yolov5 custom dataset view anchors

1. View anchors

The yolov5 model will perform autoanchor check before training its own data set. The autoanchor check will automatically perform key means clustering on the ground truth boxes in the data set, and generate new anchors to improve the mAP of model detection. You can view the new generation with the following code The anchors:

import torch
from models.experimental import attempt_load


model = attempt_load('./weights/yolov5s.pt', map_location=torch.device('cpu'))
m = model.module.model[-1] if hasattr(model, 'module') else model.model[-1]
print(m.anchor_grid)

2. Use default anchors

If you do not want to perform key means clustering of anchors, you can add it during training --noautoanchorand use the default anchors. The default anchors are as follows:

# anchors
anchors:
  - [10,13, 16,30, 33,23]  # P3/8
  - [30,61, 62,45, 59,119]  # P4/16
  - [116,90, 156,198, 373,326]  # P5/32