to be compatible with training and evaluation

ppcls/metric/metrics.py

@@ -38,6 +38,7 @@ class TopkAcc(AvgMetrics):
             topk = [topk]
         self.topk = topk
         self.reset()
+        self.warned = False
 
     def reset(self):
         self.avg_meters = {
@@ -54,17 +55,15 @@ class TopkAcc(AvgMetrics):
         metric_dict = dict()
         for idx, k in enumerate(self.topk):
             if output_dims < k:
+                if not self.warned:
                     msg = f"The output dims({output_dims}) is less than k({k}), so the Top-{k} metric is meaningless."
                     logger.warning(msg)
+                    self.warned = True
                 metric_dict[f"top{k}"] = 1
-                self.avg_meters.pop(f"top{k}")
-                continue
+            else:
                 metric_dict[f"top{k}"] = paddle.metric.accuracy(x, label, k=k)
             self.avg_meters[f"top{k}"].update(metric_dict[f"top{k}"],
                                               x.shape[0])
-
-        self.topk = list(filter(lambda k: k <= output_dims, self.topk))
-
         return metric_dict