Spaces:
Sleeping
Sleeping
| import os | |
| import sys | |
| import time | |
| import numpy as np | |
| sys.path.append(os.getcwd()) | |
| from ai.gym_env import LoveLiveCardGameEnv | |
| def benchmark_oo_steps(): | |
| print("Benchmarking OO Engine (LoveLiveCardGameEnv)...") | |
| env = LoveLiveCardGameEnv() | |
| env.reset() | |
| num_steps = 1000 | |
| start = time.time() | |
| for _ in range(num_steps): | |
| # Sample legal action | |
| masks = env.action_masks() | |
| legal_indices = np.where(masks)[0] | |
| if len(legal_indices) == 0: | |
| env.reset() | |
| continue | |
| action = np.random.choice(legal_indices) | |
| obs, reward, terminated, truncated, info = env.step(action) | |
| if terminated or truncated: | |
| env.reset() | |
| end = time.time() | |
| duration = end - start | |
| throughput = num_steps / duration | |
| print(f"Throughput: {throughput:.2f} steps/sec") | |
| if __name__ == "__main__": | |
| benchmark_oo_steps() | |