Skip to content

[single file] Detect CLIP max_position_embeddings #11306

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

Closed
wants to merge 1 commit into from
Closed

[single file] Detect CLIP max_position_embeddings #11306

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 src/diffusers/loaders/single_file_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1587,6 +1587,7 @@ def create_diffusers_clip_model_from_ldm(
model = cls(model_config)

position_embedding_dim = model.text_model.embeddings.position_embedding.weight.shape[-1]
position_embedding_size = model.text_model.embeddings.position_embedding.weight.shape[0]

if is_clip_model(checkpoint):
diffusers_format_checkpoint = convert_ldm_clip_checkpoint(checkpoint)
Expand Down Expand Up @@ -1628,6 +1629,13 @@ def create_diffusers_clip_model_from_ldm(
else:
raise ValueError("The provided checkpoint does not seem to contain a valid CLIP model.")

num_position_embeddings = diffusers_format_checkpoint["text_model.embeddings.position_embedding.weight"].shape[0]
if num_position_embeddings != position_embedding_size:
logger.warning(f"Overriding config with detected {num_position_embeddings=}.")
model_config.max_position_embeddings = num_position_embeddings
with ctx():
model = cls(model_config)

if is_accelerate_available():
load_model_dict_into_meta(model, diffusers_format_checkpoint, dtype=torch_dtype)
else:
Expand Down