Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
File size: 8,891 Bytes
970eef1 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 |
#!/usr/bin/env python3
"""
Script to run lighteval tests in parallel for multiple models
"""
import os
import sys
import json
import time
import tempfile
import asyncio
from pathlib import Path
from typing import Tuple, List, Dict, Any
# Ensure environment is properly configured
from dotenv import load_dotenv
load_dotenv()
# Import yourbench task module
sys.path.append(os.getcwd())
from tasks.yourbench_lighteval_task import create_yourbench_task
# Define models to test
INIT_MODELS = [
# 70B
("Qwen/Qwen2.5-72B-Instruct", "novita"),
("meta-llama/Llama-3.3-70B-Instruct", "novita"),
("deepseek-ai/DeepSeek-R1-Distill-Llama-70B", "novita"),
# 20 to 30B
("Qwen/QwQ-32B", "novita"),
# ("mistralai/Mistral-Small-24B-Instruct-2501", "sambanova"),
]
async def run_lighteval_test_for_model(model_info: Tuple[str, str]) -> Dict[str, Any]:
"""
Run lighteval test for a specific model
"""
model_name, provider = model_info
# Parameters
dataset_name = "yourbench_a"
organization = "yourbench"
output_dir = f"uploaded_files/test_parallel_{provider}/lighteval_results"
# Create output directory
os.makedirs(output_dir, exist_ok=True)
# Define full dataset path
dataset_path = f"{organization}/{dataset_name}"
print(f"Dataset to evaluate for {model_name}: {dataset_path}")
# Create temporary file
temp_file_path = tempfile.mktemp(suffix=".py")
print(f"Creating temporary file for {model_name}: {temp_file_path}")
with open(temp_file_path, 'w') as temp_file:
temp_file.write(f"""
import os
import sys
sys.path.append("{os.getcwd()}")
from tasks.yourbench_lighteval_task import create_yourbench_task
# Create yourbench task
yourbench = create_yourbench_task("{dataset_path}", "lighteval")
# Define TASKS_TABLE needed by lighteval
TASKS_TABLE = [yourbench]
""")
# Build lighteval command args
cmd_args = [
"lighteval",
"endpoint",
"inference-providers",
f"model={model_name},provider={provider}",
"custom|yourbench|0|0",
"--custom-tasks",
temp_file_path,
"--max-samples", "5",
"--output-dir", output_dir,
"--save-details",
"--no-push-to-hub"
]
print(f"Running command for {model_name}: {' '.join(cmd_args)}")
print(f"Start time for {model_name}: {time.strftime('%H:%M:%S')}")
results = {
"model_name": model_name,
"provider": provider,
"success": False,
"error": None,
"results": None,
"return_code": None
}
try:
# Prepare environment with needed tokens
env = os.environ.copy()
hf_token = os.getenv("HF_TOKEN")
if hf_token:
env["HF_TOKEN"] = hf_token
env["HUGGING_FACE_HUB_TOKEN"] = hf_token
env["HF_ORGANIZATION"] = organization
# Run the process asynchronously
process = await asyncio.create_subprocess_exec(
*cmd_args,
stdout=asyncio.subprocess.PIPE,
stderr=asyncio.subprocess.PIPE,
env=env
)
# Wait for the process to complete
stdout, stderr = await process.communicate()
# Store return code
exit_code = process.returncode
results["return_code"] = exit_code
# Log some output for debugging
if stdout:
stdout_lines = stdout.decode().strip().split('\n')
if stdout_lines and len(stdout_lines) > 0:
print(f"Output from {model_name}: {stdout_lines[0]}")
# Check if results were generated
results_dir = Path(output_dir) / "results"
if results_dir.exists():
result_files = list(results_dir.glob("**/*.json"))
if result_files:
# Read the first results file
with open(result_files[0], 'r') as f:
test_results = json.load(f)
results["results"] = test_results
results["success"] = True
except asyncio.CancelledError:
results["error"] = "Task cancelled"
print(f"Task cancelled for {model_name}")
except Exception as e:
