Hugging Face's logo

Datasets:
manu
/
french-30b

Modalities:
Text
Formats:
parquet
Size:
10M - 100M
Libraries:
Datasets
Dask
Dataset card Viewer Files Community
2
manu commited on
Commit
ad8031f
•
1 Parent(s): 0c07b21

Upload config.yaml with huggingface_hub

Browse files
Files changed (1) hide show
  1. config.yaml +153 -0
config.yaml ADDED
@@ -0,0 +1,153 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ data_mix:
2
+ (): dataset_construction.DataMix
3
+ name: "french_30b"
4
+ shuffle: false
5
+ compute_dataset_stats: true
6
+ local_save_dir: data/
7
+ load_from_local_save_dir: false
8
+ keep_separated_datasets_in_dataset_dict: false
9
+ deduplicate_test_set: false
10
+ ngram_path_for_extra_deduplication: null
11
+ max_shard_size: "2GB"
12
+ datasets:
13
+ # Translation datasets
14
+ # - (): dataset_construction.DatasetConfig
15
+ # dataset_path: manu/opus100-en-fr
16
+ # test_split: "test"
17
+ # build_test_set_from_train: false
18
+ # num_train_examples: 1000
19
+ # num_test_examples: 1000
20
+ # - (): dataset_construction.DatasetConfig
21
+ # dataset_path: manu/europarl-en-fr
22
+ # build_test_set_from_train: true
23
+ # num_test_examples: 1000
24
+ - (): dataset_construction.DatasetConfig
25
+ dataset_path: manu/wmt-en-fr
26
+ test_split: "test"
27
+ - (): dataset_construction.DatasetConfig
28
+ dataset_path: Nicolas-BZRD/English_French_Webpages_Scraped_Translated # Slow cause bad format
29
+ train_split: "train[:5%]"
30
+ build_test_set_from_train: true
31
+ filtering_function:
32
+ (): dataset_collection.french.web_scraped_translated.WebPageFilter
33
+ preprocessing_function:
34
+ (): dataset_collection.french.web_scraped_translated.WebPageMapper
35
+ # global voices can be added but maybe in wmt test set
36
+
37
+ # Speech datasets
38
+ - (): dataset_construction.DatasetConfig
39
+ dataset_path: manu/french_librispeech_text_only
40
+ build_test_set_from_train: true
41
+ - (): dataset_construction.DatasetConfig
42
+ dataset_path: manu/french_podcasts
43
+ build_test_set_from_train: true
44
+ num_test_examples: 100
45
+ filtering_function:
46
+ (): dataset_collection.french.french_transcribed_podcast.PodcastFilter
47
+ preprocessing_function:
48
+ (): dataset_collection.french.french_transcribed_podcast.PodcastMapper
49
+ - (): dataset_construction.DatasetConfig
50
+ dataset_path: manu/french_open_subtitles
51
+ num_test_examples: 100
52
+ build_test_set_from_train: true
53
+ - (): dataset_construction.DatasetConfig
54
+ dataset_path: Nicolas-BZRD/Original_Songs_Lyrics_with_French_Translation
55
+ build_test_set_from_train: true
56
+ filtering_function:
57
+ (): dataset_collection.french.french_translated_lyrics.LyricsFilter
58
+ preprocessing_function:
59
+ (): dataset_collection.french.french_translated_lyrics.LyricsMapper
60
+
61
+ # Books
62
+ # - (): dataset_construction.DatasetConfig
63
+ # dataset_path: manu/project_gutenberg
64
+ # train_split: fr
65
+ # build_test_set_from_train: true
66
+ # num_train_examples: 1000
67
+ # num_test_examples: 100
68
+ # here we should preprocess
69
+ - (): dataset_construction.DatasetConfig
70
+ dataset_path: manu/ProjectGutenberg_fr
71
+ build_test_set_from_train: true
72
+ - (): dataset_construction.DatasetConfig
73
