Spaces:
Sleeping
Sleeping
File size: 4,719 Bytes
f2a274f |
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 |
import gradio as gr
import time
import json
import pandas as pd
from typing import List, Dict, Any
class BenchmarkSystem:
def __init__(self):
self.results = {}
def run_benchmark(self,
model_name: str,
test_cases: List[str],
system_prompt: str = "") -> Dict[str, Any]:
"""
Run benchmark tests and measure performance metrics
"""
results = {
"model_name": model_name,
"timestamp": time.strftime("%Y-%m-%d %H:%M:%S"),
"total_tokens": 0,
"total_time": 0,
"responses": [],
"metrics": {}
}
start_time = time.time()
# Simulate processing test cases
for test in test_cases:
# Here you would add actual model inference
# This is a placeholder for demonstration
time.sleep(0.5) # Simulate processing time
results["responses"].append({
"input": test,
"output": f"Sample response for: {test}",
"tokens": len(test.split()),
"time": 0.5
})
results["total_time"] = time.time() - start_time
results["total_tokens"] = sum(r["tokens"] for r in results["responses"])
# Calculate aggregate metrics
results["metrics"] = {
"avg_response_time": results["total_time"] / len(test_cases),
"avg_tokens_per_response": results["total_tokens"] / len(test_cases)
}
self.results[model_name] = results
return results
def format_results(results: Dict[str, Any]) -> str:
"""Format benchmark results for display"""
output = f"Model: {results['model_name']}\n"
output += f"Timestamp: {results['timestamp']}\n"
output += f"Total Time: {results['total_time']:.2f}s\n"
output += f"Total Tokens: {results['total_tokens']}\n\n"
output += "Metrics:\n"
for metric, value in results["metrics"].items():
output += f"- {metric}: {value:.2f}\n"
return output
def save_results(results: Dict[str, Any], filename: str = "benchmark_results.json"):
"""Save benchmark results to a file"""
with open(filename, "w") as f:
json.dump(results, f, indent=2)
return f"Results saved to {filename}"
def run_benchmark_interface(model_name: str,
test_cases: str,
system_prompt: str) -> tuple[str, pd.DataFrame]:
"""
Gradio interface function for running benchmarks
"""
benchmark = BenchmarkSystem()
# Parse test cases (assuming one per line)
test_cases_list = [t.strip() for t in test_cases.split("\n") if t.strip()]
# Run benchmark
results = benchmark.run_benchmark(
model_name=model_name,
test_cases=test_cases_list,
system_prompt=system_prompt
)
# Create DataFrame for response details
df = pd.DataFrame([
{
"Input": r["input"],
"Output": r["output"],
"Tokens": r["tokens"],
"Time (s)": r["time"]
}
for r in results["responses"]
])
# Save results
save_results(results)
return format_results(results), df
# Create Gradio interface
with gr.Blocks(title="Model Benchmark Suite") as demo:
gr.Markdown("# Model Benchmark Suite")
gr.Markdown("Test and compare model performance across different scenarios")
with gr.Row():
with gr.Column():
model_name = gr.Textbox(
label="Model Name",
placeholder="Enter model name or identifier"
)
system_prompt = gr.Textbox(
label="System Prompt (Optional)",
placeholder="Enter system prompt if applicable",
lines=2
)
test_cases = gr.Textbox(
label="Test Cases",
placeholder="Enter test cases (one per line)",
lines=5
)
run_button = gr.Button("Run Benchmark")
with gr.Column():
results_text = gr.Textbox(
label="Benchmark Results",
lines=10,
readonly=True
)
results_table = gr.DataFrame(
label="Detailed Results",
headers=["Input", "Output", "Tokens", "Time (s)"]
)
run_button.click(
fn=run_benchmark_interface,
inputs=[model_name, test_cases, system_prompt],
outputs=[results_text, results_table]
)
if __name__ == "__main__":
demo.launch() |