์ฝ˜ํ…์ธ ๋กœ ๊ฑด๋„ˆ๋›ฐ๊ธฐ

ํ…Œ์ŠคํŠธ ์‹œ๊ฐ„ ํ™•๋Œ€(TTA)

์ด ๊ฐ€์ด๋“œ์—์„œ๋Š” YOLOv5 ๐Ÿš€์„ ์‚ฌ์šฉํ•˜์—ฌ mAP ๋ฐ ๋ฆฌ์ฝœ์„ ๊ฐœ์„ ํ•˜๊ธฐ ์œ„ํ•œ ํ…Œ์ŠคํŠธ ๋ฐ ์ถ”๋ก  ์ค‘์— TTA(ํ…Œ์ŠคํŠธ ์‹œ๊ฐ„ ํ™•๋Œ€)๋ฅผ ์‚ฌ์šฉํ•˜๋Š” ๋ฐฉ๋ฒ•์„ ์„ค๋ช…ํ•ฉ๋‹ˆ๋‹ค.

์‹œ์ž‘ํ•˜๊ธฐ ์ „

๋ฆฌํฌ์ง€ํ† ๋ฆฌ๋ฅผ ๋ณต์ œํ•˜๊ณ  ์š”๊ตฌ์‚ฌํ•ญ.txt๋ฅผ ์„ค์น˜ํ•œ๋‹ค. Python>=3.8.0 ํ™˜๊ฒฝ์—์„œ PyTorch>=1.8. ๋ชจ๋ธ๊ณผ ๋ฐ์ดํ„ฐ ์„ธํŠธ๋Š” ์ตœ์‹  YOLOv5 ๋ฆด๋ฆฌ์Šค์—์„œ ์ž๋™์œผ๋กœ ๋‹ค์šด๋กœ๋“œ๋ฉ๋‹ˆ๋‹ค.

git clone https://github.com/ultralytics/yolov5  # clone
cd yolov5
pip install -r requirements.txt  # install

์ •์ƒ์ ์œผ๋กœ ํ…Œ์ŠคํŠธ

TTA๋ฅผ ์‹œ๋„ํ•˜๊ธฐ ์ „์— ๋น„๊ตํ•  ๊ธฐ์ค€ ์„ฑ๋Šฅ์„ ์„ค์ •ํ•˜๊ณ  ์‹ถ์Šต๋‹ˆ๋‹ค. ์ด ๋ช…๋ น์€ ์ด๋ฏธ์ง€ ํฌ๊ธฐ 640ํ”ฝ์…€์˜ COCO val2017์—์„œ YOLOv5x๋ฅผ ํ…Œ์ŠคํŠธํ•ฉ๋‹ˆ๋‹ค. yolov5x.pt ์€ ๊ฐ€์žฅ ํฌ๊ณ  ์ •ํ™•ํ•œ ๋ชจ๋ธ์ž…๋‹ˆ๋‹ค. ๋‹ค๋ฅธ ์˜ต์…˜์€ ๋‹ค์Œ๊ณผ ๊ฐ™์Šต๋‹ˆ๋‹ค. yolov5s.pt, yolov5m.pt ๋ฐ yolov5l.pt๋˜๋Š” ์‚ฌ์šฉ์ž ์ง€์ • ๋ฐ์ดํ„ฐ ์„ธํŠธ ํ•™์Šต์—์„œ ์–ป์€ ์ž์ฒด ์ฒดํฌ ํฌ์ธํŠธ ./weights/best.pt. ์‚ฌ์šฉ ๊ฐ€๋Šฅํ•œ ๋ชจ๋“  ๋ชจ๋ธ์— ๋Œ€ํ•œ ์ž์„ธํ•œ ๋‚ด์šฉ์€ ์‚ฌ์šฉ ์„ค๋ช…์„œ๋ฅผ ์ฐธ์กฐํ•˜์„ธ์š”. ํ…Œ์ด๋ธ”.

