From 6a8eefc35e7a8f072730302a90747a6be6622b5e Mon Sep 17 00:00:00 2001
From: luukunn <981429396@qq.com>
Date: Fri, 15 Aug 2025 14:31:10 +0800
Subject: [PATCH] fix args

---
 fastdeploy/input/ernie_processor.py    | 7 ++++++-
 fastdeploy/input/ernie_vl_processor.py | 6 +++++-
 fastdeploy/input/text_processor.py     | 6 +++++-
 3 files changed, 16 insertions(+), 3 deletions(-)

diff --git a/fastdeploy/input/ernie_processor.py b/fastdeploy/input/ernie_processor.py
index 4ff4d5e14..766feea47 100644
--- a/fastdeploy/input/ernie_processor.py
+++ b/fastdeploy/input/ernie_processor.py
@@ -90,6 +90,7 @@ class ErnieProcessor(BaseDataProcessor):
         request = self._apply_default_parameters(request)
         if request.get("eos_token_ids") is None or len(request.eos_token_ids) == 0:
             request.eos_token_ids = self.eos_token_ids
+
         request.enable_thinking = kwargs.get("chat_template_kwargs", {}).get("enable_thinking")
         stop_sequences = request.get("stop", [])
         if stop_sequences is not None and len(stop_sequences) != 0:
@@ -141,7 +142,11 @@ class ErnieProcessor(BaseDataProcessor):
         request = self._apply_default_parameters(request)
         if not request.get("eos_token_ids"):
             request["eos_token_ids"] = self.eos_token_ids
-        request["enable_thinking"] = request.get("chat_template_kwargs", {}).get("enable_thinking")
+        chat_template_kwargs = request.get("chat_template_kwargs")
+        if chat_template_kwargs:
+            for k, v in chat_template_kwargs.items():
+                if k not in request:
+                    request[k] = v
         # processing stop_sequences
         stop_sequences = request.get("stop", [])
         if stop_sequences:
diff --git a/fastdeploy/input/ernie_vl_processor.py b/fastdeploy/input/ernie_vl_processor.py
index 756011553..04f29fa19 100644
--- a/fastdeploy/input/ernie_vl_processor.py
+++ b/fastdeploy/input/ernie_vl_processor.py
@@ -198,7 +198,11 @@ class ErnieMoEVLProcessor(ErnieProcessor):
         request = self._apply_default_parameters(request)
         if not request.get("eos_token_ids"):
             request["eos_token_ids"] = self.eos_token_ids
-        request["enable_thinking"] = request.get("chat_template_kwargs", {}).get("enable_thinking")
+        chat_template_kwargs = request.get("chat_template_kwargs")
+        if chat_template_kwargs:
+            for k, v in chat_template_kwargs.items():
+                if k not in request:
+                    request[k] = v
 
         stop_sequences = request.get("stop", [])
         if stop_sequences:
diff --git a/fastdeploy/input/text_processor.py b/fastdeploy/input/text_processor.py
index 10f5081b4..f67c922ef 100644
--- a/fastdeploy/input/text_processor.py
+++ b/fastdeploy/input/text_processor.py
@@ -207,7 +207,11 @@ class DataProcessor(BaseDataProcessor):
         request = self._apply_default_parameters(request)
         if request.get("eos_token_ids") is None or len(request.eos_token_ids) == 0:
             request.eos_token_ids = self.eos_token_ids
-        request.enable_thinking = kwargs.get("chat_template_kwargs", {}).get("enable_thinking")
+        chat_template_kwargs = request.get("chat_template_kwargs")
+        if chat_template_kwargs:
+            for k, v in chat_template_kwargs.items():
+                if k not in request:
+                    request[k] = v
         stop_sequences = request.get("stop", [])
         if stop_sequences is not None and len(stop_sequences) != 0:
             stop_seqs, stop_seqs_len = self.update_stop_seq(stop_sequences)