evaluation.py

import time
from image_generation import generate_plot
from tasks import test_tasks

def test_for_metrics(temperature, iter=3):
	session_render = []
	session_attempts = []
	session_times = []
	all_tasks_times = []

	for i in range(iter):
		print(f"Session: {i + 1}")

		cnt = 0
		results = []
		times = []

		for task in test_tasks:
			cnt += 1

			print(f"counter: {cnt} -- task: {task}")
   
			start = time.perf_counter()
			result = generate_plot(task, temperature)
			end = time.perf_counter() - start

			results.append(result)
			times.append(end)

		all_tasks_times.append(times)
		session_render.append(sum(r["success"] for r in results) / len(results))
		session_attempts.append(sum(r["attempts"] for r in results if r["success"]) / sum(r["success"] for r in results))
		succ_times = [t for t,r in zip(times, results) if r["success"]]
		session_times.append(sum(succ_times) / len(succ_times))

	return session_render, session_attempts, session_times, all_tasks_times