yolo.py

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

from ppdet.core.workspace import register, create
from .meta_arch import BaseArch

__all__ = ['YOLOv3']


@register
class YOLOv3(BaseArch):
    __category__ = 'architecture'
    __shared__ = ['data_format']
    __inject__ = ['post_process']

    def __init__(self,
                 backbone='DarkNet',
                 neck='YOLOv3FPN',
                 yolo_head='YOLOv3Head',
                 post_process='BBoxPostProcess',
                 data_format='NCHW'):
        """
        YOLOv3 network, see https://arxiv.org/abs/1804.02767

        Args:
            backbone (nn.Layer): backbone instance
            neck (nn.Layer): neck instance
            yolo_head (nn.Layer): anchor_head instance
            bbox_post_process (object): `BBoxPostProcess` instance
            data_format (str): data format, NCHW or NHWC
        """
        super(YOLOv3, self).__init__(data_format=data_format)
        self.backbone = backbone
        self.neck = neck
        self.yolo_head = yolo_head
        self.post_process = post_process

    @classmethod
    def from_config(cls, cfg, *args, **kwargs):
        # backbone
        backbone = create(cfg['backbone'])

        # fpn
        kwargs = {'input_shape': backbone.out_shape}
        neck = create(cfg['neck'], **kwargs)

        # head
        kwargs = {'input_shape': neck.out_shape}
        yolo_head = create(cfg['yolo_head'], **kwargs)

        return {
            'backbone': backbone,
            'neck': neck,
            "yolo_head": yolo_head,
        }

    def _forward(self):
        body_feats = self.backbone(self.inputs)
        body_feats = self.neck(body_feats)

        if self.training:
            return self.yolo_head(body_feats, self.inputs)
        else:
            yolo_head_outs = self.yolo_head(body_feats)
            bbox, bbox_num = self.post_process(
                yolo_head_outs, self.yolo_head.mask_anchors,
                self.inputs['im_shape'], self.inputs['scale_factor'])
            return bbox, bbox_num

    def get_loss(self):
        return self._forward()

    def get_pred(self):
        bbox_pred, bbox_num = self._forward()
        output = {'bbox': bbox_pred, 'bbox_num': bbox_num}
        return output