from dataclasses import dataclass from enum import Enum @dataclass class TaskDetails: name: str display_name: str = "" symbol: str = "" # emoji class TaskType(Enum): NLU = TaskDetails("nlu", "NLU", "🧠") NLG = TaskDetails("nlg", "NLG", "✍️") @dataclass class Task: benchmark: str metric: str col_name: str task_type: TaskType is_primary_metric: bool = True # 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("sentiment_mlt", "f1", "Sentiment Analysis (F1)", TaskType.NLU) task1 = Task("sib200_mlt", "f1", "SIB200 (F1)", TaskType.NLU) task2 = Task("taxi1500_mlt", "f1", "Taxi1500 (F1)", TaskType.NLU) task3 = Task("maltese_news_categories", "loglikelihood", "Maltese News Categories (F1)", TaskType.NLU) task4 = Task("multieurlex_mlt", "loglikelihood", "MultiEURLEX (F1)", TaskType.NLU) task5 = Task("belebele_mlt", "acc", "Belebele (Accuracy)", TaskType.NLU) task6 = Task("opus100_eng-mlt", "bleu", "OPUS-100 EN→MT (BLEU)", TaskType.NLG, False) task7 = Task("opus100_eng-mlt", "chrf", "OPUS-100 EN→MT (ChrF)", TaskType.NLG) task8 = Task("flores200_eng-mlt", "bleu", "Flores-200 EN→MT (BLEU)", TaskType.NLG, False) task9 = Task("flores200_eng-mlt", "chrf", "Flores-200 EN→MT (ChrF)", TaskType.NLG) task10 = Task("webnlg_mlt", "chrf", "WebNLG (ChrF)", TaskType.NLG) task11 = Task("webnlg_mlt", "rouge", "WebNLG (Rouge-L)", TaskType.NLG, False) task12 = Task("eurlexsum_mlt", "chrf", "EUR-Lex-Sum (ChrF)", TaskType.NLG, False) task13 = Task("eurlexsum_mlt", "rouge", "EUR-Lex-Sum (Rouge-L)", TaskType.NLG) task14 = Task("maltese_news_headlines", "chrf", "Maltese News Headlines (ChrF)", TaskType.NLG, False) task15 = Task("maltese_news_headlines", "rouge", "Maltese News Headlines (Rouge-L)", TaskType.NLG) NUM_FEWSHOT = 0 # Change with your few shot # --------------------------------------------------- # Your leaderboard name TITLE = """

🇲🇹 MELABench Leaderboard

""" # What does your leaderboard evaluate? INTRODUCTION_TEXT = """ A Maltese Evaluation Language Benchmark """ # Which evaluations are you running? how can people reproduce what you have? LLM_BENCHMARKS_TEXT = f""" ## How it works ## Reproducibility To reproduce our results, here is the commands you can run: """ EVALUATION_QUEUE_TEXT = """ To include new results on this benchmark, follow the instructions on our [GitHub Repository](https://github.com/MLRS/MELABench/tree/main/prompting). You can then upload the output files which should include the configuration/results file and all the prediction files. In addition, we ask for additional metadata about model training. """ CITATION_BUTTON_LABEL = "Copy the following snippet to cite these results" CITATION_BUTTON_TEXT = r""" """