Skip to content

feat: (fix?) try known prefixes to find correspondance for applying imatrix #1737

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
stduhpf wants to merge 1 commit into
base: master
Choose a base branch
from
Open

feat: (fix?) try known prefixes to find correspondance for applying imatrix #1737

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
27 changes: 27 additions & 0 deletions src/model_loader.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1165,6 +1165,33 @@ bool ModelLoader::load_tensors(on_new_tensor_cb_t on_new_tensor_cb,
}
std::string processed_name = convert_tensor_name(tensor_storage.name, imatrix_version);
std::vector<float> imatrix = get_imatrix_collector().get_values(processed_name);
if (imatrix.empty()) {

@leejet leejet Jul 5, 2026

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it would be better to pass in a --imat-prefix argument to specify the prefix, rather than trying to hardcode the traversal.

const std::vector<std::string> known_prefixes = {
"model.diffusion_model.",
"model.high_noise_diffusion_model.",
"model.diffusion_model.uncond.",

"cond_stage_model.transformer.",
"text_encoders.clip_l.transformer.",
"cond_stage_model.1.transformer.",
"text_encoders.clip_g.transformer.",

"text_encoders.t5xxl.transformer.",
"text_encoders.llm.",
"text_encoders.llm.visual.",

"vae.",
"tae."};
std::string processed_name_original = processed_name;

for (const auto& prefix : known_prefixes) {
processed_name = prefix + processed_name_original;
imatrix = get_imatrix_collector().get_values(processed_name);
if (!imatrix.empty()) {
break; // Exit the loop if a valid prefix is found
}
}
}
convert_tensor((void*)target_buf,
tensor_storage.type,
convert_buf,
Expand Down
Loading