main.train(params, buffer2, device, frame_idx2, exp_source2, reward_tracker2, optimizer2, net2, tgt_net2, writer2) while True: if frame // args.units % 2 == 0: frame_idx1 += 1 if main.train(params, buffer1, device, frame_idx1, exp_source1, reward_tracker1, optimizer1, net1, tgt_net1, writer1): break else: frame_idx2 += 1 if main.train(params, buffer2, device, frame_idx2, exp_source2, reward_tracker2, optimizer2, net2, tgt_net2, writer2): break frame += 1 if args.maxFrames > 0 and frame_idx1 > args.maxFrames: break if __name__ == "__main__": params = common.HYPERPARAMS['bfw'] args = utils.parse_args() device = torch.device("cuda" if args.cuda else "cpu") while True: execute(args, params, device)
import pandas as pd from lib.utils import parse_args, seconds_to_time TOTAL_TIME = 6000 def gen_prediction_every_n_const(files, n, event): y_pred = [] for file in files: for time in range(n // 2, TOTAL_TIME, n): y = { 'file_name': file, 'event_type': event, 'event_time': seconds_to_time(time) } y_pred.append(y) return pd.DataFrame(y_pred) if __name__ == '__main__': args = parse_args() output = args['output'] predict_table_path = args['predict_table'] predict_table = pd.read_csv(predict_table_path) files = predict_table['file_name'].unique() e_1 = gen_prediction_every_n_const(files, 120, 'удар по воротам') e_2 = gen_prediction_every_n_const(files, 120, 'угловой') concat = pd.concat([predict_table, e_1, e_2], ignore_index=True) concat = concat[['file_name', 'event_type', 'event_time']] concat.to_csv(output)