File size: 15,940 Bytes
788f4ae |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 |
"""
Helion-OSC Comprehensive Benchmark Suite
Performance benchmarking and comparison with other models
"""
import torch
import time
import psutil
import numpy as np
from typing import Dict, List, Any, Optional
from dataclasses import dataclass, asdict
import json
import logging
from transformers import AutoTokenizer, AutoModelForCausalLM
from tqdm import tqdm
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
@dataclass
class BenchmarkResult:
"""Single benchmark result"""
model_name: str
task: str
prompt_length: int
generation_length: int
temperature: float
inference_time: float
tokens_per_second: float
memory_used_mb: float
gpu_memory_mb: Optional[float]
success: bool
error: Optional[str] = None
@dataclass
class AggregatedResults:
"""Aggregated benchmark results"""
model_name: str
total_tests: int
successful_tests: int
failed_tests: int
avg_inference_time: float
avg_tokens_per_second: float
avg_memory_mb: float
min_inference_time: float
max_inference_time: float
std_inference_time: float
class PerformanceBenchmark:
"""Performance benchmarking utilities"""
def __init__(self, model_name: str = "DeepXR/Helion-OSC"):
self.model_name = model_name
self.device = "cuda" if torch.cuda.is_available() else "cpu"
logger.info(f"Loading model: {model_name}")
self.tokenizer = AutoTokenizer.from_pretrained(model_name)
self.model = AutoModelForCausalLM.from_pretrained(
model_name,
torch_dtype=torch.bfloat16 if self.device == "cuda" else torch.float32,
device_map="auto" if self.device == "cuda" else None
)
if self.device == "cpu":
self.model = self.model.to(self.device)
self.model.eval()
self.results: List[BenchmarkResult] = []
def get_memory_usage(self) -> tuple:
"""Get current memory usage"""
process = psutil.Process()
ram_mb = process.memory_info().rss / 1024 / 1024
gpu_mb = None
if torch.cuda.is_available():
gpu_mb = torch.cuda.memory_allocated() / 1024 / 1024
return ram_mb, gpu_mb
def benchmark_inference(
self,
prompt: str,
task: str,
max_length: int = 512,
temperature: float = 0.7,
num_runs: int = 1
) -> List[BenchmarkResult]:
"""Benchmark inference performance"""
run_results = []
for run in range(num_runs):
try:
# Tokenize
inputs = self.tokenizer(prompt, return_tensors="pt").to(self.device)
prompt_length = inputs.input_ids.shape[1]
# Warm up GPU
if run == 0 and self.device == "cuda":
with torch.no_grad():
_ = self.model.generate(**inputs, max_length=prompt_length + 10)
torch.cuda.synchronize()
# Clear cache
if self.device == "cuda":
torch.cuda.empty_cache()
# Measure memory before
ram_before, gpu_before = self.get_memory_usage()
# Generate
start_time = time.time()
with torch.no_grad():
outputs = self.model.generate(
**inputs,
max_length=max_length,
temperature=temperature,
do_sample=temperature > 0,
pad_token_id=self.tokenizer.eos_token_id
)
if self.device == "cuda":
torch.cuda.synchronize()
end_time = time.time()
# Measure memory after
ram_after, gpu_after = self.get_memory_usage()
# Calculate metrics
inference_time = end_time - start_time
generation_length = outputs.shape[1] - prompt_length
tokens_per_second = generation_length / inference_time if inference_time > 0 else 0
memory_used = ram_after - ram_before
gpu_memory = (gpu_after - gpu_before) if gpu_after and gpu_before else None
result = BenchmarkResult(
model_name=self.model_name,
task=task,
prompt_length=prompt_length,
generation_length=generation_length,
temperature=temperature,
inference_time=inference_time,
tokens_per_second=tokens_per_second,
memory_used_mb=memory_used,
gpu_memory_mb=gpu_memory,
success=True
)
run_results.append(result)
self.results.append(result)
except Exception as e:
logger.error(f"Benchmark failed: {e}")
result = BenchmarkResult(
model_name=self.model_name,
task=task,
prompt_length=0,
generation_length=0,
temperature=temperature,
inference_time=0,
tokens_per_second=0,
memory_used_mb=0,
gpu_memory_mb=None,
success=False,
error=str(e)
)
run_results.append(result)
self.results.append(result)
return run_results
def run_benchmark_suite(self) -> List[BenchmarkResult]:
"""Run comprehensive benchmark suite"""
logger.info("Starting comprehensive benchmark suite...")
