Replace hardcoded default values in argparse help strings

Author: miguelsousa

distributed/FSDP/T5_training.py

@@ -198,13 +198,13 @@ def fsdp_main(args):
     # Training settings
     parser = argparse.ArgumentParser(description='PyTorch T5 FSDP Example')
     parser.add_argument('--batch-size', type=int, default=4, metavar='N',
-                        help='input batch size for training (default: 64)')
+                        help='input batch size for training (default: %(default)s)')
     parser.add_argument('--test-batch-size', type=int, default=4, metavar='N',
-                        help='input batch size for testing (default: 1000)')
+                        help='input batch size for testing (default: %(default)s)')
     parser.add_argument('--epochs', type=int, default=2, metavar='N',
-                        help='number of epochs to train (default: 3)')
+                        help='number of epochs to train (default: %(default)s)')
     parser.add_argument('--seed', type=int, default=1, metavar='S',
-                        help='random seed (default: 1)')
+                        help='random seed (default: %(default)s)')
     parser.add_argument('--track_memory', action='store_false', default=True,
                         help='track the gpu memory')
     parser.add_argument('--run_validation', action='store_false', default=True,

distributed/ddp-tutorial-series/multigpu.py

@@ -97,7 +97,8 @@ def main(rank: int, world_size: int, save_every: int, total_epochs: int, batch_s
     parser = argparse.ArgumentParser(description='simple distributed training job')
     parser.add_argument('total_epochs', type=int, help='Total epochs to train the model')
     parser.add_argument('save_every', type=int, help='How often to save a snapshot')
-    parser.add_argument('--batch_size', default=32, type=int, help='Input batch size on each device (default: 32)')
+    parser.add_argument('--batch_size', default=32, type=int,
+                        help='Input batch size on each device (default: %(default)s)')
     args = parser.parse_args()
 
     world_size = torch.cuda.device_count()

distributed/ddp-tutorial-series/multigpu_torchrun.py

@@ -105,7 +105,8 @@ def main(save_every: int, total_epochs: int, batch_size: int, snapshot_path: str
     parser = argparse.ArgumentParser(description='simple distributed training job')
     parser.add_argument('total_epochs', type=int, help='Total epochs to train the model')
     parser.add_argument('save_every', type=int, help='How often to save a snapshot')
-    parser.add_argument('--batch_size', default=32, type=int, help='Input batch size on each device (default: 32)')
+    parser.add_argument('--batch_size', default=32, type=int,
+                        help='Input batch size on each device (default: %(default)s)')
     args = parser.parse_args()
 
     main(args.save_every, args.total_epochs, args.batch_size)

distributed/ddp-tutorial-series/multinode.py

@@ -106,7 +106,8 @@ def main(save_every: int, total_epochs: int, batch_size: int, snapshot_path: str
     parser = argparse.ArgumentParser(description='simple distributed training job')
     parser.add_argument('total_epochs', type=int, help='Total epochs to train the model')
     parser.add_argument('save_every', type=int, help='How often to save a snapshot')
-    parser.add_argument('--batch_size', default=32, type=int, help='Input batch size on each device (default: 32)')
+    parser.add_argument('--batch_size', default=32, type=int,
+                        help='Input batch size on each device (default: %(default)s)')
     args = parser.parse_args()
 
     main(args.save_every, args.total_epochs, args.batch_size)

distributed/ddp-tutorial-series/single_gpu.py

@@ -11,7 +11,7 @@ def __init__(
         train_data: DataLoader,
         optimizer: torch.optim.Optimizer,
         gpu_id: int,
-        save_every: int, 
+        save_every: int,
     ) -> None:
         self.gpu_id = gpu_id
         self.model = model.to(gpu_id)
@@ -75,8 +75,9 @@ def main(device, total_epochs, save_every, batch_size):
     parser = argparse.ArgumentParser(description='simple distributed training job')
     parser.add_argument('total_epochs', type=int, help='Total epochs to train the model')
     parser.add_argument('save_every', type=int, help='How often to save a snapshot')
-    parser.add_argument('--batch_size', default=32, type=int, help='Input batch size on each device (default: 32)')
+    parser.add_argument('--batch_size', default=32, type=int,
+                        help='Input batch size on each device (default: %(default)s)')
     args = parser.parse_args()
-    
+
     device = 0  # shorthand for cuda:0
     main(device, args.total_epochs, args.save_every, args.batch_size)

