Skip to content

Support full-precision embeddings in half precision inference mode #110

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

Merged
lstein merged 2 commits into from
Aug 28, 2022
Merged

Support full-precision embeddings in half precision inference mode #110

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
9b5101c
support full-precision embeddings in half precision mode
sajattack Aug 26, 2022
555f13e
Merge branch 'main' into half-precision-embeddings
sajattack Aug 26, 2022
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: 6 additions & 4 deletions ldm/modules/embedding_manager.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -215,11 +215,13 @@ def save(self, ckpt_path):
ckpt_path,
)

def load(self, ckpt_path):
def load(self, ckpt_path, full=True):
ckpt = torch.load(ckpt_path, map_location='cpu')

self.string_to_token_dict = ckpt['string_to_token']
self.string_to_param_dict = ckpt['string_to_param']
self.string_to_token_dict = ckpt["string_to_token"]
self.string_to_param_dict = ckpt["string_to_param"]
if not full:
for key, value in self.string_to_param_dict.items():
self.string_to_param_dict[key] = torch.nn.Parameter(value.half())

def get_embedding_norms_squared(self):
all_params = torch.cat(
Expand Down
2 changes: 1 addition & 1 deletion ldm/simplet2i.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -488,7 +488,7 @@ def load_model(self):
)
model = self._load_model_from_config(config, self.weights)
if self.embedding_path is not None:
model.embedding_manager.load(self.embedding_path)
model.embedding_manager.load(self.embedding_path, self.full_precision)
self.model = model.to(self.device)
# model.to doesn't change the cond_stage_model.device used to move the tokenizer output, so set it here
self.model.cond_stage_model.device = self.device
Expand Down