trainGT.py

from T_trainer import GTTrainer
import pandas as pd
from config import cfg, update_cfg
import time
from utils import init_random_state
# from torch.utils.tensorboard import SummaryWriter

# writer = SummaryWriter()

def run(cfg):
    seeds = [cfg.seed] if cfg.seed is not None else range(cfg.runs)
    TRAINER = GTTrainer

    all_acc = []
    start = time.time()
    for seed in seeds:
        cfg.seed = seed
        init_random_state(seed)
        trainer = TRAINER(cfg)
        trainer.train()
        acc = trainer.eval_and_save()
        all_acc.append(acc)
    end = time.time()

    if len(all_acc) > 1:
        df = pd.DataFrame(all_acc)
        print(f"GraphT: ValACC: {df['val_acc'].mean():.4f} ± {df['val_acc'].std():.4f}, TestAcc: {df['test_acc'].mean():.4f} ± {df['test_acc'].std():.4f}")
    print(f"Running time: {(end-start)/len(seeds):.2f}s")


if __name__ == '__main__':
    cfg = update_cfg(cfg)
    run(cfg)