[wip] Replace optimizers in torch.optim with the ones from torch.optim._multi_tensor#49039
Closed
izdeby wants to merge 52 commits intogh/izdeby/69/basefrom
Closed
[wip] Replace optimizers in torch.optim with the ones from torch.optim._multi_tensor#49039izdeby wants to merge 52 commits intogh/izdeby/69/basefrom
izdeby wants to merge 52 commits intogh/izdeby/69/basefrom
Conversation
[ghstack-poisoned]
This was referenced Dec 8, 2020
Closed
💊 CI failures summary and remediationsAs of commit f456748 (more details on the Dr. CI page):
🕵️ 16 new failures recognized by patternsThe following CI failures do not appear to be due to upstream breakages:
|
| Job | Step | Action |
|---|---|---|
| Ensure correct trailing newlines | 🔁 rerun |
ci.pytorch.org: 1 failed
This comment was automatically generated by Dr. CI (expand for details).
Follow this link to opt-out of these comments for your Pull Requests.Please report bugs/suggestions to the (internal) Dr. CI Users group.
….optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
….optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
….optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
….optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
This was referenced Dec 11, 2020
….optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
….optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
….optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
This was referenced Feb 8, 2021
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
This was referenced Feb 18, 2021
Closed
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
… torch.optim._multi_tensor" Differential Revision: [D25406490](https://our.internmc.facebook.com/intern/diff/D25406490) ------ ### Benchmark results SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True) Current: 201.63 ms Foreach: 56.99 ms Adam (weight_decay=1., amsgrad=True) Current: 233.27 ms Foreach: 46.89 ms AdamW (weight_decay=1., amsgrad=True) Current: 371.18 ms Foreach: 121.04 ms RMSprop (weight_decay=1, momentum=1, centered=True) Current: 364.88 ms Foreach: 47.52 ms Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50)) Current: 1.43 s Foreach: 1.26 s ASGD (weight_decay=1) Current: 165.39 ms Foreach: 40.61 ms Adamax (weight_decay=1) Current: 374.42 ms Foreach: 291.06 ms Adadelta (weight_decay=1) Current: 252.64 ms Foreach: 29.62 ms ### Benchmark script ``` import torch import torch.optim as optim import torch.nn as nn import torchvision import torch.utils.benchmark as benchmark_utils model = torchvision.models.resnet.resnet101(pretrained=True).to("cuda") targets = torch.randint(0, 1000, (100, 100), device="cuda") criterion = nn.CrossEntropyLoss() # optimizers params = dict(weight_decay=1) optimizer = optim.Adadelta(model.parameters(), **params) optimizer_mta = optim._multi_tensor.Adadelta(model.parameters(), **params) running_loss = 0.0 target = torch.empty(128, dtype=torch.long, device="cuda").random_(5) optimizer.zero_grad() inputs = torch.rand(128, 3, 100, 100, device="cuda" , requires_grad=True) outputs = model(inputs) loss = criterion(outputs, target) loss.backward() optimizer.step() running_loss += loss.item() def main(): timer = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer.step()", globals=globals(), label="str(optimizer)", ) print(f"autorange:\n{timer.blocked_autorange()}\n\n") timer_mta = benchmark_utils.Timer( stmt="torch.cuda.synchronize(); optimizer_mta.step()", globals=globals(), label="str(optimizer_mta)", ) print(f"autorange:\n{timer_mta.blocked_autorange()}\n\n") if __name__ == "__main__": main() ``` [ghstack-poisoned]
Contributor
|
Hi @izdeby! Thank you for your pull request. We require contributors to sign our Contributor License Agreement, and yours needs attention. You currently have a record in our system, but the CLA is no longer valid, and will need to be resubmitted. ProcessIn order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA. Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with If you have received this in error or have any questions, please contact us at cla@fb.com. Thanks! |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Stack from ghstack:
Differential Revision: D25406490
Benchmark results
SGD (lr=1e-3, momentum=1, dampening=0, weight_decay=1, nesterov=True)
Current: 201.63 ms
Foreach: 56.99 ms
Adam (weight_decay=1., amsgrad=True)
Current: 233.27 ms
Foreach: 46.89 ms
AdamW (weight_decay=1., amsgrad=True)
Current: 371.18 ms
Foreach: 121.04 ms
RMSprop (weight_decay=1, momentum=1, centered=True)
Current: 364.88 ms
Foreach: 47.52 ms
Rprop (lr=1e-2, etas=(0.5, 1.2), step_sizes=(1e-6, 50))
Current: 1.43 s
Foreach: 1.26 s
ASGD (weight_decay=1)
Current: 165.39 ms
Foreach: 40.61 ms
Adamax (weight_decay=1)
Current: 374.42 ms
Foreach: 291.06 ms
Adadelta (weight_decay=1)
Current: 252.64 ms
Foreach: 29.62 ms
Benchmark script