Training in progress, step 14791, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 191968
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:90032cf0dc9c856d41890f86f4685dace3e1d08c43dcf395cfb0abeb905dacff
|
| 3 |
size 191968
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 401098
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aad103e7733612d5e8f27af84dac1a020ea6533850754af991474ed8c8362f9d
|
| 3 |
size 401098
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7125b2b44ded5a58c206b318b0c5bdacc79616c40eb566c3d2a91235dce6447
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:342aa3cf2d92cb0b41d8671fb70ba6c07fc707a0a46c98b22d4a2df487835f64
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 10.318995475769043,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-12658",
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 12658,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1794,6 +1794,300 @@
|
|
| 1794 |
"eval_samples_per_second": 216.168,
|
| 1795 |
"eval_steps_per_second": 54.064,
|
| 1796 |
"step": 12658
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1797 |
}
|
| 1798 |
],
|
| 1799 |
"logging_steps": 50,
|
|
@@ -1817,12 +2111,12 @@
|
|
| 1817 |
"should_evaluate": false,
|
| 1818 |
"should_log": false,
|
| 1819 |
"should_save": true,
|
| 1820 |
-
"should_training_stop":
|
| 1821 |
},
|
| 1822 |
"attributes": {}
|
| 1823 |
}
|
| 1824 |
},
|
| 1825 |
-
"total_flos":
|
| 1826 |
"train_batch_size": 4,
|
| 1827 |
"trial_name": null,
|
| 1828 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 10.318995475769043,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-12658",
|
| 4 |
+
"epoch": 9.999830981154398,
|
| 5 |
"eval_steps": 12658,
|
| 6 |
+
"global_step": 14791,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1794 |
"eval_samples_per_second": 216.168,
|
| 1795 |
"eval_steps_per_second": 54.064,
|
| 1796 |
"step": 12658
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 8.586157356545256,
|
| 1800 |
+
"grad_norm": 0.04512249305844307,
|
| 1801 |
+
"learning_rate": 1.0667350889148373e-05,
|
| 1802 |
+
"loss": 10.3284,
|
| 1803 |
+
"step": 12700
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 8.619961125665512,
|
| 1807 |
+
"grad_norm": 0.04324914887547493,
|
| 1808 |
+
"learning_rate": 1.0171325835719605e-05,
|
| 1809 |
+
"loss": 10.328,
|
| 1810 |
+
"step": 12750
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"epoch": 8.653764894785768,
|
| 1814 |
+
"grad_norm": 0.04108593612909317,
|
| 1815 |
+
"learning_rate": 9.686541992842219e-06,
|
| 1816 |
+
"loss": 10.3283,
|
| 1817 |
+
"step": 12800
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 8.687568663906026,
|
| 1821 |
+
"grad_norm": 0.04161646589636803,
|
| 1822 |
+
"learning_rate": 9.213054782368384e-06,
|
| 1823 |
+
"loss": 10.3278,
|
| 1824 |
+
"step": 12850
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 8.721372433026282,
|
| 1828 |
+
"grad_norm": 0.0430586077272892,
|
| 1829 |
+
"learning_rate": 8.750918334687578e-06,
|
| 1830 |
+
"loss": 10.3283,
|
| 1831 |
+
"step": 12900
|
| 1832 |
+
},
|
| 1833 |
+
{
|
| 1834 |
+
"epoch": 8.75517620214654,
|
| 1835 |
+
"grad_norm": 0.042888566851615906,
|
| 1836 |
+
"learning_rate": 8.300185482538152e-06,
|
| 1837 |
+
"loss": 10.3285,
|
| 1838 |
+
"step": 12950
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 8.788979971266796,
|
| 1842 |
+
"grad_norm": 0.041527193039655685,
|
| 1843 |
+
"learning_rate": 7.860907754967288e-06,
|
| 1844 |
+
"loss": 10.3281,
|
| 1845 |
+
"step": 13000
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 8.822783740387052,
|
| 1849 |
+
"grad_norm": 0.04209504276514053,
|
| 1850 |
+
"learning_rate": 7.433135371440134e-06,
|
| 1851 |
+
"loss": 10.3281,
|
| 1852 |
+
"step": 13050
|
| 1853 |
+
},
|
| 1854 |
+
{
|
| 1855 |
+
"epoch": 8.85658750950731,
|
| 1856 |
+
"grad_norm": 0.04120711609721184,
|
| 1857 |
+
"learning_rate": 7.016917236098581e-06,
|
| 1858 |
+
"loss": 10.3284,
|
| 1859 |
+
"step": 13100
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 8.890391278627567,
|
| 1863 |
+
"grad_norm": 0.0413849912583828,
|
| 1864 |
+
"learning_rate": 6.612300932170353e-06,
|
| 1865 |
+
"loss": 10.3283,
|
| 1866 |
+
"step": 13150
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 8.924195047747824,
|
| 1870 |
+
"grad_norm": 0.04532284289598465,
|
| 1871 |
+
"learning_rate": 6.21933271652917e-06,
|
| 1872 |
+
"loss": 10.3282,
|
| 1873 |
+
"step": 13200
|
| 1874 |
+
},
|
| 1875 |
+
{
|
| 1876 |
+
"epoch": 8.95799881686808,
|
| 1877 |
+
"grad_norm": 0.04391855373978615,
|
| 1878 |
+
"learning_rate": 5.8380575144065715e-06,
|
| 1879 |
+
"loss": 10.3282,
|
| 1880 |
+
"step": 13250
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 8.991802585988339,
|
| 1884 |
+
"grad_norm": 0.04782425984740257,
|
| 1885 |
+
"learning_rate": 5.4685189142558565e-06,
|
| 1886 |
+
"loss": 10.3278,
|
| 1887 |
+
"step": 13300
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 9.025606355108595,
|
| 1891 |
+
"grad_norm": 0.06290362030267715,
|
| 1892 |
+
"learning_rate": 5.1107591627690535e-06,
|
| 1893 |
+
"loss": 10.5103,
|
| 1894 |
+
"step": 13350
|
| 1895 |
+
},
|
| 1896 |
+
{
|
| 1897 |
+
"epoch": 9.059410124228851,
|
| 1898 |
+
"grad_norm": 0.058358874171972275,
|
| 1899 |
+
"learning_rate": 4.764819160047029e-06,
|
| 1900 |
+
"loss": 10.331,
|
| 1901 |
+
"step": 13400
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 9.093213893349109,
|
| 1905 |
+
"grad_norm": 0.061768610030412674,
|
| 1906 |
+
"learning_rate": 4.4307384549237645e-06,
|
| 1907 |
+
"loss": 10.3337,
|
| 1908 |
+
"step": 13450
|
| 1909 |
+
},
|
| 1910 |
+
{
|
| 1911 |
+
"epoch": 9.127017662469365,
|
| 1912 |
+
"grad_norm": 0.05595018342137337,
|
| 1913 |
+
"learning_rate": 4.108555240444934e-06,
|
| 1914 |
+
"loss": 10.3171,
|
| 1915 |
+
"step": 13500
|
| 1916 |
+
},
|
| 1917 |
+
{
|
| 1918 |
+
"epoch": 9.160821431589623,
|
| 1919 |
+
"grad_norm": 0.05736524239182472,
|
| 1920 |
+
"learning_rate": 3.7983063495017033e-06,
|
| 1921 |
+
"loss": 10.3269,
|
| 1922 |
+
"step": 13550
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 9.19462520070988,
|
| 1926 |
+
"grad_norm": 0.05978742614388466,
|
| 1927 |
+
"learning_rate": 3.500027250619767e-06,
|
| 1928 |
+
"loss": 10.3274,
|
| 1929 |
+
"step": 13600
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 9.228428969830135,
|
| 1933 |
+
"grad_norm": 0.06201018765568733,
|
| 1934 |
+
"learning_rate": 3.2137520439045086e-06,
|
| 1935 |
+
"loss": 10.3289,
|
| 1936 |
+
"step": 13650
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 9.262232738950393,
|
| 1940 |
+
"grad_norm": 0.057245831936597824,
|
| 1941 |
+
"learning_rate": 2.9395134571426653e-06,
|
| 1942 |
+
"loss": 10.328,
|
| 1943 |
+
"step": 13700
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 9.29603650807065,
|
| 1947 |
+
"grad_norm": 0.05930756404995918,
|
| 1948 |
+
"learning_rate": 2.6773428420606484e-06,
|
| 1949 |
+
"loss": 10.3296,
|
| 1950 |
+
"step": 13750
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 9.329840277190907,
|