test_cases = [
{
"prompt": "def fibonacci(n):",
"task": "simple_function",
"max_length": 256,
"temperature": 0.7
},
{
"prompt": "Write a Python class for a binary search tree with insert, search, and delete methods:",
"task": "complex_class",
"max_length": 1024,
"temperature": 0.7
},
{
"prompt": "Implement quicksort algorithm in Python with detailed comments:",
"task": "algorithm",
"max_length": 512,
"temperature": 0.5
},
{
"prompt": "Solve: What is the derivative of f(x) = x^3 + 2x^2 - 5x + 3?",
"task": "math_simple",
"max_length": 256,
"temperature": 0.3
},
{
"prompt": "Prove using mathematical induction that the sum of first n natural numbers is n(n+1)/2:",
"task": "math_proof",
"max_length": 1024,
"temperature": 0.2
},
{
"prompt": "Design a RESTful API for a todo list application with proper documentation:",
"task": "system_design",
"max_length": 2048,
"temperature": 0.7
},
]
all_results = []
for test_case in tqdm(test_cases, desc="Running benchmarks"):
results = self.benchmark_inference(
prompt=test_case["prompt"],
task=test_case["task"],
max_length=test_case["max_length"],
temperature=test_case["temperature"],
num_runs=3
)
all_results.extend(results)
logger.info("Benchmark suite completed!")
return all_results
def aggregate_results(self) -> AggregatedResults:
"""Aggregate benchmark results"""
if not self.results:
raise ValueError("No benchmark results available")
successful = [r for r in self.results if r.success]
if not successful:
raise ValueError("No successful benchmark runs")
inference_times = [r.inference_time for r in successful]
tokens_per_sec = [r.tokens_per_second for r in successful]
memory_usage = [r.memory_used_mb for r in successful]
return AggregatedResults(
model_name=self.model_name,
total_tests=len(self.results),
successful_tests=len(successful),
failed_tests=len(self.results) - len(successful),
avg_inference_time=np.mean(inference_times),
avg_tokens_per_second=np.mean(tokens_per_sec),
avg_memory_mb=np.mean(memory_usage),
min_inference_time=np.min(inference_times),
max_inference_time=np.max(inference_times),
std_inference_time=np.std(inference_times)
)
def save_results(self, output_file: str = "benchmark_results.json"):
"""Save benchmark results to file"""
results_dict = [asdict(r) for r in self.results]
with open(output_file, 'w') as f:
json.dump(results_dict, f, indent=2)
logger.info(f"Results saved to {output_file}")
def generate_report(self, output_file: str = "benchmark_report.txt"):
"""Generate human-readable benchmark report"""
agg = self.aggregate_results()
report = f"""
{'='*80}
HELION-OSC BENCHMARK REPORT
{'='*80}
Model: {agg.model_name}
Device: {self.device}
OVERALL STATISTICS
{'='*80}
Total Tests: {agg.total_tests}
Successful: {agg.successful_tests}
Failed: {agg.failed_tests}
Success Rate: {(agg.successful_tests/agg.total_tests)*100:.2f}%
PERFORMANCE METRICS
{'='*80}
Average Inference Time: {agg.avg_inference_time:.4f} seconds
Min Inference Time: {agg.min_inference_time:.4f} seconds
Max Inference Time: {agg.max_inference_time:.4f} seconds
Std Inference Time: {agg.std_inference_time:.4f} seconds
Average Tokens/Second: {agg.avg_tokens_per_second:.2f}
Average Memory Usage: {agg.avg_memory_mb:.2f} MB
PER-TASK BREAKDOWN
{'='*80}
"""
# Group by task
df = pd.DataFrame([asdict(r) for r in self.results if r.success])
if not df.empty:
task_stats = df.groupby('task').agg({
'inference_time': ['mean', 'min', 'max'],
'tokens_per_second': 'mean',
'memory_used_mb': 'mean'
})
report += task_stats.to_string()
report += f"\n\n{'='*80}\n"
with open(output_file, 'w') as f:
f.write(report)
logger.info(f"Report saved to {output_file}")
print(report)
def plot_results(self, output_dir: str = "./benchmark_plots"):
"""Generate visualization plots"""
import os
os.makedirs(output_dir, exist_ok=True)
