Fix logit processor bugs (#427)

test/srt/example_image.png

@@ -0,0 +1 @@
+../lang/example_image.png

test/srt/test_httpserver_decode.py

@@ -3,7 +3,6 @@ Usage:
 python3 -m sglang.launch_server --model-path TinyLlama/TinyLlama-1.1B-Chat-v0.4 --port 30000
 python3 test_httpserver_decode.py
 
-
 Output:
 The capital of France is Paris.\nThe capital of the United States is Washington, D.C.\nThe capital of Canada is Ottawa.\nThe capital of Japan is Tokyo
 """
@@ -23,6 +22,7 @@ def test_decode(url, return_logprob, top_logprobs_num, return_text):
                 "temperature": 0,
                 "max_new_tokens": 32,
             },
+            "stream": False,
             "return_logprob": return_logprob,
             "top_logprobs_num": top_logprobs_num,
             "return_text_in_logprobs": return_text,

test/srt/test_httpserver_decode_stream.py

@@ -26,6 +26,7 @@ def test_decode_stream(url, return_logprob, top_logprobs_num):
             "return_logprob": return_logprob,
             "top_logprobs_num": top_logprobs_num,
             "return_text_in_logprobs": True,
+            "logprob_start_len": 0,
         },
         stream=True,
     )

test/srt/test_httpserver_llava.py

@@ -34,7 +34,7 @@ async def test_concurrent(args):
                 url + "/generate",
                 {
                     "text": "A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions. USER: <image>\nDescribe this picture ASSISTANT:",
-                    "image_data": "test_image.png",
+                    "image_data": "example_image.png",
                     "sampling_params": {
                         "temperature": 0,
                         "max_new_tokens": 16,
@@ -55,7 +55,7 @@ def test_streaming(args):
         url + "/generate",
         json={
             "text": "A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions. USER: <image>\nDescribe this picture ASSISTANT:",
-            "image_data": "test_image.png",
+            "image_data": "example_image.png",
             "sampling_params": {
                 "temperature": 0,
                 "max_new_tokens": 128,

test/srt/test_httpserver_reuse.py

@@ -6,10 +6,10 @@ The capital of France is Paris.\nThe capital of the United States is Washington,
 """
 
 import argparse
-import time
 
 import requests
 
+
 if __name__ == "__main__":
     parser = argparse.ArgumentParser()
     parser.add_argument("--host", type=str, default="http://127.0.0.1")

test/srt/test_openai_server.py

@@ -163,7 +163,7 @@ def test_regex(args):
     regex = (
         r"""\{\n"""
         + r"""   "name": "[\w]+",\n"""
-        + r"""   "population": "[\w\d\s]+"\n"""
+        + r"""   "population": [\w\d\s]+\n"""
         + r"""\}"""
     )