, Probabilistic Anchor Assignment
lr sched 18e
Memory (M) 3700.0
Backbone Layers 50
SHOW MORE
SHOW LESS
PAA (R-50-FPN, 24e, score voting=True)
Memory (M) 3700.0
Backbone Layers 50
File Size 123.47 MB
Training Data MS COCO
Training Resources 8x NVIDIA V100 GPUs
Training Time
Architecture ResNet, FPN, Probabilistic Anchor Assignment
lr sched 24e
Memory (M) 3700.0
Backbone Layers 50
SHOW MORE
SHOW LESS
PAA (R-50-FPN, 36e, score voting=True)
Memory (M) 3700.0
Backbone Layers 50
File Size 123.47 MB
Training Data MS COCO
Training Resources 8x NVIDIA V100 GPUs
Training Time
Architecture ResNet, FPN, Probabilistic Anchor Assignment
lr sched 36e
Memory (M) 3700.0
Backbone Layers 50
SHOW MORE
SHOW LESS
README.md

Probabilistic Anchor Assignment with IoU Prediction for Object Detection

[ALGORITHM]

@inproceedings{paa-eccv2020,
  title={Probabilistic Anchor Assignment with IoU Prediction for Object Detection},
  author={Kim, Kang and Lee, Hee Seok},
  booktitle = {ECCV},
  year={2020}
}

Results and Models

We provide config files to reproduce the object detection results in the ECCV 2020 paper for Probabilistic Anchor Assignment with IoU Prediction for Object Detection.

Backbone Lr schd Mem (GB) Score voting box AP Config Download
R-50-FPN 12e 3.7 True 40.4 config model | log
R-50-FPN 12e 3.7 False 40.2 -
R-50-FPN 18e 3.7 True 41.4 config model | log
R-50-FPN 18e 3.7 False 41.2 -
R-50-FPN 24e 3.7 True 41.6 config model | log
R-50-FPN 36e 3.7 True 43.3 config model | log
R-101-FPN 12e 6.2 True 42.6 config model | log
R-101-FPN 12e 6.2 False 42.4 -
R-101-FPN 24e 6.2 True 43.5 config model | log
R-101-FPN 36e 6.2 True 45.1 config model | log

Note:

  1. We find that the performance is unstable with 1x setting and may fluctuate by about 0.2 mAP. We report the best results.

Results

Object Detection on COCO minival
MODEL BOX AP
PAA (R-101-FPN, 36e, score voting=True) 45.1
PAA (R-101-FPN, 24e, score voting=True) 43.5
PAA (R-50-FPN, 36e, score voting=True) 43.3
PAA (R-101-FPN, 12e, score voting=True) 42.6
PAA (R-101-FPN, 12e, score voting=False) 42.4
PAA (R-50-FPN, 24e, score voting=True) 41.6
PAA (R-50-FPN, 18e, score voting=True) 41.4
PAA (R-50-FPN, 18e, score voting=False) 41.2
PAA (R-50-FPN, 12e, score voting=True) 40.4
PAA (R-50-FPN, 12e, score voting=False) 40.2
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