lveval.py 3.57 KB
Newer Older
1
len_levels = ['16k', '32k', '64k', '128k', '256k']
2
3

subsets_lveval_loogle_SD_mixup = [
4
    'LVEval_loogle_SD_mixup' + '_' + len_level for len_level in len_levels
5
6
]
subsets_lveval_cmrc_mixup = [
7
    'LVEval_cmrc_mixup' + '_' + len_level for len_level in len_levels
8
9
]
subsets_lveval_multifieldqa_en_mixup = [
10
    'LVEval_multifieldqa_en_mixup' + '_' + len_level
11
12
13
    for len_level in len_levels
]
subsets_lveval_multifieldqa_zh_mixup = [
14
    'LVEval_multifieldqa_zh_mixup' + '_' + len_level
15
16
17
    for len_level in len_levels
]
subsets_lveval_dureader_mixup = [
18
    'LVEval_dureader_mixup' + '_' + len_level for len_level in len_levels
19
20
]
subsets_lveval_loogle_CR_mixup = [
21
    'LVEval_loogle_CR_mixup' + '_' + len_level for len_level in len_levels
22
23
]
subsets_lveval_loogle_MIR_mixup = [
24
    'LVEval_loogle_MIR_mixup' + '_' + len_level for len_level in len_levels
25
26
]
subsets_lveval_hotpotwikiqa_mixup = [
27
    'LVEval_hotpotwikiqa_mixup' + '_' + len_level for len_level in len_levels
28
29
]
subsets_lveval_lic_mixup = [
30
    'LVEval_lic_mixup' + '_' + len_level for len_level in len_levels
31
32
]
subsets_lveval_factrecall_en = [
33
    'LVEval_factrecall_en' + '_' + len_level for len_level in len_levels
34
35
]
subsets_lveval_factrecall_zh = [
36
    'LVEval_factrecall_zh' + '_' + len_level for len_level in len_levels
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
]

subsets_lveval_single_hop_qa = (
    subsets_lveval_loogle_SD_mixup + subsets_lveval_cmrc_mixup
)
subsets_lveval_single_hop_cqa = (
    subsets_lveval_multifieldqa_en_mixup + subsets_lveval_multifieldqa_zh_mixup
)
subsets_lveval_multi_hop_qa = (
    subsets_lveval_dureader_mixup
    + subsets_lveval_loogle_CR_mixup
    + subsets_lveval_loogle_MIR_mixup
)
subsets_lveval_multi_hop_cqa = (
    subsets_lveval_hotpotwikiqa_mixup + subsets_lveval_lic_mixup
)
subsets_lveval_factrecall_cqa = (
    subsets_lveval_factrecall_en + subsets_lveval_factrecall_zh
)

subsets_lveval_qa = (
    subsets_lveval_single_hop_qa
    + subsets_lveval_single_hop_cqa
    + subsets_lveval_multi_hop_qa
    + subsets_lveval_multi_hop_cqa
    + subsets_lveval_factrecall_cqa
)

lveval_summary_groups = [
    {
67
68
        'name': 'LVEval_loogle_SD_mixup',
        'subsets': subsets_lveval_loogle_SD_mixup,
69
    },
70
    {'name': 'LVEval_cmrc_mixup', 'subsets': subsets_lveval_cmrc_mixup},
71
    {
72
73
        'name': 'LVEval_multifieldqa_en_mixup',
        'subsets': subsets_lveval_multifieldqa_en_mixup,
74
75
    },
    {
76
77
        'name': 'LVEval_multifieldqa_zh_mixup',
        'subsets': subsets_lveval_multifieldqa_zh_mixup,
78
79
    },
    {
80
81
        'name': 'LVEval_dureader_mixup',
        'subsets': subsets_lveval_dureader_mixup,
82
83
    },
    {
84
85
        'name': 'LVEval_loogle_CR_mixup',
        'subsets': subsets_lveval_loogle_CR_mixup,
86
87
    },
    {
88
89
        'name': 'LVEval_loogle_MIR_mixup',
        'subsets': subsets_lveval_loogle_MIR_mixup,
90
91
    },
    {
92
93
        'name': 'LVEval_hotpotwikiqa_mixup',
        'subsets': subsets_lveval_hotpotwikiqa_mixup,
94
    },
95
96
97
98
    {'name': 'LVEval_lic_mixup', 'subsets': subsets_lveval_lic_mixup},
    {'name': 'LVEval_factrecall_en', 'subsets': subsets_lveval_factrecall_en},
    {'name': 'LVEval_factrecall_zh', 'subsets': subsets_lveval_factrecall_zh},
    {'name': 'LVEval_single_hop_qa', 'subsets': subsets_lveval_single_hop_qa},
99
    {
100
101
        'name': 'LVEval_single_hop_cqa',
        'subsets': subsets_lveval_single_hop_cqa,
102
    },
103
104
    {'name': 'LVEval_multi_hop_qa', 'subsets': subsets_lveval_multi_hop_qa},
    {'name': 'LVEval_multi_hop_cqa', 'subsets': subsets_lveval_multi_hop_cqa},
105
    {
106
107
        'name': 'LVEval_factrecall_cqa',
        'subsets': subsets_lveval_factrecall_cqa,
108
    },
109
    {'name': 'LVEval_qa', 'subsets': subsets_lveval_qa},
110
]