distributed/rpc/batch/reinforce.py

@@ -21,11 +21,11 @@
 
 parser = argparse.ArgumentParser(description='PyTorch RPC Batch RL example')
 parser.add_argument('--gamma', type=float, default=1.0, metavar='G',
-                    help='discount factor (default: 1.0)')
+                    help='discount factor (default: %(default)s)')
 parser.add_argument('--seed', type=int, default=543, metavar='N',
-                    help='random seed (default: 543)')
+                    help='random seed (default: %(default)s)')
 parser.add_argument('--num-episode', type=int, default=10, metavar='E',
-                    help='number of episodes (default: 10)')
+                    help='number of episodes (default: %(default)s)')
 args = parser.parse_args()
 
 torch.manual_seed(args.seed)

distributed/rpc/rl/main.py

@@ -21,11 +21,11 @@
 parser.add_argument('--world-size', type=int, default=2, metavar='W',
                     help='world size for RPC, rank 0 is the agent, others are observers')
 parser.add_argument('--gamma', type=float, default=0.99, metavar='G',
-                    help='discount factor (default: 0.99)')
+                    help='discount factor (default: %(default)s)')
 parser.add_argument('--seed', type=int, default=543, metavar='N',
-                    help='random seed (default: 543)')
+                    help='random seed (default: %(default)s)')
 parser.add_argument('--log-interval', type=int, default=10, metavar='N',
-                    help='interval between training status logs (default: 10)')
+                    help='interval between training status logs (default: %(default)s)')
 args = parser.parse_args()
 
 torch.manual_seed(args.seed)

gat/main.py

@@ -292,29 +292,29 @@ def test(model, criterion, input, target, mask):
 
     parser = argparse.ArgumentParser(description='PyTorch Graph Attention Network')
     parser.add_argument('--epochs', type=int, default=300,
-                        help='number of epochs to train (default: 300)')
+                        help='number of epochs to train (default: %(default)s)')
     parser.add_argument('--lr', type=float, default=0.005,
-                        help='learning rate (default: 0.005)')
+                        help='learning rate (default: %(default)s)')
     parser.add_argument('--l2', type=float, default=5e-4,
-                        help='weight decay (default: 6e-4)')
+                        help='weight decay (default: %(default)s)')
     parser.add_argument('--dropout-p', type=float, default=0.6,
-                        help='dropout probability (default: 0.6)')
+                        help='dropout probability (default: %(default)s)')
     parser.add_argument('--hidden-dim', type=int, default=64,
-                        help='dimension of the hidden representation (default: 64)')
+                        help='dimension of the hidden representation (default: %(default)s)')
     parser.add_argument('--num-heads', type=int, default=8,
-                        help='number of the attention heads (default: 4)')
+                        help='number of the attention heads (default: %(default)s)')
     parser.add_argument('--concat-heads', action='store_true', default=False,
-                        help='wether to concatinate attention heads, or average over them (default: False)')
+                        help='wether to concatinate attention heads, or average over them (default: %(default)s)')
     parser.add_argument('--val-every', type=int, default=20,
-                        help='epochs to wait for print training and validation evaluation (default: 20)')
+                        help='epochs to wait for print training and validation evaluation (default: %(default)s)')
     parser.add_argument('--no-cuda', action='store_true', default=False,
                         help='disables CUDA training')
     parser.add_argument('--no-mps', action='store_true', default=False,
                         help='disables macOS GPU training')
     parser.add_argument('--dry-run', action='store_true', default=False,
                         help='quickly check a single pass')
     parser.add_argument('--seed', type=int, default=13, metavar='S',
-                        help='random seed (default: 13)')
+                        help='random seed (default: %(default)s)')
     args = parser.parse_args()
 
     torch.manual_seed(args.seed)
@@ -372,4 +372,4 @@ def test(model, criterion, input, target, mask):
         if args.dry_run:
             break
     loss_test, acc_test = test(gat_net, criterion, (features, adj_mat), labels, idx_test)
-    print(f'Test set results: loss {loss_test:.4f} accuracy {acc_test:.4f}')
+    print(f'Test set results: loss {loss_test:.4f} accuracy {acc_test:.4f}')

gcn/main.py

@@ -203,27 +203,27 @@ def test(model, criterion, input, target, mask):
 