+ dataset_path: manu/bnf_gallica
74
+ build_test_set_from_train: true
75
+ - (): dataset_construction.DatasetConfig
76
+ dataset_path: manu/theses_fr_2013_2023
77
+ build_test_set_from_train: true
78
+ filtering_function:
79
+ (): dataset_collection.french.theses.ThesisFilter
80
+ preprocessing_function:
81
+ (): dataset_collection.french.theses.ThesisMapper
82
+ # Dila data
83
+ - (): dataset_construction.DatasetConfig
84
+ dataset_path: Nicolas-BZRD/LEGI_opendata # manu/dila_legifrance
85
+ build_test_set_from_train: true
86
+ - (): dataset_construction.DatasetConfig
87
+ dataset_path: Nicolas-BZRD/BALO_opendata
88
+ build_test_set_from_train: true
89
+ - (): dataset_construction.DatasetConfig
90
+ dataset_path: Nicolas-BZRD/JADE_opendata
91
+ build_test_set_from_train: true
92
+ - (): dataset_construction.DatasetConfig
93
+ dataset_path: Nicolas-BZRD/DOLE_opendata
94
+ build_test_set_from_train: true
95
+ - (): dataset_construction.DatasetConfig
96
+ dataset_path: Nicolas-BZRD/SARDE_opendata
97
+ build_test_set_from_train: true
98
+ - (): dataset_construction.DatasetConfig
99
+ dataset_path: Nicolas-BZRD/QR_opendata
100
+ build_test_set_from_train: true
101
+ - (): dataset_construction.DatasetConfig
102
+ dataset_path: Nicolas-BZRD/JORF_opendata
103
+ build_test_set_from_train: true
104
+ - (): dataset_construction.DatasetConfig
105
+ dataset_path: Nicolas-BZRD/INCA_opendata
106
+ build_test_set_from_train: true
107
+ - (): dataset_construction.DatasetConfig
108
+ dataset_path: Nicolas-BZRD/ACCO_opendata
109
+ build_test_set_from_train: true
110
+ - (): dataset_construction.DatasetConfig
111
+ dataset_path: Nicolas-BZRD/KALI_opendata
112
+ build_test_set_from_train: true
113
+ - (): dataset_construction.DatasetConfig
114
+ dataset_path: Nicolas-BZRD/DEBATS_opendata
115
+ build_test_set_from_train: true
116
+ - (): dataset_construction.DatasetConfig
117
+ dataset_path: Nicolas-BZRD/CNIL_opendata
118
+ build_test_set_from_train: true
119
+ - (): dataset_construction.DatasetConfig
120
+ dataset_path: Nicolas-BZRD/CAPP_opendata
121
+ build_test_set_from_train: true
122
+ - (): dataset_construction.DatasetConfig
123
+ dataset_path: Nicolas-BZRD/CASS_opendata
124
+ build_test_set_from_train: true
125
+ - (): dataset_construction.DatasetConfig
126
+ dataset_path: Nicolas-BZRD/CONSTIT_opendata
127
+ build_test_set_from_train: true
128
+
129
+ # needs a preprocessor to have an id_column
130
+ # PDF dataset
131
+ - (): dataset_construction.DatasetConfig
132
+ dataset_path: manu/illuin_layout_dataset_text_only
133
+ build_test_set_from_train: true
134
+ # Wiki
135
+ - (): dataset_construction.DatasetConfig
136
+ dataset_path: manu/wikisource_fr
137
+ build_test_set_from_train: true
138
+ - (): dataset_construction.DatasetConfig
139
+ dataset_path: wikipedia
140
+ dataset_name: "20220301.fr"
141
+ build_test_set_from_train: true
142
+ # Internet dumps
143
+ - (): dataset_construction.DatasetConfig
144
+ dataset_path: 'oscar-corpus/OSCAR-2301'
145
+ dataset_name: fr
146
+ train_split: "train[:10%]"
147
+ build_test_set_from_train: true
148
+ num_test_examples: 10000
149
+ filtering_function:
150
+ (): dataset_collection.french.oscar.OscarFilter
151
+
152
+
153
+ tokenizer: "mistralai/Mistral-7B-v0.1"