Update to vastai package

.devcontainer/devcontainer.json

@@ -1,37 +0,0 @@
-// .devcontainer/devcontainer.json
-// Dev container for the Vast.ai serverless Ollama template.
-// Includes Docker-in-Docker so you can build and test images from inside the container.
-{
-  "name": "vast.ai-serverless-ollama",
-  "image": "mcr.microsoft.com/devcontainers/base:trixie",
-  "features": {
-    "ghcr.io/devcontainers/features/python:1": {
-      "installTools": true,
-      "version": "3.12"
-    },
-    "ghcr.io/devcontainers/features/docker-in-docker:3.0.0": {
-      "moby": false,
-      "version": "latest",
-      "installDockerBuildx": true,
-      "dockerDashComposeVersion": "v2"
-    }
-  },
-  "runArgs": ["--privileged"],
-  "containerEnv": {
-    "DOCKER_BUILDKIT": "1"
-  },
-  "postCreateCommand": "python3 -m pip install --user --upgrade pip && python3 -m pip install --user -r requirements.txt pyyaml",
-  "customizations": {
-    "vscode": {
-      "extensions": [
-        "ms-python.python",
-        "ms-azuretools.vscode-docker"
-      ],
-      "settings": {
-        "python.defaultInterpreterPath": "/usr/bin/python3",
-        "terminal.integrated.defaultProfile.linux": "bash",
-        "docker.showStartPage": false
-      }
-    }
-  }
-}

workers/openai/worker.py

@@ -7,15 +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",
-    "msg=\"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
         }