results["error"] = f"Exception: {str(e)}"
print(f"Error running test for {model_name}: {str(e)}")
finally:
# Delete temporary file
try:
os.unlink(temp_file_path)
except:
pass
print(f"End time for {model_name}: {time.strftime('%H:%M:%S')}")
return results
async def run_parallel_tests(models: List[Tuple[str, str]]) -> List[Dict[str, Any]]:
"""
Run tests in parallel for multiple models using asyncio
"""
print(f"Starting parallel tests for {len(models)} models")
# Create tasks for each model
tasks = [run_lighteval_test_for_model(model) for model in models]
# Run all tasks concurrently and gather results
model_results = await asyncio.gather(*tasks, return_exceptions=True)
# Process results
results = []
for i, result in enumerate(model_results):
if isinstance(result, Exception):
# Handle exception
model_name, provider = models[i]
print(f"Test failed for {model_name}: {str(result)}")
results.append({
"model_name": model_name,
"provider": provider,
"success": False,
"error": str(result),
"results": None,
"return_code": None
})
else:
# Valid result
results.append(result)
print(f"Test completed for {result['model_name']}")
return results
def format_comparison_results(results: List[Dict[str, Any]]) -> Dict[str, Any]:
"""
Format results for easy comparison between models
"""
comparison = {
"metadata": {
"timestamp": time.strftime("%Y-%m-%d %H:%M:%S"),
"total_models_tested": len(results),
"successful_tests": len([r for r in results if r["success"]])
},
"models_comparison": []
}
# Sort models by accuracy (if available) or name
sorted_results = sorted(
results,
key=lambda x: (
x["results"]["results"]["all"]["accuracy"] if x["success"] and x["results"] else -1,
x["model_name"]
),
reverse=True
)
for result in sorted_results:
model_result = {
"model_name": result["model_name"],
"provider": result["provider"],
"success": result["success"]
}
if result["success"] and result["results"]:
model_result.update({
"accuracy": result["results"]["results"]["all"]["accuracy"],
"accuracy_stderr": result["results"]["results"]["all"]["accuracy_stderr"],
"evaluation_time": float(result["results"]["config_general"]["total_evaluation_time_secondes"])
})
else:
model_result["error"] = result["error"]
comparison["models_comparison"].append(model_result)
return comparison
async def main_async():
"""
Async main function to run parallel tests
"""
print("Starting parallel lighteval tests")
start_time = time.time()
# Run tests in parallel
results = await run_parallel_tests(INIT_MODELS)
# Save detailed results
detailed_output_file = "parallel_test_detailed_results.json"
with open(detailed_output_file, 'w') as f:
json.dump(results, f, indent=2)
# Generate and save comparison results
comparison = format_comparison_results(results)
comparison_file = "models_comparison.json"
with open(comparison_file, 'w') as f:
json.dump(comparison, f, indent=2)
# Print summary
print("\nTest Summary:")
for model in comparison["models_comparison"]:
status = "✅" if model["success"] else "❌"
print(f"{status} {model['model_name']} ({model['provider']})")
if not model["success"]:
print(f" Error: {model['error']}")
else:
print(f" Accuracy: {model['accuracy']:.2%} (±{model['accuracy_stderr']:.2%})")
print(f" Evaluation time: {model['evaluation_time']:.2f}s")
duration = time.time() - start_time
print(f"\nTotal execution time: {duration:.2f} seconds")
print(f"Detailed results saved to: {detailed_output_file}")
print(f"Comparison results saved to: {comparison_file}")
def main():
"""
Main function to run parallel tests
"""
# Create event loop and run the async main
loop = asyncio.get_event_loop()
loop.run_until_complete(main_async())
loop.close()
if __name__ == "__main__":
main() |