Update to vastai package

workers/openai/worker.py

@@ -7,14 +7,12 @@ from vastai import Worker, WorkerConfig, HandlerConfig, LogActionConfig, Benchma
 # vLLM model configuration
 MODEL_SERVER_URL           = 'http://127.0.0.1'
 MODEL_SERVER_PORT          = 18000
-MODEL_LOG_FILE             = '/var/log/onstart.log'
+MODEL_LOG_FILE             = '/var/log/portal/vllm.log'
 MODEL_HEALTHCHECK_ENDPOINT = "/health"
 
 # vLLM-specific log messages
 MODEL_LOAD_LOG_MSG = [
     "Application startup complete.",
-    "llama runner started in",
-    "Server listening on",
 ]
 
 MODEL_ERROR_LOG_MSGS = [

workers/tgi/worker.py

@@ -35,7 +35,7 @@ def benchmark_generator() -> dict:
     benchmark_data = {
         "inputs": prompt,
         "parameters": {
-            "max_new_tokens": 500,
+            "max_new_tokens": 128,
             "temperature": 0.7,
             "return_full_text": False
         }