Skip to content

Add GraniteInstruct prompt style and Granite 3.3 instruct models #152

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

Add GraniteInstruct prompt style and Granite 3.3 instruct models #152

Show file tree
Hide file tree
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
16 changes: 16 additions & 0 deletions lcb_runner/lm_styles.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,6 +31,7 @@ class LMStyle(Enum):
QwQ = "QwQ"
LLaMa3 = "LLaMa3"
DeepSeekR1 = "DeepSeekR1"
GraniteInstruct = "GraniteInstruct"

TogetherAI = "TogetherAI"

Expand Down Expand Up @@ -861,6 +862,21 @@ def to_dict(self) -> dict:
datetime(2024, 4, 1),
"https://huggingface.co/agentica-org/DeepCoder-14B-Preview",
),
## Granite 3.3 Instruct (2B and 8B)
LanguageModel(
"ibm-granite/granite-3.3-2b-instruct",
"Granite-3.3-2B-Ins",
LMStyle.GraniteInstruct,
datetime(2025, 4, 16),
"https://huggingface.co/ibm-granite/granite-3.3-2b-instruct",
),
LanguageModel(
"ibm-granite/granite-3.3-8b-instruct",
"Granite-3.3-8B-Ins",
LMStyle.GraniteInstruct,
datetime(2025, 4, 16),
"https://huggingface.co/ibm-granite/granite-3.3-8b-instruct",
),
]

LanguageModelStore: dict[str, LanguageModel] = {
Expand Down
25 changes: 25 additions & 0 deletions lcb_runner/prompts/code_generation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -335,6 +335,31 @@ def format_prompt_generation(
prompt += f"{get_deepseek_r1_question_template_answer(question)}"
return prompt

if LanguageModelStyle == LMStyle.GraniteInstruct:
chat_messages = [
{
"role": "system",
"content": PromptConstants.SYSTEM_MESSAGE_GENERIC,
},
{
"role": "user",
"content": get_generic_question_template_answer(question),
},
]
from transformers import AutoTokenizer

# Granite 3.x instruct models share one chat template; only the template is used here.
tokenizer = AutoTokenizer.from_pretrained(
"ibm-granite/granite-3.3-2b-instruct", padding_side="left"
)
return tokenizer.apply_chat_template(
chat_messages,
tokenize=False,
add_generation_prompt=True,
truncation=False,
padding=False,
)

if LanguageModelStyle == LMStyle.GenericBase:
prompt = get_base_model_question_template_answer(question)
return prompt
Expand Down