Skip to content

update conversion script to support codellama-70b #97

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

update conversion script to support codellama-70b #97

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
4 changes: 3 additions & 1 deletion weights_conversion/hf_to_megatron.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -358,6 +358,8 @@ def main(model_name: str = "falcon", size: int = 7, out: Optional[Path] = None,
args.update({"num_attention_heads_kv": 8})
if size < 34 and not re.match(r"CodeLlama-\d+b-Python", cache_dir):
args.update({"padded_vocab_size": 32016})
if size == 70: # The vocab size of the three versions of codellama-70b is 32016
args.update({"padded_vocab_size": 32016})
else:
sys.exit(f"Model name has to be llama, llama2 or codellama, not {model_name}.")

Expand Down Expand Up @@ -440,7 +442,7 @@ def main(model_name: str = "falcon", size: int = 7, out: Optional[Path] = None,
elif args.model == "llama":
assert args.size in {7, 13, 30, 65}
elif args.model == "codellama":
assert args.size in {7, 13, 34}
assert args.size in {7, 13, 34, 70}
elif args.model == "mistral":
assert args.size in {7}
else:
Expand Down