Commit 9a09107c authored by Thor Johnsen's avatar Thor Johnsen
Browse files

Bug fix

parent 1ac28972
...@@ -84,7 +84,7 @@ class DistributedFusedLAMB(torch.optim.Optimizer): ...@@ -84,7 +84,7 @@ class DistributedFusedLAMB(torch.optim.Optimizer):
defaults = dict(lr=lr, bias_correction=bias_correction, defaults = dict(lr=lr, bias_correction=bias_correction,
betas=betas, eps=eps, weight_decay=weight_decay, betas=betas, eps=eps, weight_decay=weight_decay,
max_grad_norm=max_grad_norm) max_grad_norm=max_grad_norm)
super(DistributedFusedAdam, self).__init__(params, defaults) super(DistributedFusedLAMB, self).__init__(params, defaults)
self.eps_mode = 0 if eps_inside_sqrt else 1 self.eps_mode = 0 if eps_inside_sqrt else 1
self._overflow_buf = torch.cuda.IntTensor([0]) self._overflow_buf = torch.cuda.IntTensor([0])
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment