Update bert_layers.py

bert_layers.py

@@ -939,19 +939,19 @@ class BertForSequenceClassification(BertPreTrainedModel):
         all_attention_probs = outputs[3]
 
         try:
-            print(f'outputs[2] AFTER reassignment SHAPE: {outputs[3][0].shape} ')
+            print(f'outputs[2] AFTER reassignment probsss SHAPE: {outputs[3][0].shape} ')
         except:
-            print(print(f'outputs[2] AFTER reassignment LENGTH: {len(outputs[3][0])} '))
+            print(print(f'outputs[2] AFTER reassignment probsss LENGTH: {len(outputs[3][0])} '))
 
         
         
         try:
-            print(f'all_attention_weights last: {all_attention_probs.shape}')
+            print(f'all_attention_weights probsss last: {all_attention_probs.shape}')
         except:
             try:
-                print(f'last first except: {all_attention_probs[0].shape}')
+                print(f'last first except probsss: {all_attention_probs[0].shape}')
             except:
-                print(f'last second except: {len(all_attention_probs[0])}')
+                print(f'last second except probsss: {len(all_attention_probs[0])}')
 
 
         pooled_output = self.dropout(pooled_output)
@@ -991,9 +991,9 @@ class BertForSequenceClassification(BertPreTrainedModel):
             
         # print(outputs.attentions)
         try:
-            print(f'not stacked final attention SHAPE: {outputs[3][0].shape}')
+            print(f'not stacked final attention probsss SHAPE: {outputs[3][0].shape}')
         except:
-            print(f'not stacked final attention LEN: {len(outputs[3])}')
+            print(f'not stacked final attention probsss LEN: {len(outputs[3])}')
             
         # try:
         #     print(f'STACKED final attention SHAPE: {(outputs.attentions).shape}')