update readme

basic_demo/README.md

@@ -100,7 +100,7 @@ python trans_web_demo.py
 + 使用 Batch 推理。
 
 ```shell
-python cli_batch_request_demo.py
+python trans_batch_demo.py
 ```
 
 ### 使用 vLLM 后端代码

basic_demo/README_en.md

@@ -105,7 +105,7 @@ python trans_web_demo.py
 + Use Batch inference.
 
 ```shell
-python cli_batch_request_demo.py
+python trans_batch_demo.py
 ```
 
 ### Use vLLM backend code

finetune_demo/finetune.py

@@ -260,7 +260,6 @@ def process_batch(
         loss_masks = [False, False]
         for message in conv:
             message = process_message(message)
-
             loss_mask_val = False if message['role'] in ('system', 'user', 'observation') else True
             new_input_ids = tokenizer.apply_chat_template([message], tokenize=True, return_dict=False)[2:]
             new_loss_masks = [loss_mask_val] * len(new_input_ids)