Commit
•
8061116
1
Parent(s):
ec2b3ce
Update main.py
Browse files
main.py
CHANGED
|
@@ -175,6 +175,167 @@ def main():
|
|
| 175 |
)
|
| 176 |
|
| 177 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 178 |
dataset_comparison1 = pd.DataFrame(
|
| 179 |
{
|
| 180 |
"Dataset": [
|
|
@@ -474,6 +635,7 @@ def intro():
|
|
| 474 |
H3(
|
| 475 |
"TxT360 is the first dataset to combine both web and curated data sources commonly used in pretraining."
|
| 476 |
),
|
|
|
|
| 477 |
table_div_1,
|
| 478 |
table_div_2,
|
| 479 |
P(
|
|
|
|
| 175 |
)
|
| 176 |
|
| 177 |
|
| 178 |
+
new_dataset_comparison1 = pd.DataFrame(
|
| 179 |
+
{
|
| 180 |
+
"Data Source": [
|
| 181 |
+
"CommonCrawl",
|
| 182 |
+
"Papers",
|
| 183 |
+
"Wikipedia",
|
| 184 |
+
"FreeLaw",
|
| 185 |
+
"DM Math",
|
| 186 |
+
"USPTO",
|
| 187 |
+
"PG-19",
|
| 188 |
+
"HackerNews",
|
| 189 |
+
"Ubuntu IRC",
|
| 190 |
+
"EuroParl",
|
| 191 |
+
"StackExchange",
|
| 192 |
+
"Code",
|
| 193 |
+
|
| 194 |
+
],
|
| 195 |
+
"TxT360": [
|
| 196 |
+
"99 Snapshots",
|
| 197 |
+
"5 Sources",
|
| 198 |
+
"310+ Languages",
|
| 199 |
+
"Included",
|
| 200 |
+
"Included",
|
| 201 |
+
"Included",
|
| 202 |
+
"Included",
|
| 203 |
+
"Included",
|
| 204 |
+
"Included",
|
| 205 |
+
"Included",
|
| 206 |
+
"Included",
|
| 207 |
+
"**",
|
| 208 |
+
],
|
| 209 |
+
"FineWeb": [
|
| 210 |
+
"96 Snapshots",
|
| 211 |
+
"-",
|
| 212 |
+
"-",
|
| 213 |
+
"-",
|
| 214 |
+
"-",
|
| 215 |
+
"-",
|
| 216 |
+
"-",
|
| 217 |
+
"-",
|
| 218 |
+
"-",
|
| 219 |
+
"-",
|
| 220 |
+
"-",
|
| 221 |
+
"-",
|
| 222 |
+
],
|
| 223 |
+
"RefinedWeb": [
|
| 224 |
+
"90 Snapshots",
|
| 225 |
+
"-",
|
| 226 |
+
"-",
|
| 227 |
+
"-",
|
| 228 |
+
"-",
|
| 229 |
+
"-",
|
| 230 |
+
"-",
|
| 231 |
+
"-",
|
| 232 |
+
"-",
|
| 233 |
+
"-",
|
| 234 |
+
"-",
|
| 235 |
+
"-",
|
| 236 |
+
],
|
| 237 |
+
"PedPajama-V-2": [
|
| 238 |
+
"84 Snapshots",
|
| 239 |
+
"-",
|
| 240 |
+
"-",
|
| 241 |
+
"-",
|
| 242 |
+
"-",
|
| 243 |
+
"-",
|
| 244 |
+
"-",
|
| 245 |
+
"-",
|
| 246 |
+
"-",
|
| 247 |
+
"-",
|
| 248 |
+
"-",
|
| 249 |
+
"-",
|
| 250 |
+
],
|
| 251 |
+
"C4": [
|
| 252 |
+
"1 Snapshots",
|
| 253 |
+
"-",
|
| 254 |
+
"-",
|
| 255 |
+
"-",
|
| 256 |
+
"-",
|
| 257 |
+
"-",
|
| 258 |
+
"-",
|
| 259 |
+
"-",
|
| 260 |
+
"-",
|
| 261 |
+
"-",
|
| 262 |
+
"-",
|
| 263 |
+
"-",
|
| 264 |
+
],
|
| 265 |
+
"Dolma": [