python val.py --weights yolov5x.pt --data coco.yaml --img 640 --half

์ถœ๋ ฅ:

val: data=./data/coco.yaml, weights=['yolov5x.pt'], batch_size=32, imgsz=640, conf_thres=0.001, iou_thres=0.65, task=val, device=, single_cls=False, augment=False, verbose=False, save_txt=False, save_hybrid=False, save_conf=False, save_json=True, project=runs/val, name=exp, exist_ok=False, half=True
YOLOv5 ๐Ÿš€ v5.0-267-g6a3ee7c torch 1.9.0+cu102 CUDA:0 (Tesla P100-PCIE-16GB, 16280.875MB)

Fusing layers...
Model Summary: 476 layers, 87730285 parameters, 0 gradients

val: Scanning '../datasets/coco/val2017' images and labels...4952 found, 48 missing, 0 empty, 0 corrupted: 100% 5000/5000 [00:01<00:00, 2846.03it/s]
val: New cache created: ../datasets/coco/val2017.cache
               Class     Images     Labels          P          R     mAP@.5 mAP@.5:.95: 100% 157/157 [02:30<00:00,  1.05it/s]
                 all       5000      36335      0.746      0.626       0.68       0.49
Speed: 0.1ms pre-process, 22.4ms inference, 1.4ms NMS per image at shape (32, 3, 640, 640)  # <--- baseline speed

Evaluating pycocotools mAP... saving runs/val/exp/yolov5x_predictions.json...
...
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.504  # <--- baseline mAP
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.688
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.546
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.351
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.551
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.644
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.382
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.628
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.681  # <--- baseline mAR
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.524
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.735
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.826

TTA๋กœ ํ…Œ์ŠคํŠธํ•˜๊ธฐ

์ถ”๊ฐ€ --augment ๋ฅผ ๊ธฐ์กด val.py ๋ช…๋ น์„ ์‹คํ–‰ํ•˜์—ฌ TTA๋ฅผ ํ™œ์„ฑํ™”ํ•˜๊ณ  ์ด๋ฏธ์ง€ ํฌ๊ธฐ๋ฅผ ์•ฝ 30% ๋Š˜๋ฆฌ๋ฉด ๋” ๋‚˜์€ ๊ฒฐ๊ณผ๋ฅผ ์–ป์„ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. TTA๋ฅผ ํ™œ์„ฑํ™”ํ•œ ์ถ”๋ก ์€ ์ผ๋ฐ˜์ ์œผ๋กœ ์ด๋ฏธ์ง€๋ฅผ ์ขŒ์šฐ๋กœ ๋’ค์ง‘๊ณ  3๊ฐœ์˜ ์„œ๋กœ ๋‹ค๋ฅธ ํ•ด์ƒ๋„๋กœ ์ฒ˜๋ฆฌํ•˜๋ฉฐ ์ถœ๋ ฅ์„ NMS ์ „์— ๋ณ‘ํ•ฉํ•˜๊ธฐ ๋•Œ๋ฌธ์— ์ผ๋ฐ˜ ์ถ”๋ก ๋ณด๋‹ค ์•ฝ 2~3๋ฐฐ์˜ ์‹œ๊ฐ„์ด ์†Œ์š”๋ฉ๋‹ˆ๋‹ค. ์†๋„ ๊ฐ์†Œ์˜ ์ผ๋ถ€๋Š” ๋‹จ์ˆœํžˆ ์ด๋ฏธ์ง€ ํฌ๊ธฐ๊ฐ€ ๋” ์ปค์กŒ๊ธฐ ๋•Œ๋ฌธ์ด๊ณ (832 ๋Œ€ 640), ์ผ๋ถ€๋Š” ์‹ค์ œ TTA ์ž‘์—… ๋•Œ๋ฌธ์ด๊ธฐ๋„ ํ•ฉ๋‹ˆ๋‹ค.

