Fix bug

tokenization_chatglm.py

@@ -176,6 +176,8 @@ class ChatGLMTokenizer(PreTrainedTokenizer):
             mask_token='[MASK]',
             gmask_token='[gMASK]',
             padding_side="left",
+            pad_token="<pad>",
+            unk_token="<unk>",
             num_image_tokens=20000,
             **kwargs
     ) -> None:
@@ -188,6 +190,8 @@ class ChatGLMTokenizer(PreTrainedTokenizer):
             end_token=end_token,
             mask_token=mask_token,
             gmask_token=gmask_token,
+            pad_token=pad_token,
+            unk_token=unk_token,
             num_image_tokens=num_image_tokens,
             **kwargs
         )
@@ -402,6 +406,10 @@ class ChatGLMTokenizer(PreTrainedTokenizer):
                 encoded_inputs["attention_mask"] = attention_mask
 
             if "position_ids" not in encoded_inputs:
+                if bos_token_id in required_input:
+                    context_length = required_input.index(bos_token_id)
+                else:
+                    context_length = seq_length
                 position_ids = np.arange(seq_length, dtype=np.int64)
                 mask_token = mask_token_id if mask_token_id in required_input else gmask_token_id
                 if mask_token in required_input: