Skip to content
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

Jump to bottom

style(eltociear): fix typo in contrastive_loss.py #572

Merged
merged 1 commit into from
Jan 28, 2023
Merged

style(eltociear): fix typo in contrastive_loss.py #572

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
4 changes: 2 additions & 2 deletions ding/torch_utils/loss/contrastive_loss.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,7 +79,7 @@ def forward(self, x: torch.Tensor, y: torch.Tensor):
x_n = x.view(-1, self._encode_shape)
y_n = y.view(-1, self._encode_shape)

# Use inner product to obtain postive samples.
# Use inner product to obtain positive samples.
# [N, x_heads, encode_dim] * [N, encode_dim, y_heads] -> [N, x_heads, y_heads]
u_pos = torch.matmul(x, y.permute(0, 2, 1)).unsqueeze(2)
# Use outer product to obtain all sample permutations.
Expand All @@ -92,7 +92,7 @@ def forward(self, x: torch.Tensor, y: torch.Tensor):
u_neg = (n_mask * u_all) - (10. * (1 - n_mask))
u_neg = u_neg.view(N, N * x_heads, y_heads).unsqueeze(dim=1).expand(-1, x_heads, -1, -1)

# Concatenate postive and negative samples and apply log softmax.
# Concatenate positive and negative samples and apply log softmax.
pred_lgt = torch.cat([u_pos, u_neg], dim=2)
pred_log = F.log_softmax(pred_lgt * self._temperature, dim=2)

Expand Down