pointcloud_samples=args.pc_samples,
res=args.res,
sample_distribution=args.sample_distribution,
sample_sigmas=args.sample_sigmas,
num_sample_points=50000,
batch_size=args.batch_size,
num_workers=30)
val_dataset = voxelized_data.VoxelizedDataset(
'val',
voxelized_pointcloud=args.pointcloud,
pointcloud_samples=args.pc_samples,
res=args.res,
sample_distribution=args.sample_distribution,
sample_sigmas=args.sample_sigmas,
num_sample_points=50000,
batch_size=args.batch_size,
num_workers=30)
exp_name = 'i{}_dist-{}sigmas-{}v{}_m{}'.format(
'PC' + str(args.pc_samples) if args.pointcloud else 'Voxels',
''.join(str(e) + '_' for e in args.sample_distribution),
''.join(str(e) + '_' for e in args.sample_sigmas), args.res, args.model)
trainer = training.Trainer(net,
torch.device("cuda"),
train_dataset,
val_dataset,
exp_name,
optimizer=args.optimizer)
trainer.train_model(1500)
sample_sigmas=args.sample_sigmas,
num_sample_points=50000,
batch_size=args.batch_size,
num_workers=30)
val_dataset = voxelized_data.VoxelizedDataset(
'val',
voxelized_pointcloud=args.pointcloud,
pointcloud_samples=args.pc_samples,
data_path=args.input_dir_val,
res=args.res,
sample_distribution=args.sample_distribution,
sample_sigmas=args.sample_sigmas,
num_sample_points=50000,
batch_size=args.batch_size,
num_workers=30)
exp_name = 'i{}_dist-{}sigmas-{}v{}_m{}_{}'.format(
'PC' + str(args.pc_samples) if args.pointcloud else 'Voxels',
''.join(str(e) + '_' for e in args.sample_distribution),
''.join(str(e) + '_'
for e in args.sample_sigmas), args.res, args.model, args.name)
trainer = training.Trainer(net,
Device,
train_dataset,
val_dataset,
exp_name,
optimizer=args.optimizer)
trainer.train_model(args.epochs)
import models.local_model as model
import models.dataloader as dataloader
from models import training
import argparse
import torch
import config.config_loader as cfg_loader
parser = argparse.ArgumentParser(description='Train Model')
parser.add_argument('config', type=str, help='Path to config file.')
args = parser.parse_args()
cfg = cfg_loader.load(args.config)
net = model.get_models()[cfg['model']]()
train_dataset = dataloader.VoxelizedDataset('train', cfg)
val_dataset = dataloader.VoxelizedDataset('val', cfg)
trainer = training.Trainer(net,
torch.device("cuda"),
train_dataset,
val_dataset,
cfg['folder_name'],
optimizer=cfg['training']['optimizer'])
trainer.train_model(1500)
pointcloud_samples=cfg.num_points,
data_path=cfg.data_dir,
split_file=cfg.split_file,
batch_size=cfg.batch_size,
num_sample_points=cfg.num_sample_points_training,
num_workers=30,
sample_distribution=cfg.sample_ratio,
sample_sigmas=cfg.sample_std_dev)
val_dataset = voxelized_data.VoxelizedDataset('val',
res=cfg.input_res,
pointcloud_samples=cfg.num_points,
data_path=cfg.data_dir,
split_file=cfg.split_file,
batch_size=cfg.batch_size,
num_sample_points=cfg.num_sample_points_training,
num_workers=30,
sample_distribution=cfg.sample_ratio,
sample_sigmas=cfg.sample_std_dev)
trainer = training.Trainer(net,
torch.device("cuda"),
train_dataset,
val_dataset,
cfg.exp_name,
optimizer=cfg.optimizer,
lr=cfg.lr)
trainer.train_model(cfg.num_epochs)