Skip to content

Commit b7fa598

Browse files
miguelsousamiguelsousa
committed
Replace hardcoded default values in argparse help strings
1 parent cdef4d4 commit b7fa598

File tree

21 files changed

+91
-87
lines changed

21 files changed

+91
-87
lines changed

Diff for: ‎distributed/FSDP/T5_training.py

+4-4
Original file line numberDiff line numberDiff line change
@@ -198,13 +198,13 @@ def fsdp_main(args):
198198
# Training settings
199199
parser = argparse.ArgumentParser(description='PyTorch T5 FSDP Example')
200200
parser.add_argument('--batch-size', type=int, default=4, metavar='N',
201-
help='input batch size for training (default: 64)')
201+
help='input batch size for training (default: %(default)s)')
202202
parser.add_argument('--test-batch-size', type=int, default=4, metavar='N',
203-
help='input batch size for testing (default: 1000)')
203+
help='input batch size for testing (default: %(default)s)')
204204
parser.add_argument('--epochs', type=int, default=2, metavar='N',
205-
help='number of epochs to train (default: 3)')
205+
help='number of epochs to train (default: %(default)s)')
206206
parser.add_argument('--seed', type=int, default=1, metavar='S',
207-
help='random seed (default: 1)')
207+
help='random seed (default: %(default)s)')
208208
parser.add_argument('--track_memory', action='store_false', default=True,
209209
help='track the gpu memory')
210210
parser.add_argument('--run_validation', action='store_false', default=True,

Diff for: ‎distributed/ddp-tutorial-series/multigpu.py

+2-1
Original file line numberDiff line numberDiff line change
@@ -97,7 +97,8 @@ def main(rank: int, world_size: int, save_every: int, total_epochs: int, batch_s
9797
parser = argparse.ArgumentParser(description='simple distributed training job')
9898
parser.add_argument('total_epochs', type=int, help='Total epochs to train the model')
9999
parser.add_argument('save_every', type=int, help='How often to save a snapshot')
100-
parser.add_argument('--batch_size', default=32, type=int, help='Input batch size on each device (default: 32)')
100+
parser.add_argument('--batch_size', default=32, type=int,
101+
help='Input batch size on each device (default: %(default)s)')
101102
args = parser.parse_args()
102103

103104
world_size = torch.cuda.device_count()

Diff for: ‎distributed/ddp-tutorial-series/multigpu_torchrun.py

+2-1
Original file line numberDiff line numberDiff line change
@@ -105,7 +105,8 @@ def main(save_every: int, total_epochs: int, batch_size: int, snapshot_path: str
105105
parser = argparse.ArgumentParser(description='simple distributed training job')
106106
parser.add_argument('total_epochs', type=int, help='Total epochs to train the model')
107107
parser.add_argument('save_every', type=int, help='How often to save a snapshot')
108-
parser.add_argument('--batch_size', default=32, type=int, help='Input batch size on each device (default: 32)')
108+
parser.add_argument('--batch_size', default=32, type=int,
109+
help='Input batch size on each device (default: %(default)s)')
109110
args = parser.parse_args()
110111

111112
main(args.save_every, args.total_epochs, args.batch_size)

Diff for: ‎distributed/ddp-tutorial-series/multinode.py

+2-1
Original file line numberDiff line numberDiff line change
@@ -106,7 +106,8 @@ def main(save_every: int, total_epochs: int, batch_size: int, snapshot_path: str
106106
parser = argparse.ArgumentParser(description='simple distributed training job')
107107
parser.add_argument('total_epochs', type=int, help='Total epochs to train the model')
108108
parser.add_argument('save_every', type=int, help='How often to save a snapshot')
109-
parser.add_argument('--batch_size', default=32, type=int, help='Input batch size on each device (default: 32)')
109+
parser.add_argument('--batch_size', default=32, type=int,
110+
help='Input batch size on each device (default: %(default)s)')
110111
args = parser.parse_args()
111112

112113
main(args.save_every, args.total_epochs, args.batch_size)

