def main():
# Get command line arguments
args = parse_arguments()
hyperparameters = {"epochs": args.epochs, "batch_size": args.batch_size}
# Create path for training summaries
summary_path = None
if args.logdir is not None:
summary_path = f"{SUMMARIES_PATH}/{args.logdir}"
os.makedirs(summary_path, exist_ok=True)
# TODO: Add GPU support. This line of code might be helpful.
# device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
print("Summary path:", summary_path)
print("Epochs:", args.epochs)
print("Batch size:", args.batch_size)
# Initalize dataset and model. Then train the model!
train_dataset = StartingDataset()
val_dataset = StartingDataset()
model = StartingNetwork()
starting_train(
train_dataset=train_dataset,
val_dataset=val_dataset,
model=model,
hyperparameters=hyperparameters,
n_eval=args.n_eval,
summary_path=summary_path,
)
def main():
# Get command line arguments
args = parse_arguments()
hyperparameters = {"epochs": args.epochs, "batch_size": args.batch_size}
# Create path for training summaries
label = f"cassava__{int(time.time())}"
summary_path = f"{SUMMARIES_PATH}/{label}"
os.makedirs(summary_path, exist_ok=True)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
print("Summary path:", summary_path)
print("Epochs:", args.epochs)
print("Batch size:", args.batch_size)
# Initalize dataset and model. Then train the model!
train_dataset = TransferTrainDataset()
val_dataset = TransferValidationDataset()
model = StartingNetwork(3, 224, 224, 5)
starting_train(
train_dataset=train_dataset,
val_dataset=val_dataset,
model=model,
hyperparameters=hyperparameters,
n_eval=args.n_eval,
summary_path=summary_path,
)
def main():
images_dir = "./cassava-leaf-disease-classification/train_images"
# Get command line arguments
args = parse_arguments()
hyperparameters = {
"epochs": args.epochs,
"batch_size": args.batch_size,
"weight_decay": args.weight_decay
}
# Create path for training summaries
label = f"cassava__{int(time.time())}"
summary_path = f"{SUMMARIES_PATH}/{label}"
os.makedirs(summary_path, exist_ok=True)
# TODO: Add GPU support. This line of code might be helpful.
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
#print(torch.cuda.get_device_name(0))
print("Summary path:", summary_path)
print("Epochs:", args.epochs)
print("Batch size:", args.batch_size)
# Initalize dataset and model. Then train the model!
train_dataset = StartingDataset(images_dir, 'train.csv')
val_dataset = StartingDataset(images_dir, 'val.csv')
model = ConvNet(num_classes, args.network, args.p_dropout)
#train_dataset.__showitem__(0)
#print(model)
starting_train(train_dataset=train_dataset,
val_dataset=val_dataset,
model=model,
hyperparameters=hyperparameters,
n_eval=args.n_eval,
summary_path=summary_path,
device=device,
usePretrained=args.use_pretrained)
def main():
# Get command line arguments
args = parse_arguments()
hyperparameters = {
"epochs": args.epochs,
"batch_size": args.batch_size,
"weight_decay": args.l2
}
# Create path for training summaries
label = f"cassava__{int(time.time())}"
summary_path = f"{SUMMARIES_PATH}/{label}"
os.makedirs(summary_path, exist_ok=True)
if torch.cuda.is_available():
device = torch.device("cuda:0")
else:
device = torch.device("cpu")
print("Summary path:", summary_path)
print("Epochs:", args.epochs)
print("Batch size:", args.batch_size)
# Initalize dataset and model. Then train the model!
train_dataset = AugmentedDataset(train=True)
val_dataset = AugmentedDataset(train=False)
model = FullResNextModel(5).to(device)
starting_train(
train_dataset=train_dataset,
val_dataset=val_dataset,
model=model,
hyperparameters=hyperparameters,
n_eval=args.n_eval,
summary_path=summary_path,
device=device,
)
torch.save(model.state_dict(),
f'./model-{datetime.now().strftime("%m-%d-%Y_%H-%M-%S")}.pt')