Update app.py

app.py

@@ -188,7 +188,31 @@ def prepare_features_for_qa_inference(examples, tokenizer, pad_on_right, max_seq
                     raise ValueError(f"在 prepare_features_for_qa_inference 中，{key_to_check} 的第 {i} 個特徵列表為 None！")
                 if any(x is None for x in lst):
                     raise ValueError(f"在 prepare_features_for_qa_inference 中，{key_to_check} 的第 {i} 個特徵列表內部包含 None！內容: {lst[:20]}")
-                    
+
+    for key_to_check in ["input_ids", "attention_mask", "token_type_ids"]:
+        if key_to_check in final_batch:
+            new_list_of_lists = []
+            for single_feature_list in final_batch[key_to_check]:
+                if single_feature_list is None: # 如果整個特徵的這個字段是 None
+                    # logger.error(f"Critical error: {key_to_check} list for a feature is None. Reconstructing a default.")
+                    # 根據 key_to_check 類型創建一個安全的默認值
+                    if key_to_check == "input_ids":
+                        safe_list = [tokenizer.cls_token_id or 101, tokenizer.sep_token_id or 102] + \
+                                    [tokenizer.pad_token_id or 0] * (max_seq_len - 2)
+                        new_list_of_lists.append(safe_list[:max_seq_len])
+                    elif key_to_check == "attention_mask":
+                        safe_list = [1,1] + [0] * (max_seq_len-2)
+                        new_list_of_lists.append(safe_list[:max_seq_len])
+                    elif key_to_check == "token_type_ids":
+                        new_list_of_lists.append([0] * max_seq_len)
+                elif not all(isinstance(x, int) for x in single_feature_list): # 如果列表內包含非整數
+                    # logger.error(f"Critical error: {key_to_check} list for a feature contains non-integers: {single_feature_list[:10]}. Fixing.")
+                    default_val = tokenizer.pad_token_id if key_to_check == "input_ids" else 0
+                    new_list_of_lists.append([default_val if not isinstance(x, int) else x for x in single_feature_list])
+                else:
+                    new_list_of_lists.append(single_feature_list) # 原本就是好的
+            final_batch[key_to_check] = new_list_of_lists
+            
     return final_batch
 
 # postprocess_qa_predictions 函數也需要從 utils_qa.py 複製或導入