Skip to content

log #262

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
EthanReid wants to merge 2 commits into
base: main
Choose a base branch
from
Open

log #262

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
36d1b7e
log
EthanReid Feb 26, 2025
d5f1c18
500 to 50 points
EthanReid Mar 27, 2025
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
13 changes: 11 additions & 2 deletions moondream/torch/moondream.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -369,6 +369,7 @@ def _generate_points(
next_token.item() != self.config.tokenizer.eos_id
and len(out) < max_points
):

x_logits = decode_coordinate(hidden, self.region)
x_center = torch.argmax(x_logits, dim=-1) / x_logits.size(-1)
next_emb = encode_coordinate(
Expand All @@ -390,9 +391,17 @@ def _generate_points(
mask[:, :, pos], pos_ids[0] = 1, pos
logits, hidden = self._decode_one_tok(next_emb, mask, pos_ids)
pos += 1

size_logits = decode_size(hidden, self.region)
w = torch.argmax(size_logits[0], dim=-1) / size_logits.size(-1)
h = torch.argmax(size_logits[1], dim=-1) / size_logits.size(-1)

w_bin = torch.argmax(size_logits[0], dim=-1).float()
w_log2 = w_bin / 1023.0 * 10.0 - 10.0
w = 2.0**w_log2

h_bin = torch.argmax(size_logits[1], dim=-1).float()
h_log2 = h_bin / 1023.0 * 10.0 - 10.0
h = 2.0**h_log2

next_emb = encode_size(
torch.tensor(
[w, h], device=self.device, dtype=size_logits.dtype
Expand Down