chore: import upstream snapshot with attribution

2026-07-13 13:34:58 +08:00
commit a203934033
1368 changed files with 175001 additions and 0 deletions
@@ -0,0 +1,44 @@
+# Copyright (c) ModelScope Contributors. All rights reserved.
+import os
+from openai import OpenAI
+
+os.environ['CUDA_VISIBLE_DEVICES'] = '0'
+
+
+def infer(client, model: str, messages):
+    resp = client.chat.completions.create(model=model, messages=messages)
+    classify = resp.choices[0].message.content
+    print(f'messages: {messages}')
+    print(f'classify: {classify}')
+    return classify
+
+
+def run_client(host: str = '127.0.0.1', port: int = 8000):
+    client = OpenAI(
+        api_key='EMPTY',
+        base_url=f'http://{host}:{port}/v1',
+    )
+    model = client.models.list().data[0].id
+    print(f'model: {model}')
+
+    messages = [{
+        'role': 'user',
+        'content': 'What is the capital of China?',
+    }, {
+        'role': 'assistant',
+        'content': 'Beijing',
+    }]
+    infer(client, model, messages)
+
+
+if __name__ == '__main__':
+    from swift import DeployArguments, run_deploy
+    with run_deploy(
+            DeployArguments(
+                model='/your/seq_cls/checkpoint-xxx',
+                task_type='seq_cls',
+                infer_backend='vllm',
+                num_labels=2,
+                verbose=False,
+                log_interval=-1)) as port:
+        run_client(port=port)
@@ -0,0 +1,9 @@
+# GME/GTE models or your checkpoints are also supported
+# transformers/vllm/sglang supported
+CUDA_VISIBLE_DEVICES=0 swift deploy \
+    --host 0.0.0.0 \
+    --port 8000 \
+    --model /your/seq_cls/checkpoint-xxx \
+    --infer_backend vllm \
+    --task_type seq_cls \
+    --num_labels 2 \