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 TasksRGB(Enum):
    # task_key in the json file, metric_key in the json file, name to display in the leaderboard
    task0 = Task("mRNA", "RMSE", "mRNA (RMSE)")
    task1 = Task("SNMD", "AUC", "SNMD (AUC)")
    task2 = Task("SNMR", "F1", "SNMR (F1)")
    task3 = Task("ArchiveII", "F1", "ArchiveII (F1)")
    task4 = Task("bpRNA", "F1", "bpRNA (F1)")
    task5 = Task("RNAStralign", "F1", "RNAStralign (F1)")

class TasksPGB(Enum):
    # task_key in the json file, metric_key in the json file, name to display in the leaderboard
    task0 = Task("PolyA", "F1", "PolyA (F1)")
    task1 = Task("LncRNA", "F1", "LncRNA (F1)")
    task2 = Task("Chrom Acc", "F1", "Chrom Acc (F1)")
    task3 = Task("Prom Str", "RMSE", "Prom Str (RMSE)")
    task4 = Task("Term Str", "RMSE", "Term Str (RMSE)")
    task5 = Task("Splice", "F1", "Splice (F1)")
    task6 = Task("Gene Exp", "RMSE", "Gene Exp (RMSE)")
    task7 = Task("Enhancer", "F1", "Enhancer (F1)")

class TasksGUE(Enum):
    # task_key in the json file, metric_key in the json file, name to display in the leaderboard
    task0 = Task("Yeast EMP", "F1", "Yeast EMP (F1)")
    task1 = Task("Mouse TF-M", "F1", "Mouse TF-M (F1)")
    task2 = Task("Virus CVC", "F1", "Virus CVC (F1)")
    task3 = Task("Human TF-H", "F1", "Human TF-H (F1)")
    task4 = Task("Human PD", "F1", "Human PD (F1)")
    task5 = Task("Human CPD", "F1", "Human CPD (F1)")
    task6 = Task("Human SSP", "F1", "Human SSP (F1)")

class TasksGB(Enum):
    # task_key in the json file, metric_key in the json file, name to display in the leaderboard
    task0 = Task("DEM", "F1", "DEM (F1)")
    task1 = Task("DOW", "F1", "DOW (F1)")
    task2 = Task("DRE", "F1", "DRE (F1)")
    task3 = Task("DME", "F1", "DME (F1)")
    task4 = Task("HCE", "F1", "HCE (F1)")
    task5 = Task("HEE", "F1", "HEE (F1)")
    task6 = Task("HRE", "F1", "HRE (F1)")
    task7 = Task("HNP", "F1", "HNP (F1)")
    task8 = Task("HOR", "F1", "HOR (F1)")


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

# Your leaderboard name
TITLE = """<h1 align="center" id="space-title">Genomic Modelling Leaderboard</h1>"""

# What does your leaderboard evaluate?
INTRODUCTION_TEXT = """
"""

# Which evaluations are you running? how can people reproduce what you have?
LLM_BENCHMARKS_TEXT = f"""
## Why do we need this benchmark?
Large-scale foundation models for molecular biology constitute a vital and rapidly developing change in the computational biology and AI4Science landscape.
As key parts of biology, such as DNA, RNA sequences, secondary structures, have a large effect on each other, the usage of this information within large-scale models allows for foundation models to be adapted and suited to multiple key tasks.
However, with this trend comes significant issues, the primary one being the difficulty to comprehensively evaluate these models and compare them fairly.
Here, we refer to the specific lack of real-world data to reflect the true performance of the models, rather than in-silico experiments only.
This issue forces repeated benchmark testing and models being trained and adapted for a specific task that may not have any real-world benefit.
Given the importance of this, we propose this genomic leaderboard on meticulously curated real-world datasets, to allow for a fair and comprehensive benchmark on the most important genomic downstream tasks.
## Evaluation Datasets
TODO HERE
## Reported Scores and Ranking
TODO HERE
## How it works
Do we need this?
## Reproducibility
To reproduce our results, here are the commands you can run:
"""

EVALUATION_QUEUE_TEXT = """
## Some good practices before submitting a model
### 1) Make sure you can load your model and tokenizer using AutoClasses:
```python
from transformers import AutoConfig, AutoModel, AutoTokenizer
config = AutoConfig.from_pretrained("your model name", revision=revision)
model = AutoModel.from_pretrained("your model name", revision=revision)
tokenizer = AutoTokenizer.from_pretrained("your model name", revision=revision)
```
If this step fails, follow the error messages to debug your model before submitting it. It's likely your model has been improperly uploaded.
Note: make sure your model is public!
Note: if your model needs `use_remote_code=True`, we do not support this option yet but we are working on adding it, stay posted!
### 2) Convert your model weights to [safetensors](https://huggingface.co/docs/safetensors/index)
It's a new format for storing weights which is safer and faster to load and use. It will also allow us to add the number of parameters of your model to the `Extended Viewer`!
### 3) Make sure your model has an open license!
This is a leaderboard for Open LLMs, and we'd love for as many people as possible to know they can use your model 🤗
### 4) Fill up your model card
When we add extra information about models to the leaderboard, it will be automatically taken from the model card
## In case of model failure
If your model is displayed in the `FAILED` category, its execution stopped.
Make sure you have followed the above steps first.
If everything is done, check you can launch the EleutherAIHarness on your model locally, using the above command without modifications (you can add `--limit` to limit the number of examples per task).
"""

CITATION_BUTTON_LABEL = "Copy the following snippet to cite these results"
CITATION_BUTTON_TEXT = r"""
@article{Yang2024,
  author = {Yang, Heng and Li, Ke},
  title = {OmniGenome: Aligning RNA Sequences with Secondary Structures in Genomic Foundation Models},
  journal = {arXiv},
  year = {2024},
  note = {arXiv preprint arXiv:2407.11242}
  url = {https://arxiv.org/abs/2407.11242}
}
"""