TxT360 / data /txt360_eval /CKPT Eval - NQ.csv
hunterhector's picture
fix data columns
e74bc72
5-shot,FineWeb-1.5T,Ours-Base,Ours-Upsampling2,All-Upsampling1
time: 22 min,Llama-8x8B-seq8192,Llama-8x8B-seq8192,Llama-8x8B-seq8192,Llama-8x8B-seq8192
5k,0.0341,0.0416,0.0565,0.0526
10k,0.0715,,0.0931,0.0767
15k,0.0765,,0.1061,0.1127
20k,0.0787,,0.1183,0.1247
25k,0.0892,0.115,0.1352,0.1343
30k,0.0911,0.1366,0.1271,0.1421
35k,0.097,0.1488,0.1485,0.1524
40k,0.1028,0.1355,0.1488,0.1562
45k,0.1078,0.1488,0.162,0.1598
50k,0.105,0.154,0.159,0.1698
55k,0.1097,0.1607,0.1662,0.1704
60k,0.1211,0.1654,0.1612,0.1801
65k,0.1089,0.1573,0.1693,0.1823
70k,0.1222,0.1634,0.1679,0.1767
75k,0.1097,0.1709,0.1881,0.1762
80k,0.1277,0.1573,0.1776,0.1964
85k,0.128,0.1776,0.1889,0.1889
90k,0.1158,0.1598,0.1806,0.1773
95k,0.1235,0.1762,0.1781,0.1917
100k,0.1258,,0.1928,0.1947
105k,0.1366,,0.1814,0.2094
110k,0.1377,0.1756,0.1859,
115k,0.1346,0.1831,0.1947,0.2119
120k,0.1402,0.2014,,0.2119
125k,0.1307,0.203,0.1992,0.1787
130k,0.1368,0.1997,0.1994,0.2086
135k,0.1363,,0.2014,0.2069
140k,0.1435,,0.1986,0.2058
145k,0.1532,,0.1953,0.2102
150k,0.1404,,,0.2075
155k,0.1418,,0.1931,0.2205
160k,0.1346,,0.2116,0.2208
165k,0.1524,,0.2139,0.2213
170k,0.1388,,,0.2169
175k,0.1438,,0.2222,0.2321
180k,0.1471,,0.2249,0.236
185k,0.1499,,0.2222,0.2366
190k,0.1504,,,0.2274
195k,0.1554,,,0.2454
200k,0.1565,,,0.2346
205k,0.1726,,,0.2316
210k,0.1623,,,0.2493
215k,0.1576,,,0.2355
220k,0.1693,,,0.2427
225k,0.1596,,,0.244
230k,0.1693,,,0.2554
235k,0.172,,,0.2535
240k,0.1712,,,
245k,0.1704,,,
250k,0.1784,,,
255k,0.174,,,
260k,0.1756,,,
265k,0.1886,,,
270k,0.182,,,
275k,0.187,,,
280k,0.1704,,,
285k,0.1903,,,
290k,,,,
300k,,,,
305k,,,,
310k,,,,
315k,,,,
320k,,,,
325k,,,,
330k,,,,
335k,,,,