     parser = argparse.ArgumentParser(description='PyTorch Graph Convolutional Network')
     parser.add_argument('--epochs', type=int, default=200,
-                        help='number of epochs to train (default: 200)')
+                        help='number of epochs to train (default: %(default)s)')
     parser.add_argument('--lr', type=float, default=0.01,
-                        help='learning rate (default: 0.01)')
+                        help='learning rate (default: %(default)s)')
     parser.add_argument('--l2', type=float, default=5e-4,
-                        help='weight decay (default: 5e-4)')
+                        help='weight decay (default: %(default)s)')
     parser.add_argument('--dropout-p', type=float, default=0.5,
-                        help='dropout probability (default: 0.5)')
+                        help='dropout probability (default: %(default)s)')
     parser.add_argument('--hidden-dim', type=int, default=16,
-                        help='dimension of the hidden representation (default: 16)')
+                        help='dimension of the hidden representation (default: %(default)s)')
     parser.add_argument('--val-every', type=int, default=20,
-                        help='epochs to wait for print training and validation evaluation (default: 20)')
+                        help='epochs to wait for print training and validation evaluation (default: %(default)s)')
     parser.add_argument('--include-bias', action='store_true', default=False,
-                        help='use bias term in convolutions (default: False)')
+                        help='use bias term in convolutions (default: %(default)s)')
     parser.add_argument('--no-cuda', action='store_true', default=False,
                         help='disables CUDA training')
     parser.add_argument('--no-mps', action='store_true', default=False,
                         help='disables macOS GPU training')
     parser.add_argument('--dry-run', action='store_true', default=False,
                         help='quickly check a single pass')
     parser.add_argument('--seed', type=int, default=42, metavar='S',
-                        help='random seed (default: 42)')
+                        help='random seed (default: %(default)s)')
     args = parser.parse_args()
 
     use_cuda = not args.no_cuda and torch.cuda.is_available()
@@ -260,4 +260,4 @@ def test(model, criterion, input, target, mask):
             break
 
     loss_test, acc_test = test(gcn, criterion, (features, adj_mat), labels, idx_test)
-    print(f'Test set results: loss {loss_test:.4f} accuracy {acc_test:.4f}')
+    print(f'Test set results: loss {loss_test:.4f} accuracy {acc_test:.4f}')

imagenet/main.py

@@ -27,34 +27,34 @@
 
 parser = argparse.ArgumentParser(description='PyTorch ImageNet Training')
 parser.add_argument('data', metavar='DIR', nargs='?', default='imagenet',
-                    help='path to dataset (default: imagenet)')
+                    help='path to dataset (default: %(default)s)')
 parser.add_argument('-a', '--arch', metavar='ARCH', default='resnet18',
                     choices=model_names,
                     help='model architecture: ' +
                         ' | '.join(model_names) +
-                        ' (default: resnet18)')
+                        ' (default: %(default)s)')
 parser.add_argument('-j', '--workers', default=4, type=int, metavar='N',
-                    help='number of data loading workers (default: 4)')
+                    help='number of data loading workers (default: %(default)s)')
 parser.add_argument('--epochs', default=90, type=int, metavar='N',
                     help='number of total epochs to run')
 parser.add_argument('--start-epoch', default=0, type=int, metavar='N',
                     help='manual epoch number (useful on restarts)')
 parser.add_argument('-b', '--batch-size', default=256, type=int,
                     metavar='N',
-                    help='mini-batch size (default: 256), this is the total '
+                    help='mini-batch size (default: %(default)s), this is the total '
                          'batch size of all GPUs on the current node when '
                          'using Data Parallel or Distributed Data Parallel')
 parser.add_argument('--lr', '--learning-rate', default=0.1, type=float,
                     metavar='LR', help='initial learning rate', dest='lr')
 parser.add_argument('--momentum', default=0.9, type=float, metavar='M',
                     help='momentum')
 parser.add_argument('--wd', '--weight-decay', default=1e-4, type=float,
-                    metavar='W', help='weight decay (default: 1e-4)',
+                    metavar='W', help='weight decay (default: %(default)s)',
                     dest='weight_decay')
 parser.add_argument('-p', '--print-freq', default=10, type=int,
-                    metavar='N', help='print frequency (default: 10)')
+                    metavar='N', help='print frequency (default: %(default)s)')
 parser.add_argument('--resume', default='', type=str, metavar='PATH',
-                    help='path to latest checkpoint (default: none)')
+                    help='path to latest checkpoint (default: %(default)s)')
 parser.add_argument('-e', '--evaluate', dest='evaluate', action='store_true',
                     help='evaluate model on validation set')
 parser.add_argument('--pretrained', dest='pretrained', action='store_true',

