Skip to content

[sft] load chat_template.jinja if available #509

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
merged 4 commits into from
Nov 4, 2025
+19 −3
Merged

[sft] load chat_template.jinja if available #509

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
6fa9816
chat_template.jinja file
jaysonfrancis Oct 29, 2025
875c450
lint
jaysonfrancis Oct 29, 2025
43bb944
check if exist
jaysonfrancis Oct 29, 2025
e88f941
refactor
jaysonfrancis Nov 3, 2025
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
3 changes: 3 additions & 0 deletions apps/sft/main.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -154,6 +154,9 @@ def setup_data(self):
generation_config_path=os.path.join(
self.job_config.model.hf_assets_path, "generation_config.json"
),
chat_template_path=(
path if os.path.exists(path := os.path.join(self.job_config.model.hf_assets_path, "chat_template.jinja")) else None
),
)

dataset = sft_iterable_dataset(
Expand Down
7 changes: 7 additions & 0 deletions src/forge/data/tokenizer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -42,6 +42,7 @@ def __init__(
*,
tokenizer_config_json_path: str | None = None,
generation_config_path: str | None = None,
chat_template_path: str | None = None,
):
self.tokenizer = Tokenizer.from_file(tokenizer_json_path)
if not (tokenizer_config_json_path or generation_config_path):
Expand All @@ -51,6 +52,10 @@ def __init__(
if tokenizer_config_json_path:
with open(tokenizer_config_json_path, "rb") as f:
self.config = json.load(f)
if chat_template_path:
with open(chat_template_path, "r") as f:
# TODO: warning in the case of overwrite?
self.config["chat_template"] = f.read()
else:
self.config = None
if generation_config_path:
Expand Down Expand Up @@ -227,12 +232,14 @@ def __init__(
*,
tokenizer_config_json_path: str | None = None,
generation_config_path: str | None = None,
chat_template_path: str | None = None,
truncation_type: str = "right",
):
self.base_tokenizer = HuggingFaceBaseTokenizer(
tokenizer_json_path=tokenizer_json_path,
tokenizer_config_json_path=tokenizer_config_json_path,
generation_config_path=generation_config_path,
chat_template_path=chat_template_path,
)

# Contents of the tokenizer_config.json
Expand Down