refactor: Remove coupled Unsloth service, upgrade to vLLM 0.11+ (#497)

Author: bradhilton

dev/playwright_agent/playwright_agent.py

@@ -4,13 +4,13 @@
 
 import argparse
 import asyncio
+from dataclasses import dataclass
 import json
 import math
 import os
 import time
-import uuid
-from dataclasses import dataclass
 from typing import Any, Dict, List
+import uuid
 
 from dotenv import load_dotenv
 from mcp import ClientSession, StdioServerParameters

dev/swebench/logs.py

@@ -1,10 +1,10 @@
 import logging
 from logging import Handler, LogRecord
 
-import litellm
 from langfuse import Langfuse
 from langfuse.decorators import langfuse_context
 from langfuse.types import SpanLevel
+import litellm
 from sweagent.agent.agents import DefaultAgent
 from sweagent.run.hooks.apply_patch import SaveApplyPatchHook
 

dev/swebench/rollout.py

@@ -1,19 +1,19 @@
 import asyncio
-import json
 from http.client import RemoteDisconnected
+import json
 from pathlib import Path
 from typing import Any, Literal, overload
 
-import modal
-import requests
 from aiolimiter import AsyncLimiter
 from config import get_config
 from eval import eval_instance
 from grpclib.exceptions import StreamTerminatedError
 from instances import Instance
 from langfuse.decorators import observe
 from logs import setup_agent_logger
+import modal
 from pydantic import BaseModel
+import requests
 from requests import adapters as requests_adapters
 from requests.exceptions import ConnectionError, ConnectTimeout, SSLError
 from run import run

dev/swebench/sandbox/test.py

@@ -1,5 +1,5 @@
-import pytest
 from dotenv import load_dotenv
+import pytest
 
 from ..instances import as_instances_iter, get_filtered_swe_smith_instances_df
 from .new import new_sandbox

dev/swebench/train.ipynb

@@ -36,8 +36,8 @@
    "source": [
     "import os\n",
     "\n",
-    "import nest_asyncio\n",
     "from dotenv import load_dotenv\n",
+    "import nest_asyncio\n",
     "\n",
     "# PREPARE ENVIRONMENT\n",
     "os.environ[\"MODAL_IMAGE_BUILDER_VERSION\"] = \"2024.10\"\n",
@@ -60,8 +60,8 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import torch\n",
     "from rollout import ModelConfig\n",
+    "import torch\n",
     "\n",
     "import art\n",
     "from art.local import LocalBackend\n",

dev/tau-bench/analyze_model_errors.py

@@ -2,13 +2,13 @@
 
 import argparse
 import asyncio
+from datetime import datetime
 import json
 import os
-from datetime import datetime
 from typing import Any, Dict, List
 
-import litellm
 from dotenv import load_dotenv
+import litellm
 from litellm import provider_list
 
 # Import evaluate_model and rollout functions from run_rl

dev/tau-bench/auto_error_identification.py

@@ -1,9 +1,9 @@
 # Copyright Sierra
 
 import argparse
-import json
 from concurrent.futures import ThreadPoolExecutor
 from enum import Enum
+import json
 from typing import Any, Dict, List
 
 from pydantic import BaseModel

dev/tau-bench/benchmark_rl.py

@@ -2,13 +2,13 @@
 
 import argparse
 import asyncio
+from datetime import datetime
 import json
 import os
-from datetime import datetime
 from typing import Any, Dict, List
 
-import litellm
 from dotenv import load_dotenv
+import litellm
 from litellm import provider_list
 
 # Import evaluate_model and rollout functions from run_rl

dev/tau-bench/run.ipynb

@@ -37,11 +37,11 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import torch\n",
     "from dotenv import load_dotenv\n",
     "from run import RunConfig\n",
     "from run_rl import train\n",
     "from tau_bench.types import TauBenchPolicyConfig, TauBenchTrainingConfig\n",
+    "import torch\n",
     "\n",
     "import art\n",
     "\n",

dev/tau-bench/run_training.py

@@ -4,8 +4,8 @@
 import textwrap
 import traceback
 
-import sky
 from dotenv import dotenv_values
+import sky
 from sky import ClusterStatus
 from tau_bench.types import RunConfig, TauBenchPolicyConfig, TauBenchTrainingConfig