legacy/snli/util.py

@@ -23,18 +23,18 @@ def get_args():
     parser.add_argument('--epochs', type=int, default=50,
                         help='the number of total epochs to run.')
     parser.add_argument('--batch_size', type=int, default=128,
-                        help='batch size. (default: 128)')
+                        help='batch size. (default: %(default)s)')
     parser.add_argument('--d_embed', type=int, default=100,
                         help='the size of each embedding vector.')
     parser.add_argument('--d_proj', type=int, default=300,
                         help='the size of each projection layer.')
     parser.add_argument('--d_hidden', type=int, default=300,
                         help='the number of features in the hidden state.')
     parser.add_argument('--n_layers', type=int, default=1,
-                        help='the number of recurrent layers. (default: 50)')
+                        help='the number of recurrent layers. (default: %(default)s)')
     parser.add_argument('--log_every', type=int, default=50,
                         help='iteration period to output log.')
-    parser.add_argument('--lr',type=float, default=.001,
+    parser.add_argument('--lr', type=float, default=.001,
                         help='initial learning rate.')
     parser.add_argument('--dev_every', type=int, default=1000,
                         help='log period of validation results.')
@@ -51,7 +51,7 @@ def get_args():
     parser.add_argument('--train_embed', action='store_false', dest='fix_emb',
                         help='enable embedding word training.')
     parser.add_argument('--gpu', type=int, default=0,
-                        help='gpu id to use. (default: 0)')
+                        help='gpu id to use. (default: %(default)s)')
     parser.add_argument('--save_path', type=str, default='results',
                         help='save path of results.')
     parser.add_argument('--vector_cache', type=str, default=os.path.join(os.getcwd(), '.vector_cache/input_vectors.pt'),

mnist/main.py

@@ -73,23 +73,23 @@ def main():
     # Training settings
     parser = argparse.ArgumentParser(description='PyTorch MNIST Example')
     parser.add_argument('--batch-size', type=int, default=64, metavar='N',
-                        help='input batch size for training (default: 64)')
+                        help='input batch size for training (default: %(default)s)')
     parser.add_argument('--test-batch-size', type=int, default=1000, metavar='N',
-                        help='input batch size for testing (default: 1000)')
+                        help='input batch size for testing (default: %(default)s)')
     parser.add_argument('--epochs', type=int, default=14, metavar='N',
-                        help='number of epochs to train (default: 14)')
+                        help='number of epochs to train (default: %(default)s)')
     parser.add_argument('--lr', type=float, default=1.0, metavar='LR',
-                        help='learning rate (default: 1.0)')
+                        help='learning rate (default: %(default)s)')
     parser.add_argument('--gamma', type=float, default=0.7, metavar='M',
-                        help='Learning rate step gamma (default: 0.7)')
+                        help='Learning rate step gamma (default: %(default)s)')
     parser.add_argument('--no-cuda', action='store_true', default=False,
                         help='disables CUDA training')
     parser.add_argument('--no-mps', action='store_true', default=False,
                         help='disables macOS GPU training')
     parser.add_argument('--dry-run', action='store_true', default=False,
                         help='quickly check a single pass')
     parser.add_argument('--seed', type=int, default=1, metavar='S',
-                        help='random seed (default: 1)')
+                        help='random seed (default: %(default)s)')
     parser.add_argument('--log-interval', type=int, default=10, metavar='N',
                         help='how many batches to wait before logging training status')
     parser.add_argument('--save-model', action='store_true', default=False,

mnist_forward_forward/main.py

@@ -92,14 +92,14 @@ def train(self, x_pos, x_neg):
         type=int,
         default=1000,
         metavar="N",
-        help="number of epochs to train (default: 1000)",
+        help="number of epochs to train (default: %(default)s)",
     )
     parser.add_argument(
         "--lr",
         type=float,
         default=0.03,
         metavar="LR",
-        help="learning rate (default: 0.03)",
+        help="learning rate (default: %(default)s)",
     )
     parser.add_argument(
         "--no_cuda", action="store_true", default=False, help="disables CUDA training"
@@ -108,7 +108,7 @@ def train(self, x_pos, x_neg):
         "--no_mps", action="store_true", default=False, help="disables MPS training"
     )
     parser.add_argument(
-        "--seed", type=int, default=1, metavar="S", help="random seed (default: 1)"
+        "--seed", type=int, default=1, metavar="S", help="random seed (default: %(default)s)"
     )
     parser.add_argument(
         "--save_model",