python val.py --weights yolov5x.pt --data coco.yaml --img 832 --augment --half

์ถœ๋ ฅ:

val: data=./data/coco.yaml, weights=['yolov5x.pt'], batch_size=32, imgsz=832, conf_thres=0.001, iou_thres=0.6, task=val, device=, single_cls=False, augment=True, verbose=False, save_txt=False, save_hybrid=False, save_conf=False, save_json=True, project=runs/val, name=exp, exist_ok=False, half=True
YOLOv5 ๐Ÿš€ v5.0-267-g6a3ee7c torch 1.9.0+cu102 CUDA:0 (Tesla P100-PCIE-16GB, 16280.875MB)

Fusing layers...
/usr/local/lib/python3.7/dist-packages/torch/nn/functional.py:718: UserWarning: Named tensors and all their associated APIs are an experimental feature and subject to change. Please do not use them for anything important until they are released as stable. (Triggered internally at  /pytorch/c10/core/TensorImpl.h:1156.)
  return torch.max_pool2d(input, kernel_size, stride, padding, dilation, ceil_mode)
Model Summary: 476 layers, 87730285 parameters, 0 gradients
val: Scanning '../datasets/coco/val2017' images and labels...4952 found, 48 missing, 0 empty, 0 corrupted: 100% 5000/5000 [00:01<00:00, 2885.61it/s]
val: New cache created: ../datasets/coco/val2017.cache
               Class     Images     Labels          P          R     mAP@.5 mAP@.5:.95: 100% 157/157 [07:29<00:00,  2.86s/it]
                 all       5000      36335      0.718      0.656      0.695      0.503
Speed: 0.2ms pre-process, 80.6ms inference, 2.7ms NMS per image at shape (32, 3, 832, 832)  # <--- TTA speed

Evaluating pycocotools mAP... saving runs/val/exp2/yolov5x_predictions.json...
...
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.516  # <--- TTA mAP
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.701
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.562
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.361
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.564
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.656
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.388
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.640
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.696  # <--- TTA mAR
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.553
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.744
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.833

TTA๋ฅผ ํ†ตํ•œ ์ถ”๋ก 

detect.py TTA ์ถ”๋ก ์€ ๋‹ค์Œ๊ณผ ๋™์ผํ•˜๊ฒŒ ์ž‘๋™ํ•ฉ๋‹ˆ๋‹ค. val.py TTA: ๊ฐ„๋‹จํžˆ ์ถ”๊ฐ€ --augment ๋ฅผ ๊ธฐ์กด detect.py ๋ช…๋ น์ž…๋‹ˆ๋‹ค:

python detect.py --weights yolov5s.pt --img 832 --source data/images --augment

์ถœ๋ ฅ:

YOLOv5 ๐Ÿš€ v5.0-267-g6a3ee7c torch 1.9.0+cu102 CUDA:0 (Tesla P100-PCIE-16GB, 16280.875MB)

Downloading https://github.com/ultralytics/yolov5/releases/download/v5.0/yolov5s.pt to yolov5s.pt...
100% 14.1M/14.1M [00:00<00:00, 81.9MB/s]

Fusing layers...
Model Summary: 224 layers, 7266973 parameters, 0 gradients
image 1/2 /content/yolov5/data/images/bus.jpg: 832x640 4 persons, 1 bus, 1 fire hydrant, Done. (0.029s)
image 2/2 /content/yolov5/data/images/zidane.jpg: 480x832 3 persons, 3 ties, Done. (0.024s)
Results saved to runs/detect/exp
Done. (0.156s)

YOLOv5 ํ…Œ์ŠคํŠธ ์‹œ๊ฐ„ ํ™•๋Œ€

PyTorch Hub TTA

TTA๋Š” ๋ชจ๋“  YOLOv5 PyTorch Hub ๋ชจ๋ธ์— ์•ก์„ธ์Šคํ•  ์ˆ˜ ์žˆ์œผ๋ฉฐ, ๋‹ค์Œ๊ณผ ๊ฐ™์ด augment=True ๋กœ ์„ค์ •ํ•ฉ๋‹ˆ๋‹ค.