Diff for: ‎distributed/ddp-tutorial-series/single_gpu.py

+4-3
Original file line numberDiff line numberDiff line change
@@ -11,7 +11,7 @@ def __init__(
1111
train_data: DataLoader,
1212
optimizer: torch.optim.Optimizer,
1313
gpu_id: int,
14-
save_every: int,
14+
save_every: int,
1515
) -> None:
1616
self.gpu_id = gpu_id
1717
self.model = model.to(gpu_id)
@@ -75,8 +75,9 @@ def main(device, total_epochs, save_every, batch_size):
7575
parser = argparse.ArgumentParser(description='simple distributed training job')
7676
parser.add_argument('total_epochs', type=int, help='Total epochs to train the model')
7777
parser.add_argument('save_every', type=int, help='How often to save a snapshot')
78-
parser.add_argument('--batch_size', default=32, type=int, help='Input batch size on each device (default: 32)')
78+
parser.add_argument('--batch_size', default=32, type=int,
79+
help='Input batch size on each device (default: %(default)s)')
7980
args = parser.parse_args()
80-
81+
8182
device = 0 # shorthand for cuda:0
8283
main(device, args.total_epochs, args.save_every, args.batch_size)

Diff for: ‎distributed/rpc/batch/reinforce.py

+3-3
Original file line numberDiff line numberDiff line change
@@ -21,11 +21,11 @@
2121

2222
parser = argparse.ArgumentParser(description='PyTorch RPC Batch RL example')
2323
parser.add_argument('--gamma', type=float, default=1.0, metavar='G',
24-
help='discount factor (default: 1.0)')
24+
help='discount factor (default: %(default)s)')
2525
parser.add_argument('--seed', type=int, default=543, metavar='N',
26-
help='random seed (default: 543)')
26+
help='random seed (default: %(default)s)')
2727
parser.add_argument('--num-episode', type=int, default=10, metavar='E',
28-
help='number of episodes (default: 10)')
28+
help='number of episodes (default: %(default)s)')
2929
args = parser.parse_args()
3030

3131
torch.manual_seed(args.seed)

Diff for: ‎distributed/rpc/rl/main.py

+3-3
Original file line numberDiff line numberDiff line change
@@ -21,11 +21,11 @@
2121
parser.add_argument('--world-size', type=int, default=2, metavar='W',
2222
help='world size for RPC, rank 0 is the agent, others are observers')
2323
parser.add_argument('--gamma', type=float, default=0.99, metavar='G',
24-
help='discount factor (default: 0.99)')
24+
help='discount factor (default: %(default)s)')
2525
parser.add_argument('--seed', type=int, default=543, metavar='N',
26-
help='random seed (default: 543)')
26+
help='random seed (default: %(default)s)')
2727
parser.add_argument('--log-interval', type=int, default=10, metavar='N',
28-
help='interval between training status logs (default: 10)')
28+
help='interval between training status logs (default: %(default)s)')
2929
args = parser.parse_args()
3030

3131
torch.manual_seed(args.seed)

Diff for: ‎gat/main.py

+10-10
Original file line numberDiff line numberDiff line change
@@ -292,29 +292,29 @@ def test(model, criterion, input, target, mask):
292292

