base complete

Author: chibu

examples/guidellm_example.py

@@ -14,7 +14,9 @@
     #target="http://fed73cc1-us-east.lb.appdomain.cloud/v1",
     data_type="emulated",
     max_seconds=30,
-    data="prompt_tokens=512,generated_tokens=256",
+    #data="{'prompt_tokens': 512, 'generated_tokens': 256, 'output_tokens' : 256}",
+    data="prompt_tokens=512,generated_tokens=256,output_tokens=256",
+    #data="prompt_tokens=512,generated_tokens=256",
     branch = "update_guidellm",
     #vllm_kwargs={"enable-chunked-prefill": True}
 )

src/automation/tasks/guidellm.py

@@ -4,8 +4,8 @@
 import os
 
 DEFAULT_SERVER_WAIT_TIME = 600 # 600 seconds = 10 minutes
-#GUIDELLM_PACKAGE = "git+https://github.com/neuralmagic/guidellm.git"
-GUIDELLM_PACKAGE = "git+https://github.com/neuralmagic/guidellm.git@clearml-guidellm"
+GUIDELLM_PACKAGE = "git+https://github.com/neuralmagic/guidellm.git"
+#GUIDELLM_PACKAGE = "git+https://github.com/neuralmagic/guidellm.git@clearml-guidellm"
 #GUIDELLM_PACKAGE = "git+https://github.com/neuralmagic/guidellm.git@clearml-guidellm#egg=guidellm"
 #GUIDELLM_PACKAGE = "git+https://github.com/neuralmagic/guidellm.git@main#egg=guidellm[dev]"