Skip to content

Fix LTX2 prompt relay frame quantization #2018

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
GOvEy1nw wants to merge 1 commit into
base: main
Choose a base branch
from
+5 −5
Open

Fix LTX2 prompt relay frame quantization #2018

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
10 changes: 5 additions & 5 deletions shared/prompt_relay.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -117,7 +117,7 @@ def __call__(self, state: Any, frame_indices: torch.Tensor | None, context: Any)

raw_query_frames = frame_indices.to(torch.float32)
raw_max_frame = raw_query_frames.amax(dim=1, keepdim=True).clamp_min(1.0)
visible_start = raw_max_frame * self.visible_start_ratio
visible_start = torch.round(raw_max_frame * self.visible_start_ratio)
query_frames = raw_query_frames - visible_start
max_frame = (raw_max_frame - visible_start).clamp_min(1.0)
sigma_sq = 2.0 * self.sigma * self.sigma
Expand All @@ -126,11 +126,11 @@ def __call__(self, state: Any, frame_indices: torch.Tensor | None, context: Any)
end_key = min(segment.key_end, positive_len)
if start_key >= end_key:
continue
start = torch.tensor(segment.start, device=device, dtype=torch.float32) * max_frame
end = torch.tensor(segment.end, device=device, dtype=torch.float32) * max_frame
start = torch.round(torch.tensor(segment.start, device=device, dtype=torch.float32) * max_frame)
end = torch.round(torch.tensor(segment.end, device=device, dtype=torch.float32) * max_frame)
length = (end - start).clamp_min(1.0)
midpoint = (start + end) * 0.5
window = (length * 0.5 - 2.0).clamp_min(0.0)
midpoint = torch.floor((start + end) * 0.5)
window = (torch.floor(length * 0.5) - 2.0).clamp_min(0.0)
distance = (query_frames - midpoint).abs()
cost = torch.relu(distance - window).square() / sigma_sq
mask[:, :, start_key:end_key] = -cost.unsqueeze(-1)
Expand Down