Bladeren bron

fix large_dataset epoch shuffle bug

haoneng.lhn 2 jaren geleden
bovenliggende
commit
9a7ea0f10e

+ 1 - 1
funasr/datasets/large_datasets/datapipes/batch.py

@@ -39,7 +39,7 @@ class MaxTokenBucketizerIterDataPipe(IterableDataset):
         self.batch_mode = batch_mode
 
     def set_epoch(self, epoch):
-        self.epoch = epoch
+        self.datapipe.set_epoch(epoch)
 
     def __iter__(self):
         buffer = []

+ 2 - 2
funasr/datasets/large_datasets/datapipes/filter.py

@@ -13,7 +13,7 @@ class FilterIterDataPipe(IterableDataset):
         self.fn = fn
 
     def set_epoch(self, epoch):
-        self.epoch = epoch
+        self.datapipe.set_epoch(epoch)
 
     def __iter__(self):
         assert callable(self.fn)
@@ -21,4 +21,4 @@ class FilterIterDataPipe(IterableDataset):
             if self.fn(data):
                 yield data
             else:
-                continue
+                continue

+ 1 - 1
funasr/datasets/large_datasets/datapipes/map.py

@@ -14,7 +14,7 @@ class MapperIterDataPipe(IterableDataset):
         self.fn = fn
 
     def set_epoch(self, epoch):
-        self.epoch = epoch
+        self.datapipe.set_epoch(epoch)
 
     def __iter__(self):
         assert callable(self.fn)