更换文档检测模型

paddle_detection/ppdet/modeling/architectures/retinanet.py

@@ -0,0 +1,84 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+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
+import paddle
+import paddle.nn.functional as F
+
+__all__ = ['RetinaNet']
+
+
+@register
+class RetinaNet(BaseArch):
+    __category__ = 'architecture'
+
+    def __init__(self, backbone, neck, head):
+        super(RetinaNet, self).__init__()
+        self.backbone = backbone
+        self.neck = neck
+        self.head = head
+
+    @classmethod
+    def from_config(cls, cfg, *args, **kwargs):
+        backbone = create(cfg['backbone'])
+
+        kwargs = {'input_shape': backbone.out_shape}
+        neck = create(cfg['neck'], **kwargs)
+
+        kwargs = {'input_shape': neck.out_shape}
+        head = create(cfg['head'], **kwargs)
+
+        return {
+            'backbone': backbone,
+            'neck': neck,
+            'head': head,
+        }
+
+    def _forward(self):
+        body_feats = self.backbone(self.inputs)
+        neck_feats = self.neck(body_feats)
+
+        if self.training:
+            return self.head(neck_feats, self.inputs)
+        else:
+            head_outs = self.head(neck_feats)
+            bbox, bbox_num, nms_keep_idx = self.head.post_process(
+                head_outs, self.inputs['im_shape'], self.inputs['scale_factor'])
+
+            if self.use_extra_data:
+                extra_data = {}  # record the bbox output before nms, such like scores and nms_keep_idx
+                """extra_data:{
+                            'scores': predict scores,
+                            'nms_keep_idx': bbox index before nms,
+                           }
+                           """
+                preds_logits = self.head.decode_cls_logits(head_outs[0])
+                preds_scores = F.sigmoid(preds_logits)
+                extra_data['logits'] = preds_logits
+                extra_data['scores'] = preds_scores
+                extra_data['nms_keep_idx'] = nms_keep_idx  # bbox index before nms
+                return {'bbox': bbox, 'bbox_num': bbox_num, "extra_data": extra_data}
+            else:
+                return {'bbox': bbox, 'bbox_num': bbox_num}
+
+    def get_loss(self):
+        return self._forward()
+
+    def get_pred(self):
+        return self._forward()