match new eval function

ppcls/configs/Attr/StrongBaselineAttr.yaml

@@ -5,7 +5,7 @@ Global:
   output_dir: "./output/"
   device: "gpu"
   save_interval: 5
-  eval_during_train: False
+  eval_during_train: True
   eval_interval: 1
   epochs: 30
   print_batch_step: 20

ppcls/engine/evaluation/classification.py

@@ -18,7 +18,7 @@ import time
 import platform
 import paddle
 
-from ppcls.utils.misc import AverageMeter, AttrMeter
+from ppcls.utils.misc import AverageMeter
 from ppcls.utils import logger
 
 
@@ -34,10 +34,6 @@ def classification_eval(engine, epoch_id=0):
     }
     print_batch_step = engine.config["Global"]["print_batch_step"]
 
-    if engine.eval_metric_func is not None and "ATTRMetric" in engine.config[
-            "Metric"]["Eval"][0]:
-        output_info["attr"] = AttrMeter(threshold=0.5)
-
     metric_key = None
     tic = time.time()
     accum_samples = 0
@@ -162,7 +158,7 @@ def classification_eval(engine, epoch_id=0):
     if "ATTRMetric" in engine.config["Metric"]["Eval"][0]:
         metric_msg = ", ".join([
             "evalres: ma: {:.5f} label_f1: {:.5f} label_pos_recall: {:.5f} label_neg_recall: {:.5f} instance_f1: {:.5f} instance_acc: {:.5f} instance_prec: {:.5f} instance_recall: {:.5f}".
-            format(*output_info["attr"].res())
+            format(*engine.eval_metric_func.attr_res())
         ])
         logger.info("[Eval][Epoch {}][Avg]{}".format(epoch_id, metric_msg))
 
@@ -170,7 +166,7 @@ def classification_eval(engine, epoch_id=0):
         if engine.eval_metric_func is None:
             return -1
         # return 1st metric in the dict
-        return output_info["attr"].res()[0]
+        return engine.eval_metric_func.attr_res()[0]
     else:
         metric_msg = ", ".join([
             "{}: {:.5f}".format(key, output_info[key].avg)

ppcls/metric/__init__.py

@@ -56,6 +56,9 @@ class CombinedMetrics(AvgMetrics):
     def avg(self):
         return self.metric_func_list[0].avg
 
+    def attr_res(self):
+        return self.metric_func_list[0].attrmeter.res()
+
     def reset(self):
         for metric in self.metric_func_list:
             if hasattr(metric, "reset"):

ppcls/metric/metrics.py

@@ -25,7 +25,7 @@ from sklearn.preprocessing import binarize
 from easydict import EasyDict
 
 from ppcls.metric.avg_metrics import AvgMetrics
-from ppcls.utils.misc import AverageMeter
+from ppcls.utils.misc import AverageMeter, AttrMeter
 
 
 class TopkAcc(AvgMetrics):
@@ -438,7 +438,11 @@ class ATTRMetric(nn.Layer):
         super().__init__()
         self.threshold = threshold
 
+    def reset(self):
+        self.attrmeter = AttrMeter(threshold=0.5)
+
     def forward(self, output, target):
         metric_dict = get_attr_metrics(target[:, 0, :].numpy(),
                                        output.numpy(), self.threshold)
+        self.attrmeter.update(metric_dict)
         return metric_dict