f8ee76eb82bdffaacde7ab29ba3d8c3669007642,samples/a2c.py,,,#,73

Before Change

    mean_games_bound = run.getint("stop", "mean_games")
    mean_reward_bound = run.getfloat("stop", "mean_reward")

    decimate_params_after_score = run.getint("learning", "decimate_params_after_score", fallback=None)

    try:
        for exp in exp_source:
            batch.append(exp)

After Change

            if rewards and mean_rewards > mean_reward_bound:
                break
            updated_options = run.check_and_reload()
            if updated_options:
                if ("learning", "lr") in updated_options:
                    lr = run.getfloat("learning", "lr")
                    optimizer = optim.RMSprop(model.parameters(), lr=lr)
                    log.info("LR updated to %s", lr)
    finally:
        if args.plot:
            plot_charts(titles=GRAPH_TITLES, data=graph_data, page_name=args.plot)

Instances