train.py

import torch
import torch.optim as optim
from tensorboardX import SummaryWriter
import numpy as np
import os
import sys, importlib
import argparse
import time, datetime
import matplotlib; matplotlib.use('Agg')
from src import config, data
from src.checkpoints import CheckpointIO
from collections import defaultdict
import shutil


# Arguments
parser = argparse.ArgumentParser(
    description='Train Implicit Autoencoder.'
)
parser.add_argument('config', type=str, help='Path to config file.')

args = parser.parse_args()
cfg = config.load_config(args.config, 'configs/default.yaml')
is_cuda = torch.cuda.is_available()
device = torch.device("cuda" if is_cuda else "cpu")
# Set t0
t0 = time.time()

# Shorthands
out_dir = cfg['training']['out_dir']
batch_size = cfg['training']['batch_size']
backup_every = cfg['training']['backup_every']

model_selection_metric = cfg['training']['model_selection_metric']
if cfg['training']['model_selection_mode'] == 'maximize':
    model_selection_sign = 1
elif cfg['training']['model_selection_mode'] == 'minimize':
    model_selection_sign = -1
else:
    raise ValueError('model_selection_mode must be '
                     'either maximize or minimize.')

# Output directory
if not os.path.exists(out_dir):
    os.makedirs(out_dir)

shutil.copyfile(args.config, os.path.join(out_dir, 'config.yaml'))

# Dataset
train_dataset = config.get_dataset('train', cfg)
val_dataset = config.get_dataset('val', cfg, return_idx=True)

train_loader = torch.utils.data.DataLoader(
    train_dataset, batch_size=batch_size, num_workers=cfg['training']['n_workers'], shuffle=True,
    collate_fn=data.collate_remove_none,
    worker_init_fn=data.worker_init_fn)

val_loader = torch.utils.data.DataLoader(
        val_dataset, batch_size=1, num_workers=cfg['training']['n_workers_val'], shuffle=False,
    collate_fn=data.collate_remove_none,
    worker_init_fn=data.worker_init_fn)

print('Loaded data.')

# Model
model = config.get_model(cfg, device=device, dataset=train_dataset)

# Intialize training
optimizer = optim.Adam(model.parameters(), lr=1e-4)
# optimizer = optim.SGD(model.parameters(), lr=1e-4, momentum=0.9)

trainer = config.get_trainer(model, optimizer, cfg, device=device)
checkpoint_io = CheckpointIO(out_dir, model=model, optimizer=optimizer)
try:
    if cfg['training']['model_file']:
        load_dict = checkpoint_io.load(cfg['training']['model_file'])
    else:
        load_dict = checkpoint_io.load('model.pt')
except FileExistsError:
    print('No file load...')
    load_dict = dict()

epoch_it = load_dict.get('epoch_it', 0)
it = load_dict.get('it', 0)
metric_val_best = load_dict.get(
    'loss_val_best', -model_selection_sign * np.inf)

if metric_val_best == np.inf or metric_val_best == -np.inf:
    metric_val_best = -model_selection_sign * np.inf
print('Current best validation metric (%s): %.8f'
      % (model_selection_metric, metric_val_best))
logger = SummaryWriter(os.path.join(out_dir, 'logs'))

# Shorthands
print_every = cfg['training']['print_every']
checkpoint_every = cfg['training']['checkpoint_every']
validate_every = cfg['training']['validate_every']

print('output path: ', cfg['training']['out_dir'])

while True:
    epoch_it += 1
    for batch in train_loader:
        it += 1
        loss = trainer.train_step(batch)
        logger.add_scalar('train/loss', loss, it)

        # Print output
        if print_every > 0 and (it % print_every) == 0:
            t = datetime.datetime.now()
            print('[Epoch %02d] it=%03d, loss=%.4f, time: %.2fs, %02d:%02d'
                     % (epoch_it, it, loss, time.time() - t0, t.hour, t.minute))

        # Save checkpoint
        if (checkpoint_every > 0 and (it % checkpoint_every) == 0):
            print('Saving checkpoint')
            checkpoint_io.save('model.pt', epoch_it=epoch_it, it=it,
                               loss_val_best=metric_val_best)
        
        # Backup if necessary
        if (backup_every > 0 and (it % backup_every) == 0):
            print('Backup checkpoint')
            checkpoint_io.save('model_%d.pt' % it, epoch_it=epoch_it, it=it,
                               loss_val_best=metric_val_best)
        # Run validation
        if validate_every > 0 and (it % validate_every) == 0:
        #if 1:
            eval_dict = trainer.evaluate(val_loader)
            metric_val = eval_dict[model_selection_metric]
            print('Validation metric (%s): %.4f'
                  % (model_selection_metric, metric_val))

            for k, v in eval_dict.items():
                logger.add_scalar('val/%s' % k, v, it)

            if model_selection_sign * (metric_val - metric_val_best) > 0:
                metric_val_best = metric_val
                print('New best model (loss %.4f)' % metric_val_best)
                checkpoint_io.save('model_best.pt', epoch_it=epoch_it, it=it,
                                   loss_val_best=metric_val_best)
            #import ipdb; ipdb.set_trace()