File size: 5,810 Bytes
1c7fe13
 
 
 
 
 
 
 
 
 
 
 
 
e759b31
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
230ca5c
 
 
 
 
 
 
 
e759b31
 
230ca5c
 
 
 
 
 
 
 
e759b31
 
230ca5c
 
 
 
 
 
 
 
e759b31
 
230ca5c
 
 
 
 
 
 
 
e759b31
 
230ca5c
 
 
 
 
 
 
 
e759b31
 
230ca5c
 
 
 
 
 
 
 
e759b31
 
230ca5c
 
 
 
 
 
 
 
e759b31
 
230ca5c
 
 
 
 
 
 
 
e759b31
 
230ca5c
 
 
 
 
 
 
 
e759b31
 
230ca5c
 
 
 
 
 
 
 
e759b31
 
 
 
 
 
1c7fe13
1703b06
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
89deee5
e759b31
 
 
 
 
17b2190
 
6afc890
e759b31
17b2190
 
1c7fe13
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
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
from fasthtml.common import *
from fasthtml.components import *
from fasthtml.components import D_title, D_article, D_front_matter, D_contents, D_byline
from plotly import graph_objects as go
from fh_plotly import plotly2fasthtml
import pandas as pd
import json
from rich import print
import curated
import web
import common
import results

dataset_comparison = pd.DataFrame(
        {
            "Dataset": [
                "TxT360",
                "FineWeb",
                "RefinedWeb",
                "RedPajama-v2",
                "C4",
                "Dolma",
                "RedPajama-v1",
                "The Pile",
            ],
            "CommonCrawl": [
                "99 Snapshots",
                "96 Snapshots",
                "90 Snapshots",
                "84 Snapshots",
                "1 Snapshots",
                "24 Snapshots",
                "5 Snapshots",
                "0.6% of 74 Snapshots",
            ],
            "Papers": [
                "5 Sources",
                "-",
                "-",
                "-",
                "-",
                "1 Source",
                "1 Source",
                "4 Sources",
            ],
            "Wikipedia": [
                "310+ Languages",
                "-",
                "-",
                "-",
                "-",
                "what does a check mark mean?",
                "what does a check mark mean?",
                "English Only",
            ],
            "FreeLaw": [
                "Included",
                "-",
                "-",
                "-",
                "-",
                "-",
                "-",
                "Included",
            ],
            "DM Math": [
                "Included",
                "-",
                "-",
                "-",
                "-",
                "-",
                "-",
                "Included",
            ],
            "USPTO": [
                "Included",
                "-",
                "-",
                "-",
                "-",
                "-",
                "-",
                "Included",
            ],
            "PG-19": [
                "Included",
                "-",
                "-",
                "-",
                "-",
                "Included",
                "Included",
                "Included",
            ],
            "HackerNews": [
                "Included",
                "-",
                "-",
                "-",
                "-",
                "-",
                "-",
                "Included",
            ],
            "Ubuntu IRC": [
                "Included",
                "-",
                "-",
                "-",
                "-",
                "-",
                "-",
                "Included",
            ],
            "EuroParl": [
                "Included",
                "-",
                "-",
                "-",
                "-",
                "-",
                "-",
                "Included",
            ],
            "StackExchange": [
                "Included",
                "-",
                "-",
                "-",
                "-",
                "-",
                "Included",
                "Included",
            ],
            "Code": [
                "- what is this?",
                "-",
                "-",
                "-",
                "-",
                "Included",
                "Included",
                "Included",
            ],
        }
    )

table_html = dataset_comparison.to_html(index=False, border=0)
table_div = Div(NotStr(table_html), style="margin: 40px;")

dataset_sources = pd.DataFrame(
        {
            "Data Source": [
                "CommonCrawl",
                "Papers",
                "Wikipedia",
                "Freelaw",
                "DM Math",
                "USPTO",
                "PG-19",
                "HackerNews",
                "Ubuntu IRC",
                "Europarl",
                "StackExchange",
            ],
            "Raw Data Size": [
                "11 TB",
                "712 GB",
                "210 GB",
                "23 GB",
                "22 GB",
                "45 GB",
                "11 GB",
                "4.1 GB",
                "4.7 GB",
                "6.1 GB",
                "45 GB",
            ],
            "Token Count": [
                "5.71T",
                "154.96B",
                "4.75B",
                "7.34B",
                "5.23B",
                "4.95B",
                "2.94B",
                "1.08B",
                "1.54B",
                "1.96B",
                "8.37B",
            ],
            "Cut-Off Date": [
                "2024-30",
                "Q4 2023",
                "-",
                "Q1 2024",
                "-",
                "Q4 2023",
                "-",
                "Q4 2023",
                "Q4 2023",
                "-",
                "Q4 2023",
            ],
        }
    )

table_html = dataset_sources.to_html(index=False, border=0)
table_div1 = Div(NotStr(table_html), style="margin: 40px;")

def overview():
    return Div(Section(
            H2("Combining the Best of Web and Curated Sources"),
            H3("Why combine the web and highly curated sources? Isn't the web-only data enough?"),
            P("Table 1: TxT360 combines both the web data and highly-curated sources, which none of the existing datasets have covered. The following table shows TxT360 and other well-known datasets on the coverage and size of data sources."),
            table_div,
            P("Table 2: Statistics of TxT360. The basic statistics of TxT360 are presented."),
            table_div1,
            id="inner-text",
        )
    )