diff --git a/mmedit/apis/train.py b/mmedit/apis/train.py
index cffb26c4ba..c955b7aa6f 100644
--- a/mmedit/apis/train.py
+++ b/mmedit/apis/train.py
@@ -182,11 +182,6 @@ def _non_dist_train(model,
         meta (dict | None): Meta dict to record some important information.
             Default: None.
     """
-    if validate:
-        raise NotImplementedError('Built-in validation is not implemented '
-                                  'yet in not-distributed training. Use '
-                                  'distributed training or test.py and '
-                                  '*eval.py scripts instead.')
     # prepare data loaders
     dataset = dataset if isinstance(dataset, (list, tuple)) else [dataset]
     data_loaders = [
diff --git a/mmedit/core/evaluation/eval_hooks.py b/mmedit/core/evaluation/eval_hooks.py
index e8ff8c0acc..95dcb93f5b 100644
--- a/mmedit/core/evaluation/eval_hooks.py
+++ b/mmedit/core/evaluation/eval_hooks.py
@@ -36,6 +36,7 @@ def after_train_iter(self, runner):
         """
         if not self.every_n_iters(runner, self.interval):
             return
+        runner.log_buffer.clear()
         from mmedit.apis import single_gpu_test
         results = single_gpu_test(
             runner.model,