|
| 266 |
+
"24 Snapshots",
|
| 267 |
+
"1 Source",
|
| 268 |
+
"checkmark",
|
| 269 |
+
"-",
|
| 270 |
+
"-",
|
| 271 |
+
"-",
|
| 272 |
+
"Included",
|
| 273 |
+
"-",
|
| 274 |
+
"-",
|
| 275 |
+
"-",
|
| 276 |
+
"-",
|
| 277 |
+
"Included",
|
| 278 |
+
],
|
| 279 |
+
"RedPajama-V-1": [
|
| 280 |
+
"5 Snapshots",
|
| 281 |
+
"1 Source",
|
| 282 |
+
"checkmark",
|
| 283 |
+
"",
|
| 284 |
+
" ",
|
| 285 |
+
"",
|
| 286 |
+
"Included",
|
| 287 |
+
"-",
|
| 288 |
+
"-",
|
| 289 |
+
"-",
|
| 290 |
+
"Included",
|
| 291 |
+
"Included",
|
| 292 |
+
],
|
| 293 |
+
"The Pile": [
|
| 294 |
+
"0.6% of 74 Snapshots",
|
| 295 |
+
"4 Sources",
|
| 296 |
+
"English Only",
|
| 297 |
+
"Included",
|
| 298 |
+
"Included",
|
| 299 |
+
"Included",
|
| 300 |
+
"Included",
|
| 301 |
+
"Included",
|
| 302 |
+
"Included",
|
| 303 |
+
"Included",
|
| 304 |
+
"Included",
|
| 305 |
+
"Included",
|
| 306 |
+
],
|
| 307 |
+
}
|
| 308 |
+
)
|
| 309 |
+
|
| 310 |
+
|
| 311 |
+
# Apply table styling: Light green for the header, alternating white and light grey for rows
|
| 312 |
+
styled_table = (
|
| 313 |
+
new_dataset_comparison1.style.set_properties(
|
| 314 |
+
**{"background-color": "#E1EEDB"},
|
| 315 |
+
subset=pd.IndexSlice[0, :], # Row 0 with a light green background
|
| 316 |
+
)
|
| 317 |
+
.apply(
|
| 318 |
+
lambda x: [
|
| 319 |
+
"background-color: #E1EEDB"
|
| 320 |
+
if i == 0
|
| 321 |
+
else (
|
| 322 |
+
"background-color: rgb(237, 242, 251)"
|
| 323 |
+
if i % 2 == 0
|
| 324 |
+
else "background-color: white"
|
| 325 |
+
)
|
| 326 |
+
for i in range(len(x))
|
| 327 |
+
],
|
| 328 |
+
axis=0,
|
| 329 |
+
)
|
| 330 |
+
.hide(axis="index")
|
| 331 |
+
) # Hide the row index
|
| 332 |
+
|
| 333 |
+
# Use _repr_html_() method to get the HTML representation of the styled DataFrame
|
| 334 |
+
table_html = styled_table._repr_html_()
|
| 335 |
+
# table_html = dataset_comparison1.to_html(index=False, border=0)
|
| 336 |
+
new_table_div_1 = Div(NotStr(table_html), style="margin: 40px;")
|
| 337 |
+
|
| 338 |
+
|
| 339 |
dataset_comparison1 = pd.DataFrame(
|
| 340 |
{
|
| 341 |
"Dataset": [
|
|
|
|
| 635 |
H3(
|
| 636 |
"TxT360 is the first dataset to combine both web and curated data sources commonly used in pretraining."
|
| 637 |
),
|
| 638 |
+
new_table_div_1,
|
| 639 |
table_div_1,
|
| 640 |
table_div_2,
|
| 641 |
P(
|