df = pd.DataFrame([asdict(r) for r in self.results if r.success])
if df.empty:
logger.warning("No data to plot")
return
# Set style
sns.set_style("whitegrid")
# Plot 1: Inference time by task
plt.figure(figsize=(12, 6))
sns.barplot(data=df, x='task', y='inference_time')
plt.xticks(rotation=45, ha='right')
plt.title('Inference Time by Task')
plt.ylabel('Time (seconds)')
plt.tight_layout()
plt.savefig(f"{output_dir}/inference_time_by_task.png", dpi=300)
plt.close()
# Plot 2: Tokens per second by task
plt.figure(figsize=(12, 6))
sns.barplot(data=df, x='task', y='tokens_per_second')
plt.xticks(rotation=45, ha='right')
plt.title('Tokens Per Second by Task')
plt.ylabel('Tokens/Second')
plt.tight_layout()
plt.savefig(f"{output_dir}/tokens_per_second_by_task.png", dpi=300)
plt.close()
# Plot 3: Memory usage by task
plt.figure(figsize=(12, 6))
sns.barplot(data=df, x='task', y='memory_used_mb')
plt.xticks(rotation=45, ha='right')
plt.title('Memory Usage by Task')
plt.ylabel('Memory (MB)')
plt.tight_layout()
plt.savefig(f"{output_dir}/memory_usage_by_task.png", dpi=300)
plt.close()
# Plot 4: Scatter plot - generation length vs inference time
plt.figure(figsize=(10, 6))
sns.scatterplot(data=df, x='generation_length', y='inference_time', hue='task', s=100)
plt.title('Generation Length vs Inference Time')
plt.xlabel('Generation Length (tokens)')
plt.ylabel('Inference Time (seconds)')
plt.legend(bbox_to_anchor=(1.05, 1), loc='upper left')
plt.tight_layout()
plt.savefig(f"{output_dir}/length_vs_time.png", dpi=300)
plt.close()
logger.info(f"Plots saved to {output_dir}")
class ComparisonBenchmark:
"""Compare multiple models"""
def __init__(self, model_names: List[str]):
self.model_names = model_names
self.benchmarks = {}
def run_comparison(self):
"""Run benchmarks for all models"""
for model_name in self.model_names:
logger.info(f"\nBenchmarking {model_name}...")
try:
benchmark = PerformanceBenchmark(model_name)
benchmark.run_benchmark_suite()
self.benchmarks[model_name] = benchmark
except Exception as e:
logger.error(f"Failed to benchmark {model_name}: {e}")
def generate_comparison_report(self, output_file: str = "comparison_report.txt"):
"""Generate comparison report"""
report = f"""
{'='*80}
MODEL COMPARISON REPORT
{'='*80}
"""
for model_name, benchmark in self.benchmarks.items():
agg = benchmark.aggregate_results()
report += f"""
Model: {model_name}
{'='*80}
Avg Inference Time: {agg.avg_inference_time:.4f}s
Avg Tokens/Second: {agg.avg_tokens_per_second:.2f}
Avg Memory Usage: {agg.avg_memory_mb:.2f} MB
Success Rate: {(agg.successful_tests/agg.total_tests)*100:.2f}%
"""
with open(output_file, 'w') as f:
f.write(report)
print(report)
logger.info(f"Comparison report saved to {output_file}")
def main():
"""Main benchmark script"""
import argparse
parser = argparse.ArgumentParser(description="Benchmark Helion-OSC model")
parser.add_argument("--model", type=str, default="DeepXR/Helion-OSC")
parser.add_argument("--output-dir", type=str, default="./benchmark_results")
parser.add_argument("--compare", nargs='+', help="List of models to compare")
parser.add_argument("--plot", action="store_true", help="Generate plots")
args = parser.parse_args()
import os
os.makedirs(args.output_dir, exist_ok=True)
if args.compare:
# Comparison mode
comparison = ComparisonBenchmark(args.compare)
comparison.run_comparison()
comparison.generate_comparison_report(
os.path.join(args.output_dir, "comparison_report.txt")
)
else:
# Single model benchmark
benchmark = PerformanceBenchmark(args.model)
benchmark.run_benchmark_suite()
benchmark.save_results(os.path.join(args.output_dir, "benchmark_results.json"))
benchmark.generate_report(os.path.join(args.output_dir, "benchmark_report.txt"))
if args.plot:
benchmark.plot_results(os.path.join(args.output_dir, "plots"))
if __name__ == "__main__":
main() |