293293
parser = argparse.ArgumentParser(description='PyTorch Graph Attention Network')
294294
parser.add_argument('--epochs', type=int, default=300,
295-
help='number of epochs to train (default: 300)')
295+
help='number of epochs to train (default: %(default)s)')
296296
parser.add_argument('--lr', type=float, default=0.005,
297-
help='learning rate (default: 0.005)')
297+
help='learning rate (default: %(default)s)')
298298
parser.add_argument('--l2', type=float, default=5e-4,
299-
help='weight decay (default: 6e-4)')
299+
help='weight decay (default: %(default)s)')
300300
parser.add_argument('--dropout-p', type=float, default=0.6,
301-
help='dropout probability (default: 0.6)')
301+
help='dropout probability (default: %(default)s)')
302302
parser.add_argument('--hidden-dim', type=int, default=64,
303-
help='dimension of the hidden representation (default: 64)')
303+
help='dimension of the hidden representation (default: %(default)s)')
304304
parser.add_argument('--num-heads', type=int, default=8,
305-
help='number of the attention heads (default: 4)')
305+
help='number of the attention heads (default: %(default)s)')
306306
parser.add_argument('--concat-heads', action='store_true', default=False,
307-
help='wether to concatinate attention heads, or average over them (default: False)')
307+
help='wether to concatinate attention heads, or average over them (default: %(default)s)')
308308
parser.add_argument('--val-every', type=int, default=20,
309-
help='epochs to wait for print training and validation evaluation (default: 20)')
309+
help='epochs to wait for print training and validation evaluation (default: %(default)s)')
310310
parser.add_argument('--no-cuda', action='store_true', default=False,
311311
help='disables CUDA training')
312312
parser.add_argument('--no-mps', action='store_true', default=False,
313313
help='disables macOS GPU training')
314314
parser.add_argument('--dry-run', action='store_true', default=False,
315315
help='quickly check a single pass')
316316
parser.add_argument('--seed', type=int, default=13, metavar='S',
317-
help='random seed (default: 13)')
317+
help='random seed (default: %(default)s)')
318318
args = parser.parse_args()
319319

320320
torch.manual_seed(args.seed)
@@ -372,4 +372,4 @@ def test(model, criterion, input, target, mask):
372372
if args.dry_run:
373373
break
374374
loss_test, acc_test = test(gat_net, criterion, (features, adj_mat), labels, idx_test)
375-
print(f'Test set results: loss {loss_test:.4f} accuracy {acc_test:.4f}')
375+
print(f'Test set results: loss {loss_test:.4f} accuracy {acc_test:.4f}')

Diff for: ‎gcn/main.py

+9-9
Original file line numberDiff line numberDiff line change
@@ -203,27 +203,27 @@ def test(model, criterion, input, target, mask):
203203

204204
parser = argparse.ArgumentParser(description='PyTorch Graph Convolutional Network')
205205
parser.add_argument('--epochs', type=int, default=200,
206-
help='number of epochs to train (default: 200)')
206+
help='number of epochs to train (default: %(default)s)')
207207
parser.add_argument('--lr', type=float, default=0.01,
208-
help='learning rate (default: 0.01)')
208+
help='learning rate (default: %(default)s)')
209209
parser.add_argument('--l2', type=float, default=5e-4,
210-
help='weight decay (default: 5e-4)')
210+
help='weight decay (default: %(default)s)')
211211
parser.add_argument('--dropout-p', type=float, default=0.5,
212-
help='dropout probability (default: 0.5)')
212+
help='dropout probability (default: %(default)s)')
213213
parser.add_argument('--hidden-dim', type=int, default=16,
214-
help='dimension of the hidden representation (default: 16)')
214+
help='dimension of the hidden representation (default: %(default)s)')
215215
parser.add_argument('--val-every', type=int, default=20,
216-
help='epochs to wait for print training and validation evaluation (default: 20)')
216+
help='epochs to wait for print training and validation evaluation (default: %(default)s)')
217217
parser.add_argument('--include-bias', action='store_true', default=False,
218-
help='use bias term in convolutions (default: False)')
218+
help='use bias term in convolutions (default: %(default)s)')
219219
parser.add_argument('--no-cuda', action='store_true', default=False,
220220
help='disables CUDA training')
221221
parser.add_argument('--no-mps', action='store_true', default=False,
222222
help='disables macOS GPU training')
223223
parser.add_argument('--dry-run', action='store_true', default=False,
224224
help='quickly check a single pass')
225225
parser.add_argument('--seed', type=int, default=42, metavar='S',
226-
help='random seed (default: 42)')
226+
help='random seed (default: %(default)s)')
227227
args = parser.parse_args()
228228

229229
use_cuda = not args.no_cuda and torch.cuda.is_available()
@@ -260,4 +260,4 @@ def test(model, criterion, input, target, mask):
260260
break
261261

