train.py

# Copyright 2019-present NAVER Corp.
# CC BY-NC-SA 3.0
# Available only for non-commercial use

import os, pdb
import torch
import torch.optim as optim

from tools import common, trainer
from tools.dataloader import *
from nets.patchnet import *
from nets.losses import *

default_net = "Fast_Quad_L2Net_ConfCFS()" # original Quad_L2Net_ConfCFS()

toy_db_debug = """SyntheticPairDataset(
    ImgFolder('imgs'), 
            'RandomScale(256,1024,can_upscale=True)', 
            'RandomTilting(0.5), PixelNoise(25)')"""

db_web_images = """SyntheticPairDataset(
    web_images, 
        'RandomScale(256,1024,can_upscale=True)',
        'RandomTilting(0.5), PixelNoise(25)')"""

db_aachen_images = """SyntheticPairDataset(
    aachen_db_images, 
        'RandomScale(256,1024,can_upscale=True)', 
        'RandomTilting(0.5), PixelNoise(25)')"""

db_aachen_style_transfer = """TransformedPairs(
    aachen_style_transfer_pairs,
            'RandomScale(256,1024,can_upscale=True), RandomTilting(0.5), PixelNoise(25)')"""

db_aachen_flow = "aachen_flow_pairs"

data_sources = dict(
    D = toy_db_debug,
    W = db_web_images,
    A = db_aachen_images,
    F = db_aachen_flow,
    S = db_aachen_style_transfer,
    )

default_dataloader = """PairLoader(CatPairDataset(`data`),
    scale   = 'RandomScale(256,1024,can_upscale=True)',
    distort = 'ColorJitter(0.2,0.2,0.2,0.1)',
    crop    = 'RandomCrop(192)')"""

default_sampler = """NghSampler2(ngh=7, subq=-8, subd=1, pos_d=3, neg_d=5, border=16,
                            subd_neg=-8,maxpool_pos=True)"""

default_loss = """MultiLoss(
        1, ReliabilityLoss(`sampler`, base=0.5, nq=20),
        1, CosimLoss(N=`N`),
        1, PeakyLoss(N=`N`))"""

default_loss2 = """MultiLoss(
        1, ReliabilityLoss(`sampler`, base=0.5, nq=20),
        1, DicesimLoss(N=`N`),
        1, PeakyLoss(N=`N`))"""

default_loss3 = """MultiLoss(
        1, ReliabilityLoss(`sampler`, base=0.5, nq=20),
        1, JaccardsimLoss(N=`N`),
        1, PeakyLoss(N=`N`))"""

default_loss4 = """MultiLoss(
        1, ReliabilityLoss(`sampler`, base=0.5, nq=20),
        1, CzksimLoss(N=`N`),
        1, PeakyLoss(N=`N`))"""

default_loss5 = """MultiLoss(
        1, ReliabilityLoss(`sampler`, base=0.5, nq=20),
        1, SqcsimLoss(N=`N`),
        1, PeakyLoss(N=`N`))"""

default_loss6 = """MultiLoss(
        1, HardnetLoss(),
        1, SqcsimLoss(N=`N`),
        1, PeakyLoss(N=`N`))"""



class MyTrainer(trainer.Trainer):
    """ This class implements the network training.
        Below is the function I need to overload to explain how to do the backprop.
    """
    def forward_backward(self, inputs):
        output = self.net(imgs=[inputs.pop('img1'),inputs.pop('img2')])
        allvars = dict(inputs, **output)
        loss, details = self.loss_func(**allvars)
        if torch.is_grad_enabled(): loss.backward()
        return loss, details



if __name__ == '__main__':
    import argparse
    parser = argparse.ArgumentParser("Train R2D2")

    parser.add_argument("--data-loader", type=str, default=default_dataloader)
    parser.add_argument("--train-data", type=str, default=list('WASF'), nargs='+', 
        choices = set(data_sources.keys()))
    parser.add_argument("--net", type=str, default=default_net, help='network architecture')

    parser.add_argument("--pretrained", type=str, default="./models/faster2d2_WASF_N16.pt", help='pretrained model path') # ./models/faster2d2_WASF_N16.pt
    parser.add_argument("--save-path", type=str, required=True, help='model save_path path')
    
    parser.add_argument("--loss", type=str, default=default_loss2, help="loss function") # original default_loss
    parser.add_argument("--sampler", type=str, default=default_sampler, help="AP sampler")
    parser.add_argument("--N", type=int, default=16, help="patch size for repeatability")

    parser.add_argument("--epochs", type=int, default=5, help='number of training epochs') # original 25
    parser.add_argument("--batch-size", "--bs", type=int, default=6, help="batch size") # original 8
    parser.add_argument("--learning-rate", "--lr", type=str, default=1e-4)
    parser.add_argument("--weight-decay", "--wd", type=float, default=5e-4)
    
    parser.add_argument("--threads", type=int, default=8, help='number of worker threads')
    parser.add_argument("--gpu", type=int, nargs='+', default=[0], help='-1 for CPU')
    
    args = parser.parse_args()
    
    iscuda = common.torch_set_gpu(args.gpu)
    common.mkdir_for(args.save_path)

    # Create data loader
    from datasets import *
    db = [data_sources[key] for key in args.train_data]
    db = eval(args.data_loader.replace('`data`',','.join(db)).replace('\n',''))
    print("Training image database =", db)
    loader = threaded_loader(db, iscuda, args.threads, args.batch_size, shuffle=True)

    # create network
    print("\n>> Creating net = " + args.net) 
    net = eval(args.net)
    print(f" ( Model size: {common.model_size(net)/1000:.0f}K parameters )")

    # initialization
    if args.pretrained:
        checkpoint = torch.load(args.pretrained, lambda a,b:a)
        net.load_state_dict(checkpoint['state_dict']) # load_pretrained initially
        
    # create losses
    loss = args.loss.replace('`sampler`',args.sampler).replace('`N`',str(args.N))
    print("\n>> Creating loss = " + loss)
    loss = eval(loss.replace('\n',''))
    
    # create optimizer
    optimizer = optim.Adam( [p for p in net.parameters() if p.requires_grad], 
                            lr=args.learning_rate, weight_decay=args.weight_decay)

    train = MyTrainer(net, loader, loss, optimizer)
    if iscuda: train = train.cuda()

    # Training loop #
    for epoch in range(args.epochs):
        print(f"\n>> Starting epoch {epoch}...")
        train()

    print(f"\n>> Saving model to {args.save_path}")
    torch.save({'net': args.net, 'state_dict': net.state_dict()}, args.save_path)