Skip to content

Add _apply_decay_mask_logits function to avoid numerical instability #4

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
boxin-wbx wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Add _apply_decay_mask_logits function to avoid numerical instability #4

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 8 additions & 0 deletions generation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -57,6 +57,14 @@ def _apply_decay_mask(self, logits: torch.Tensor, decay_mask: torch.Tensor) -> t
probabilities = probabilities / probabilities.sum(dim=-1)
return probabilities

def _apply_decay_mask_logits(self, logits: torch.Tensor, decay_mask: torch.Tensor) -> torch.Tensor:
"""Applies decay to a tensor of logits in the log space"""
decay_mask = torch.exp(- decay_mask * self.decay_constant)
decay_mask = torch.max(decay_mask, torch.tensor([self.epsilon], device=decay_mask.device))
log_decay_mask = torch.log(decay_mask)
logits += log_decay_mask
return logits

def _generate_decay_mask(self, logits_regular: torch.FloatTensor, logits_biased_list: List[torch.FloatTensor]) -> torch.Tensor:
"""Computes the alpha values (see paper) for each token and stores them in a mask tensor"""
p_regular = logits_regular.softmax(dim=-1)
Expand Down