262262
loss_test, acc_test = test(gcn, criterion, (features, adj_mat), labels, idx_test)
263-
print(f'Test set results: loss {loss_test:.4f} accuracy {acc_test:.4f}')
263+
print(f'Test set results: loss {loss_test:.4f} accuracy {acc_test:.4f}')

Diff for: ‎imagenet/main.py

+7-7
Original file line numberDiff line numberDiff line change
@@ -27,34 +27,34 @@
2727

2828
parser = argparse.ArgumentParser(description='PyTorch ImageNet Training')
2929
parser.add_argument('data', metavar='DIR', nargs='?', default='imagenet',
30-
help='path to dataset (default: imagenet)')
30+
help='path to dataset (default: %(default)s)')
3131
parser.add_argument('-a', '--arch', metavar='ARCH', default='resnet18',
3232
choices=model_names,
3333
help='model architecture: ' +
3434
' | '.join(model_names) +
35-
' (default: resnet18)')
35+
' (default: %(default)s)')
3636
parser.add_argument('-j', '--workers', default=4, type=int, metavar='N',
37-
help='number of data loading workers (default: 4)')
37+
help='number of data loading workers (default: %(default)s)')
3838
parser.add_argument('--epochs', default=90, type=int, metavar='N',
3939
help='number of total epochs to run')
4040
parser.add_argument('--start-epoch', default=0, type=int, metavar='N',
4141
help='manual epoch number (useful on restarts)')
4242
parser.add_argument('-b', '--batch-size', default=256, type=int,
4343
metavar='N',
44-
help='mini-batch size (default: 256), this is the total '
44+
help='mini-batch size (default: %(default)s), this is the total '
4545
'batch size of all GPUs on the current node when '
4646
'using Data Parallel or Distributed Data Parallel')
4747
parser.add_argument('--lr', '--learning-rate', default=0.1, type=float,
4848
metavar='LR', help='initial learning rate', dest='lr')
4949
parser.add_argument('--momentum', default=0.9, type=float, metavar='M',
5050
help='momentum')
5151
parser.add_argument('--wd', '--weight-decay', default=1e-4, type=float,
52-
metavar='W', help='weight decay (default: 1e-4)',
52+
metavar='W', help='weight decay (default: %(default)s)',
5353
dest='weight_decay')
5454
parser.add_argument('-p', '--print-freq', default=10, type=int,
55-
metavar='N', help='print frequency (default: 10)')
55+
metavar='N', help='print frequency (default: %(default)s)')
5656
parser.add_argument('--resume', default='', type=str, metavar='PATH',
57-
help='path to latest checkpoint (default: none)')
57+
help='path to latest checkpoint (default: %(default)s)')
5858
parser.add_argument('-e', '--evaluate', dest='evaluate', action='store_true',
5959
help='evaluate model on validation set')
6060
parser.add_argument('--pretrained', dest='pretrained', action='store_true',

Diff for: ‎legacy/snli/util.py

+4-4
Original file line numberDiff line numberDiff line change
@@ -23,18 +23,18 @@ def get_args():
2323
parser.add_argument('--epochs', type=int, default=50,
2424
help='the number of total epochs to run.')
2525
parser.add_argument('--batch_size', type=int, default=128,
26-
help='batch size. (default: 128)')
26+
help='batch size. (default: %(default)s)')
2727
parser.add_argument('--d_embed', type=int, default=100,
2828
help='the size of each embedding vector.')
2929
parser.add_argument('--d_proj', type=int, default=300,
3030
help='the size of each projection layer.')
3131
parser.add_argument('--d_hidden', type=int, default=300,
3232
help='the number of features in the hidden state.')
3333
parser.add_argument('--n_layers', type=int, default=1,
34-
help='the number of recurrent layers. (default: 50)')
34+
help='the number of recurrent layers. (default: %(default)s)')
3535
parser.add_argument('--log_every', type=int, default=50,
3636
help='iteration period to output log.')
37-
parser.add_argument('--lr',type=float, default=.001,
37+
parser.add_argument('--lr', type=float, default=.001,
3838
help='initial learning rate.')
3939
parser.add_argument('--dev_every', type=int, default=1000,
4040
help='log period of validation results.')
@@ -51,7 +51,7 @@ def get_args():
5151
parser.add_argument('--train_embed', action='store_false', dest='fix_emb',
5252
help='enable embedding word training.')
5353
parser.add_argument('--gpu', type=int, default=0,
54-
help='gpu id to use. (default: 0)')
54+
help='gpu id to use. (default: %(default)s)')
5555
parser.add_argument('--save_path', type=str, default='results',
5656
help='save path of results.')
5757
parser.add_argument('--vector_cache', type=str, default=os.path.join(os.getcwd(), '.vector_cache/input_vectors.pt'),

Diff for: ‎mnist/main.py

+6-6
Original file line numberDiff line numberDiff line change
@@ -73,23 +73,23 @@ def main():
7373
# Training settings
7474
parser = argparse.ArgumentParser(description='PyTorch MNIST Example')
7575
parser.add_argument('--batch-size', type=int, default=64, metavar='N',
76-
help='input batch size for training (default: 64)')
76+
help='input batch size for training (default: %(default)s)')
7777
parser.add_argument('--test-batch-size', type=int, default=1000, metavar='N',
78-
help='input batch size for testing (default: 1000)')
78+
help='input batch size for testing (default: %(default)s)')
7979
parser.add_argument('--epochs', type=int, default=14, metavar='N',
80-
help='number of epochs to train (default: 14)')
80+
help='number of epochs to train (default: %(default)s)')
8181
parser.add_argument('--lr', type=float, default=1.0, metavar='LR',
82-
help='learning rate (default: 1.0)')
82+
help='learning rate (default: %(default)s)')
8383
parser.add_argument('--gamma', type=float, default=0.7, metavar='M',
84-
help='Learning rate step gamma (default: 0.7)')
84+
help='Learning rate step gamma (default: %(default)s)')
8585
parser.add_argument('--no-cuda', action='store_true', default=False,
8686
help='disables CUDA training')
8787
parser.add_argument('--no-mps', action='store_true', default=False,
8888
help='disables macOS GPU training')
8989
parser.add_argument('--dry-run', action='store_true', default=False,
9090
help='quickly check a single pass')
9191
parser.add_argument('--seed', type=int, default=1, metavar='S',
92-
help='random seed (default: 1)')
92+
help='random seed (default: %(default)s)')
9393
parser.add_argument('--log-interval', type=int, default=10, metavar='N',
9494
help='how many batches to wait before logging training status')
9595
parser.add_argument('--save-model', action='store_true', default=False,

Diff for: ‎mnist_forward_forward/main.py

+3-3
Original file line numberDiff line numberDiff line change
@@ -92,14 +92,14 @@ def train(self, x_pos, x_neg):
9292
type=int,
9393
default=1000,
9494
metavar="N",
95-
help="number of epochs to train (default: 1000)",
95+
help="number of epochs to train (default: %(default)s)",
9696
)
9797
parser.add_argument(
9898
"--lr",
9999
type=float,
100100
default=0.03,
101101
metavar="LR",
102-
help="learning rate (default: 0.03)",
102+
help="learning rate (default: %(default)s)",
103103
)
104104
parser.add_argument(
105105
"--no_cuda", action="store_true", default=False, help="disables CUDA training"
@@ -108,7 +108,7 @@ def train(self, x_pos, x_neg):
108108
"--no_mps", action="store_true", default=False, help="disables MPS training"
109109
)
110110
parser.add_argument(
111-
"--seed", type=int, default=1, metavar="S", help="random seed (default: 1)"
111+
"--seed", type=int, default=1, metavar="S", help="random seed (default: %(default)s)"
112112
)
113113
parser.add_argument(
114114
"--save_model",

0 commit comments

Comments
 (0)