File size: 2,993 Bytes
9d5b280
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
"""
Take in a YAML, and output all other splits with this YAML
"""

import argparse
import os

import requests
import yaml
from tqdm import tqdm

from lm_eval.utils import logging


API_URL = "https://datasets-server.huggingface.co/splits?dataset=facebook/belebele"


def parse_args():
    parser = argparse.ArgumentParser()
    parser.add_argument("--base_yaml_path", required=True)
    parser.add_argument("--save_prefix_path", default="belebele")
    parser.add_argument("--cot_prompt_path", default=None)
    parser.add_argument("--task_prefix", default="")
    return parser.parse_args()


if __name__ == "__main__":
    args = parse_args()

    # get filename of base_yaml so we can `"include": ` it in our other YAMLs.
    base_yaml_name = os.path.split(args.base_yaml_path)[-1]
    with open(args.base_yaml_path, encoding="utf-8") as f:
        base_yaml = yaml.full_load(f)

    if args.cot_prompt_path is not None:
        import json

        with open(args.cot_prompt_path, encoding="utf-8") as f:
            cot_file = json.load(f)

    def query():
        response = requests.get(API_URL)
        return response.json()["splits"]

    print(query())
    languages = [split["split"] for split in query()]

    for lang in tqdm([lang for lang in languages if "default" not in lang]):
        yaml_dict = {
            "include": base_yaml_name,
            "task": f"belebele_{args.task_prefix}_{lang}"
            if args.task_prefix != ""
            else f"belebele_{lang}",
            "test_split": lang,
            "fewshot_split": lang,
        }

        file_save_path = args.save_prefix_path + f"_{lang}.yaml"
        logging.info(f"Saving yaml for subset {lang} to {file_save_path}")
        with open(file_save_path, "w", encoding="utf-8") as yaml_file:
            yaml.dump(
                yaml_dict,
                yaml_file,
                width=float("inf"),
                allow_unicode=True,
                default_style='"',
            )

    # write group config out

    group_yaml_dict = {
        "group": f"belebele_{args.task_prefix}"
        if args.task_prefix != ""
        else "belebele",
        "task": [
            (
                f"belebele_{args.task_prefix}_{lang}"
                if args.task_prefix != ""
                else f"belebele_{lang}"
            )
            for lang in languages
            if "default" not in lang
        ],
        "aggregate_metric_list": [
            {"metric": "acc", "aggregation": "mean", "weight_by_size": False},
            {"metric": "acc_norm", "aggregation": "mean", "weight_by_size": False},
        ],
        "metadata": {"version": 0.0},
    }

    file_save_path = "_" + args.save_prefix_path + f"{args.task_prefix}.yaml"

    with open(file_save_path, "w", encoding="utf-8") as group_yaml_file:
        yaml.dump(
            group_yaml_dict,
            group_yaml_file,
            width=float("inf"),
            allow_unicode=True,
            default_style='"',
        )