src/about.py

from dataclasses import dataclass
from enum import Enum

@dataclass
class Task:
    benchmark: str
    metric: str
    col_name: str


# Select your tasks here
# ---------------------------------------------------
class Tasks(Enum):
    # task_key in the json file, metric_key in the json file, name to display in the leaderboard 
    task0 = Task("blimp", "acc", "BLiMP")
    task1 = Task("blimp_supplement", "acc", "BLiMP Supplement")
    task2 = Task("glue", "acc", "(Super)GLUE")
    task3 = Task("ewok", "acc", "EWoK")

class TasksMultimodal(Enum):
    task0 = Task("blimp", "acc", "BLiMP")
    task1 = Task("blimp_supplement", "acc", "BLiMP Supplement")
    task2 = Task("glue", "acc", "(Super)GLUE")
    task3 = Task("ewok", "acc", "EWoK")
    task4 = Task("vqa", "acc", "VQA")
    task5 = Task("winoground", "acc", "Winoground")
    task6 = Task("devbench", "acc", "DevBench")

NUM_FEWSHOT = 0 # Change with your few shot
# ---------------------------------------------------



# Your leaderboard name
TITLE = """<h1 align="center" id="space-title">BabyLM 2024 Leaderboards</h1>"""

# What does your leaderboard evaluate?
INTRODUCTION_TEXT = """
The leaderboards for each track of the 2024 BabyLM Challenge.
"""

# Which evaluations are you running? how can people reproduce what you have?
LLM_BENCHMARKS_TEXT = f"""
This leaderboard displays scores from the 2024 BabyLM Challenge. Each track has its own tab.
"""

EVALUATION_QUEUE_TEXT = """
## Some good practices before requesting a model upload:

Make sure you can get scores from your prediction using the `score_predictions.py` script.
```bash
git clone https://github.com/babylm/evaluation-pipeline-2024/
cd evaluation-pipeline-2024
python score_predictions.py path/to/your/predictions.json.gz
```
If this step fails, follow the error messages to debug your model before submitting it. It's likely that either (i) some results are missing, or (ii) the results are incorrectly formatted.

Make sure your model has an open license! This is a leaderboard that is meant to advance research on language modeling, and we'd love for as many people as possible to know they can use your model.

Once these steps have been followed, get in touch with the organizers with your predictions file(s), and the scores you've obtained.
We'll verify that we can match your scores, and then upload to the leaderboard. Optionally, you can give us your preferred model display name for the leaderboard, and a link to your model on HuggingFace.
"""

CITATION_BUTTON_LABEL = "If you would like to cite these results, please cite the 2024 BabyLM Findings paper, as well as the authors of the model(s) whose results you cite!"
CITATION_BUTTON_TEXT = r"""
@article{hu2024findingssecondbabylmchallenge,
      title={Findings of the Second BabyLM Challenge: Sample-Efficient Pretraining on Developmentally Plausible Corpora}, 
      author={Michael Y. Hu and Aaron Mueller and Candace Ross and Adina Williams and Tal Linzen and Chengxu Zhuang and Ryan Cotterell and Leshem Choshen and Alex Warstadt and Ethan Gotlieb Wilcox},
      year={2024},
      journal={Computing Research Repository},
      volume={arXiv:2412.05149},
      url={https://arxiv.org/abs/2412.05149}, 
}
"""