import torch

# Model
model = torch.hub.load('ultralytics/yolov5', 'yolov5s')  # or yolov5m, yolov5x, custom

# Images
img = 'https://ultralytics.com/images/zidane.jpg'  # or file, PIL, OpenCV, numpy, multiple

# Inference
results = model(img, augment=True)  # <--- TTA inference

# Results
results.print()  # or .show(), .save(), .crop(), .pandas(), etc.

์‚ฌ์šฉ์ž ์ง€์ •

์— ์ ์šฉ๋œ TTA ์ž‘์—…์„ ์‚ฌ์šฉ์ž ์ง€์ •ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. YOLOv5 forward_augment() ๋ฉ”์„œ๋“œ ์—ฌ๊ธฐ.

์ง€์› ํ™˜๊ฒฝ

Ultralytics ๋Š” ๋ฐ”๋กœ ์‚ฌ์šฉํ•  ์ˆ˜ ์žˆ๋Š” ๋‹ค์–‘ํ•œ ํ™˜๊ฒฝ์„ ์ œ๊ณตํ•˜๋ฉฐ, ๊ฐ ํ™˜๊ฒฝ์—๋Š” CUDA, CUDNN๊ณผ ๊ฐ™์€ ํ•„์ˆ˜ ์ข…์† ์š”์†Œ๊ฐ€ ์‚ฌ์ „ ์„ค์น˜๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค, Python, ๋ฐ PyTorch์™€ ๊ฐ™์€ ํ•„์ˆ˜ ์ข…์†์„ฑ์ด ์‚ฌ์ „ ์„ค์น˜๋˜์–ด ์žˆ์–ด ํ”„๋กœ์ ํŠธ๋ฅผ ์‹œ์ž‘ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

ํ”„๋กœ์ ํŠธ ์ƒํƒœ

YOLOv5 CI

์ด ๋ฐฐ์ง€๋Š” ๋ชจ๋“  YOLOv5 GitHub Actions ์ง€์†์  ํ†ตํ•ฉ(CI) ํ…Œ์ŠคํŠธ๊ฐ€ ์„ฑ๊ณต์ ์œผ๋กœ ํ†ต๊ณผ๋˜์—ˆ์Œ์„ ๋‚˜ํƒ€๋ƒ…๋‹ˆ๋‹ค. ์ด๋Ÿฌํ•œ CI ํ…Œ์ŠคํŠธ๋Š” ๊ต์œก, ๊ฒ€์ฆ, ์ถ”๋ก , ๋‚ด๋ณด๋‚ด๊ธฐ, ๋ฒค์น˜๋งˆํฌ ๋“ฑ ๋‹ค์–‘ํ•œ ์ฃผ์š” ์ธก๋ฉด์— ๊ฑธ์ณ YOLOv5 ์˜ ๊ธฐ๋Šฅ๊ณผ ์„ฑ๋Šฅ์„ ์—„๊ฒฉํ•˜๊ฒŒ ํ™•์ธํ•ฉ๋‹ˆ๋‹ค. 24์‹œ๊ฐ„๋งˆ๋‹ค, ๊ทธ๋ฆฌ๊ณ  ์ƒˆ๋กœ์šด ์ปค๋ฐ‹์ด ์žˆ์„ ๋•Œ๋งˆ๋‹ค ํ…Œ์ŠคํŠธ๋ฅผ ์ˆ˜ํ–‰ํ•˜์—ฌ macOS, Windows ๋ฐ Ubuntu์—์„œ ์ผ๊ด€๋˜๊ณ  ์•ˆ์ •์ ์ธ ์ž‘๋™์„ ๋ณด์žฅํ•ฉ๋‹ˆ๋‹ค.



์ƒ์„ฑ๋จ 2023-11-12, ์—…๋ฐ์ดํŠธ๋จ 2023-12-03
์ž‘์„ฑ์ž: glenn-jocher (4)

๋Œ“๊ธ€