Add distributed_checkpoint for more efficient checkpoints

This commit is contained in:
James Betker 2020-10-06 20:38:38 -06:00
parent e4b89a172f
commit 4290918359
2 changed files with 52 additions and 0 deletions

View File

@ -105,6 +105,7 @@ if __name__ == "__main__":
screen=True, tofile=True)
logger = logging.getLogger('base')
logger.info(option.dict2str(opt))
util.loaded_options = opt
#### Create test dataset and dataloader
test_loaders = []

View File

@ -0,0 +1,51 @@
import torch
import warnings
def detach_variable(inputs):
if isinstance(inputs, tuple):
out = []
for inp in inputs:
x = inp.detach()
x.requires_grad = inp.requires_grad
out.append(x)
return tuple(out)
else:
raise RuntimeError(
"Only tuple of tensors is supported. Got Unsupported input type: ", type(inputs).__name__)
def check_backward_validity(inputs):
if not any(inp.requires_grad for inp in inputs):
warnings.warn("None of the inputs have requires_grad=True. Gradients will be None")
class CheckpointFunction(torch.autograd.Function):
@staticmethod
def forward(ctx, run_function, length, *args):
ctx.run_function = run_function
ctx.input_tensors = list(args[:length])
ctx.input_params = list(args[length:])
with torch.no_grad():
output_tensors = ctx.run_function(*ctx.input_tensors)
return output_tensors
@staticmethod
def backward(ctx, *output_grads):
for i in range(len(ctx.input_tensors)):
temp = ctx.input_tensors[i]
ctx.input_tensors[i] = temp.detach()
ctx.input_tensors[i].requires_grad = temp.requires_grad
with torch.enable_grad():
output_tensors = ctx.run_function(*ctx.input_tensors)
input_grads = torch.autograd.grad(output_tensors, ctx.input_tensors + ctx.input_params, output_grads, allow_unused=True)
return (None, None) + input_grads
def checkpoint(module, *params):
differentiable_params = tuple(filter(lambda p: p.requires_grad, module.parameters()))
if len(differentiable_params) > 0:
args = params + differentiable_params
return CheckpointFunction.apply(module, len(params), *args)
else:
return module(*params)