Improve the control of streaming and improve the first token latency in streaming (#117)

2024-01-29 17:05:42 -08:00
parent cd6872334e
commit 6f560c761b
12 changed files with 46 additions and 23 deletions
--- a/test/srt/test_httpserver_decode.py
+++ b/test/srt/test_httpserver_decode.py
@@ -1,5 +1,8 @@
 """
+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