Update tokenization_chatglm.py (#3)

- Update tokenization_chatglm.py (5c8ffabbfc8ba950a8f450939138bc5254b6ee1f)
- Update config.json (73a1403930e64d11e421f8eb7b1d19eb49cfeeeb)


Co-authored-by: Niklas Muennighoff <[email protected]>

config.json

@@ -7,6 +7,7 @@
   "auto_map": {
     "AutoConfig": "configuration_chatglm.ChatGLMConfig",
     "AutoModel": "modeling_chatglm.ChatGLMForConditionalGeneration",
+    "AutoModelForCausalLM": "modeling_chatglm.ChatGLMForConditionalGeneration",
     "AutoModelForSeq2SeqLM": "modeling_chatglm.ChatGLMForConditionalGeneration"
   },
   "add_bias_linear": false,

tokenization_chatglm.py

@@ -225,7 +225,7 @@ class ChatGLMTokenizer(PreTrainedTokenizer):
                 (optional) Set to False to avoid returning attention mask (default: set to model specifics)
         """
         # Load from model defaults
-        assert self.padding_side == "left"
+        # assert self.padding_side == "left"
 
         required_input = encoded_inputs[self.model_input_names[0]]
         seq_length = len(required_input)
@@ -248,10 +248,17 @@ class ChatGLMTokenizer(PreTrainedTokenizer):
         if needs_to_be_padded:
             difference = max_length - len(required_input)
 
-            if "attention_mask" in encoded_inputs:
-                encoded_inputs["attention_mask"] = [0] * difference + encoded_inputs["attention_mask"]
-            if "position_ids" in encoded_inputs:
-                encoded_inputs["position_ids"] = [0] * difference + encoded_inputs["position_ids"]
-            encoded_inputs[self.model_input_names[0]] = [self.pad_token_id] * difference + required_input
+            if self.padding_side == "left":
+                if "attention_mask" in encoded_inputs:
+                    encoded_inputs["attention_mask"] = [0] * difference + encoded_inputs["attention_mask"]
+                if "position_ids" in encoded_inputs:
+                    encoded_inputs["position_ids"] = [0] * difference + encoded_inputs["position_ids"]
+                encoded_inputs[self.model_input_names[0]] = [self.pad_token_id] * difference + required_input
+            else:
+                if "attention_mask" in encoded_inputs:
+                    encoded_inputs["attention_mask"] = encoded_inputs["attention_mask"] + [0] * difference
+                if "position_ids" in encoded_inputs:
+                    encoded_inputs["position_ids"] = encoded_inputs["position_ids"] + [0] * difference
+                encoded_inputs[self.model_input_names[0]] = required_input + [self.pad_token_id] * difference
 
         return encoded_inputs