liangheming / fcosv1

FCOS ,640px(max side),mAP 38.2,fps45.45(RTX 2080TI).faster! better! <<Fully Convolutional One-Stage Object Detection>>

FCOS

This is an unofficial pytorch implementation of FCOS object detection as described in FCOS: Fully Convolutional One-Stage Object Detection by Zhi Tian,Chunhua Shen, Hao Chen, Kaiming He and Tong He.

requirement

tqdm
pyyaml
numpy
opencv-python
pycocotools
torch >= 1.6
torchvision >=0.7.0

result

we trained this repo on 4 GPUs with batch size 32(8 image per node).the total epoch is 24(about 180k iter),Adam with cosine lr decay is used for optimizing. finally, this repo achieves 38.0 mAp at 640px(max side) resolution with resnet50 backbone(no center sample).you can update the param "radius" to activate this setting.

 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.382
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.565
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.409
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.183
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.434
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.542
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.313
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.497
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.534
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.278
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.615
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.725

difference from original implement

the main difference is about the input resolution.the original implement use min_thresh and max_thresh to keep the short side of the input image larger than min_thresh while keep the long side smaller than max_thresh.for simplicity we fix the long side a certain size, then we resize the input image while keep the width/height ratio, next we pad the short side.the final width and height of the input are same.

training

for now we only support coco detection data.

COCO

• modify main.py (modify config file path)

from solver.ddp_mix_solver import DDPMixSolver
if __name__ == '__main__':
    processor = DDPMixSolver(cfg_path="your own config path") 
    processor.run()

• custom some parameters in config.yaml

model_name: fcos
data:
  train_annotation_path: data/annotations/instances_train2017.json
#  train_annotation_path: data/annotations/instances_val2017.json
  val_annotation_path: data/annotations/instances_val2017.json
  train_img_root: data/train2017
#  train_img_root: data/val2017
  val_img_root: data/val2017
  max_thresh: 640
  use_crowd: False
  batch_size: 8
  num_workers: 4
  debug: False
  remove_blank: Ture

model:
  num_cls: 80
  strides: [8, 16, 32, 64, 128]
  backbone: resnet50
  pretrained: True
  alpha: 0.25
  gamma: 2.0
  iou_type: giou
  radius: 0
  layer_limits: [64, 128, 256, 512]
  iou_loss_weight: 0.5
  reg_loss_weight: 1.3
  conf_thresh: 0.05
  nms_iou_thresh: 0.6
  max_det: 300
optim:
  optimizer: Adam
  lr: 0.0001
  milestones: [18,24]
  warm_up_epoch: 0
  weight_decay: 0.0001
  epochs: 24
  sync_bn: True
  amp: True
val:
  interval: 1
  weight_path: weights


gpus: 0,1,2,3

• run train scripts

nohup python -m torch.distributed.launch --nproc_per_node=4 main.py >>train.log 2>&1 &

TODO

• Color Jitter
• Perspective Transform
• Mosaic Augment
• MixUp Augment
• IOU GIOU DIOU CIOU
• Warming UP
• Cosine Lr Decay
• Center Sample
• EMA(Exponential Moving Average)
• Mixed Precision Training (torch native amp)
• Sync Batch Normalize
• PANet(neck)
• BiFPN(EfficientDet neck)
• VOC data train\test scripts
• custom data train\test scripts
• MobileNet Backbone support