Update quick start examples (#120)

examples/quick_start/srt_example_llava.py

@@ -10,29 +10,53 @@ def image_qa(s, image_path, question):
     s += sgl.assistant(sgl.gen("answer"))
 
 
-runtime = sgl.Runtime(model_path="liuhaotian/llava-v1.5-7b",
-                      tokenizer_path="llava-hf/llava-1.5-7b-hf")
-sgl.set_default_backend(runtime)
+def single():
+    state = image_qa.run(
+        image_path="images/cat.jpeg",
+        question="What is this?",
+        max_new_tokens=64)
+    print(state["answer"], "\n")
 
 
-# Single
-state = image_qa.run(
-    image_path="images/cat.jpeg",
-    question="What is this?",
-    max_new_tokens=64)
-print(state["answer"], "\n")
+def stream():
+    state = image_qa.run(
+        image_path="images/cat.jpeg",
+        question="What is this?",
+        max_new_tokens=64,
+        stream=True)
+
+    for out in state.text_iter("answer"):
+        print(out, end="", flush=True)
+    print()
 
 
-# Batch
-states = image_qa.run_batch(
-    [
-        {"image_path": "images/cat.jpeg", "question":"What is this?"},
-        {"image_path": "images/dog.jpeg", "question":"What is this?"},
-    ],
-    max_new_tokens=64,
-)
-for s in states:
-    print(s["answer"], "\n")
+def batch():
+    states = image_qa.run_batch(
+        [
+            {"image_path": "images/cat.jpeg", "question":"What is this?"},
+            {"image_path": "images/dog.jpeg", "question":"What is this?"},
+        ],
+        max_new_tokens=64,
+    )
+    for s in states:
+        print(s["answer"], "\n")
 
 
-runtime.shutdown()
+if __name__ == "__main__":
+    runtime = sgl.Runtime(model_path="liuhaotian/llava-v1.5-7b",
+                          tokenizer_path="llava-hf/llava-1.5-7b-hf")
+    sgl.set_default_backend(runtime)
+
+    # Run a single request
+    print("\n========== single ==========\n")
+    single()
+
+    # Stream output
+    print("\n========== stream ==========\n")
+    stream()
+
+    # Run a batch of requests
+    print("\n========== batch ==========\n")
+    batch()
+
+    runtime.shutdown()