Merge https://github.com/PaddlePaddle/PaddleOCR into dygraph

ppocr/losses/det_db_loss.py

@@ -47,11 +47,12 @@ class DBLoss(nn.Layer):
             negative_ratio=ohem_ratio)
 
     def forward(self, predicts, labels):
+        predict_maps = predicts['maps']
         label_threshold_map, label_threshold_mask, label_shrink_map, label_shrink_mask = labels[
             1:]
-        shrink_maps = predicts[:, 0, :, :]
-        threshold_maps = predicts[:, 1, :, :]
-        binary_maps = predicts[:, 2, :, :]
+        shrink_maps = predict_maps[:, 0, :, :]
+        threshold_maps = predict_maps[:, 1, :, :]
+        binary_maps = predict_maps[:, 2, :, :]
 
         loss_shrink_maps = self.bce_loss(shrink_maps, label_shrink_map,
                                          label_shrink_mask)

ppocr/modeling/heads/det_db_head.py

@@ -120,9 +120,9 @@ class DBHead(nn.Layer):
     def forward(self, x):
         shrink_maps = self.binarize(x)
         if not self.training:
-            return shrink_maps
+            return {'maps': shrink_maps}
 
         threshold_maps = self.thresh(x)
         binary_maps = self.step_function(shrink_maps, threshold_maps)
         y = paddle.concat([shrink_maps, threshold_maps, binary_maps], axis=1)
-        return y
+        return {'maps': y}

ppocr/postprocess/db_postprocess.py

@@ -40,7 +40,8 @@ class DBPostProcess(object):
         self.max_candidates = max_candidates
         self.unclip_ratio = unclip_ratio
         self.min_size = 3
-        self.dilation_kernel = None if not use_dilation else np.array([[1, 1], [1, 1]])
+        self.dilation_kernel = None if not use_dilation else np.array(
+            [[1, 1], [1, 1]])
 
     def boxes_from_bitmap(self, pred, _bitmap, dest_width, dest_height):
         '''
@@ -132,7 +133,8 @@ class DBPostProcess(object):
         cv2.fillPoly(mask, box.reshape(1, -1, 2).astype(np.int32), 1)
         return cv2.mean(bitmap[ymin:ymax + 1, xmin:xmax + 1], mask)[0]
 
-    def __call__(self, pred, shape_list):
+    def __call__(self, outs_dict, shape_list):
+        pred = outs_dict['maps']
         if isinstance(pred, paddle.Tensor):
             pred = pred.numpy()
         pred = pred[:, 0, :, :]

ppocr/utils/save_load.py

@@ -102,7 +102,6 @@ def init_model(config, model, logger, optimizer=None, lr_scheduler=None):
             best_model_dict = states_dict.get('best_model_dict', {})
             if 'epoch' in states_dict:
                 best_model_dict['start_epoch'] = states_dict['epoch'] + 1
-            best_model_dict['start_epoch'] = best_model_dict['best_epoch'] + 1
 
         logger.info("resume from {}".format(checkpoints))
     elif pretrained_model:

tools/infer/predict_det.py

@@ -65,12 +65,12 @@ class TextDetector(object):
             postprocess_params["unclip_ratio"] = args.det_db_unclip_ratio
             postprocess_params["use_dilation"] = True
         elif self.det_algorithm == "EAST":
-            postprocess_params['name'] = 'EASTPostProcess'      
+            postprocess_params['name'] = 'EASTPostProcess'
             postprocess_params["score_thresh"] = args.det_east_score_thresh
             postprocess_params["cover_thresh"] = args.det_east_cover_thresh
             postprocess_params["nms_thresh"] = args.det_east_nms_thresh
         elif self.det_algorithm == "SAST":
-            postprocess_params['name'] = 'SASTPostProcess'      
+            postprocess_params['name'] = 'SASTPostProcess'
             postprocess_params["score_thresh"] = args.det_sast_score_thresh
             postprocess_params["nms_thresh"] = args.det_sast_nms_thresh
             self.det_sast_polygon = args.det_sast_polygon
@@ -177,8 +177,10 @@ class TextDetector(object):
             preds['f_score'] = outputs[1]
             preds['f_tco'] = outputs[2]
             preds['f_tvo'] = outputs[3]
+        elif self.det_algorithm == 'DB':
+            preds['maps'] = outputs[0]
         else:
-            preds = outputs[0]
+            raise NotImplementedError
 
         post_result = self.postprocess_op(preds, shape_list)
         dt_boxes = post_result[0]['points']