diff --git a/src/model_meta/app.py b/src/model_meta/app.py
new file mode 100755
index 0000000..33ce16b
--- /dev/null
+++ b/src/model_meta/app.py
@@ -0,0 +1,57 @@
+import os
+import json
+
+from flask import Flask, request
+
+
+model_task = os.getenv('TASK', '')
+app = Flask(__name__)
+
+def get_task_input_examples(task):
+    current_work_dir = os.path.dirname(__file__)
+    with open(current_work_dir + '/data/data/pipeline_inputs.json', 'r') as f:
+        input_examples = json.load(f)
+    if task in input_examples:
+        return input_examples[task]
+    return None
+
+
+@app.route('/schemas', methods=['GET'])
+def get_schemas():
+    request_id = request.headers.get("x-fc-request-id", "")
+    schemas_task = request.args.get("task", model_task)
+
+    if not schemas_task:
+        err_ret = {
+            'Code': 400,
+            'Message': "failed",
+            'Data': "param task is null",
+            "RequestId": request_id,
+            "Success": False
+        }
+        return err_ret, 400, [("x-fc-status", "400")]
+
+    print("[INFO] get schemas by task: " + schemas_task)
+    result = get_task_input_examples(schemas_task)
+    if not result:
+        err_ret = {
+            'Code': 400,
+            'Message': "failed",
+            'Data': "param task is invalid.",
+            "RequestId": request_id,
+            "Success": False
+        }
+        return err_ret, 400, [("x-fc-status", "400")]
+
+    return {
+               'Code': 200,
+               'Message': "",
+               'Data': result,
+               "RequestId": request_id,
+               "Success": True
+           }, 200, [("Content-Type", "application/json")]
+
+
+if __name__ == '__main__':
+    os.system('mkdir data; cd data; wget https://modelscope.oss-cn-beijing.aliyuncs.com/swingdeploy/deploy.tar; tar xvf deploy.tar')
+    app.run(debug=False, host='0.0.0.0', port=9000)
diff --git a/src/s.yaml b/src/s.yaml
index 6b4caf4..0401493 100644
--- a/src/s.yaml
+++ b/src/s.yaml
@@ -82,17 +82,23 @@ resources:
       memorySize: 32768
       cpu: 8
       diskSize: 512
-      gpuConfig:
-        gpuMemorySize: 16384
-        gpuType: fc.gpu.tesla.1
       instanceConcurrency: 1
-      runtime: custom-container
+      runtime: python3.9
+      code: ./model_meta
       internetAccess: true
-      customContainerConfig:
-        image: registry.${vars.region}.aliyuncs.com/modelscope-repo/modelscope:${vars.version}
-        port: 9000
+      initializer: 'true'
       environmentVariables:
-        MODELSCOPE_CACHE: ${resources.model_cache.output.nasConfig.mountPoints.0.mountDir}
+        MODEL_ID: ${vars.modelID}
+        MODEL_VERSION: ${vars.modelRevision}
+        TASK: ${vars.modelTask}
+        PATH: /opt/python3.9/bin:/usr/local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/opt/bin
+        PYTHONPATH: /opt/python
+      customRuntimeConfig:
+        command:
+          - python3.9
+        args:
+          - "app.py"
+        port: 9000
       triggers:
         - triggerName: httpTrigger
           triggerType: http