Skip to content
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

extract_lora.py improvements and fixes #333

Merged
merged 9 commits into from
Aug 5, 2024
Merged

extract_lora.py improvements and fixes #333

Show file tree
Hide file tree
Changes from 8 commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
2812332
added min_fraction_variance option and truncate if new_weight.shape !…
jukofyork May 24, 2024
25f2a27
Fixed passing null rank_pattern
jukofyork May 24, 2024
8ba6ede
Merge branch 'arcee-ai:main' into extract_lora.py-changes
jukofyork May 26, 2024
b280c34
Update extract_lora.py
jukofyork May 26, 2024
d74a458
[mergekit-extract-lora] added flag to allow vocab extension
thomasgauthier Jul 18, 2024
0a622ae
[mergekit-extract-lora] added vocab extension details to generated mo…
thomasgauthier Jul 18, 2024
17ff289
[mergekit-extract-lora] renamed a method for clarity
thomasgauthier Jul 18, 2024
9f9b9d4
Merge branch 'main' into extract_lora.py-changes
thomasgauthier Jul 18, 2024
cd7968d
Merge branch 'main' into extract_lora.py-changes
thomasgauthier Aug 5, 2024
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: 9 additions & 1 deletion mergekit/card.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -202,6 +202,8 @@ def generate_card_lora(
base_model_ref: ModelReference,
finetuned_model_ref: ModelReference,
invocation: str,
extended: bool,
vocab_size: int,
name: str,
) -> str:
"""
Expand All @@ -218,7 +220,13 @@ def generate_card_lora(
hf_bases = list(extract_hf_paths([base_model_ref, finetuned_model_ref]))
tags = ["mergekit", "peft"]

details = f"This LoRA adapter was extracted from {modelref_md(finetuned_model_ref)} and uses {modelref_md(base_model_ref)} as a base."
finetuned_ref_md = modelref_md(finetuned_model_ref)
basemodel_ref_md = modelref_md(base_model_ref)

details = f"This LoRA adapter was extracted from {finetuned_ref_md} and uses {basemodel_ref_md} as a base."

if extended:
details += f"\n\n> [!WARNING]\n> This LoRA adapter has an extended vocabulary. Make sure to call `model.resize_token_embeddings({vocab_size})` before applying the adapter to {basemodel_ref_md}"

if os.path.isdir(base_model_ref.model.path) or os.path.isdir(
finetuned_model_ref.model.path
Expand Down
Loading
Loading