| 1954 |
+
"grad_norm": 0.0608067624270916,
|
| 1955 |
+
"learning_rate": 2.4272701707404724e-06,
|
| 1956 |
+
"loss": 10.3306,
|
| 1957 |
+
"step": 13800
|
| 1958 |
+
},
|
| 1959 |
+
{
|
| 1960 |
+
"epoch": 9.363644046311164,
|
| 1961 |
+
"grad_norm": 0.05809257924556732,
|
| 1962 |
+
"learning_rate": 2.1893240321931975e-06,
|
| 1963 |
+
"loss": 10.3233,
|
| 1964 |
+
"step": 13850
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 9.39744781543142,
|
| 1968 |
+
"grad_norm": 0.061754010617733,
|
| 1969 |
+
"learning_rate": 1.963531629090508e-06,
|
| 1970 |
+
"loss": 10.332,
|
| 1971 |
+
"step": 13900
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 9.431251584551678,
|
| 1975 |
+
"grad_norm": 0.06163324415683746,
|
| 1976 |
+
"learning_rate": 1.749918774654905e-06,
|
| 1977 |
+
"loss": 10.3353,
|
| 1978 |
+
"step": 13950
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 9.465055353671934,
|
| 1982 |
+
"grad_norm": 0.06020689383149147,
|
| 1983 |
+
"learning_rate": 1.548509889708657e-06,
|
| 1984 |
+
"loss": 10.3174,
|
| 1985 |
+
"step": 14000
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 9.498859122792192,
|
| 1989 |
+
"grad_norm": 0.05851924419403076,
|
| 1990 |
+
"learning_rate": 1.3593279998818836e-06,
|
| 1991 |
+
"loss": 10.3277,
|
| 1992 |
+
"step": 14050
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 9.532662891912448,
|
| 1996 |
+
"grad_norm": 0.054948098957538605,
|
| 1997 |
+
"learning_rate": 1.182394732980232e-06,
|
| 1998 |
+
"loss": 10.3299,
|
| 1999 |
+
"step": 14100
|
| 2000 |
+
},
|
| 2001 |
+
{
|
| 2002 |
+
"epoch": 9.566466661032706,
|
| 2003 |
+
"grad_norm": 0.06071977689862251,
|
| 2004 |
+
"learning_rate": 1.0177303165123553e-06,
|
| 2005 |
+
"loss": 10.3287,
|
| 2006 |
+
"step": 14150
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 9.600270430152962,
|
| 2010 |
+
"grad_norm": 0.06051783263683319,
|
| 2011 |
+
"learning_rate": 8.653535753774341e-07,
|
| 2012 |
+
"loss": 10.3278,
|
| 2013 |
+
"step": 14200
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 9.634074199273218,
|
| 2017 |
+
"grad_norm": 0.060153182595968246,
|
| 2018 |
+
"learning_rate": 7.252819297130155e-07,
|
| 2019 |
+
"loss": 10.3266,
|
| 2020 |
+
"step": 14250
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 9.667877968393476,
|
| 2024 |
+
"grad_norm": 0.06272024661302567,
|
| 2025 |
+
"learning_rate": 5.975313929035335e-07,
|
| 2026 |
+
"loss": 10.339,
|
| 2027 |
+
"step": 14300
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 9.701681737513733,
|
| 2031 |
+
"grad_norm": 0.06360357999801636,
|
| 2032 |
+
"learning_rate": 4.821165697496309e-07,
|
| 2033 |
+
"loss": 10.3309,
|
| 2034 |
+
"step": 14350
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 9.73548550663399,
|
| 2038 |
+
"grad_norm": 0.059037111699581146,
|
| 2039 |
+
"learning_rate": 3.7905065479846173e-07,
|
| 2040 |
+
"loss": 10.3234,
|
| 2041 |
+
"step": 14400
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 9.769289275754247,
|
| 2045 |
+
"grad_norm": 0.0558004267513752,
|
| 2046 |
+
"learning_rate": 2.883454308352653e-07,
|
| 2047 |
+
"loss": 10.3263,
|
| 2048 |
+
"step": 14450
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 9.803093044874503,
|
| 2052 |
+
"grad_norm": 0.05941386893391609,
|
| 2053 |
+
"learning_rate": 2.100112675363192e-07,
|
| 2054 |
+
"loss": 10.3226,
|
| 2055 |
+
"step": 14500
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 9.83689681399476,
|
| 2059 |
+
"grad_norm": 0.0531686469912529,
|
| 2060 |
+
"learning_rate": 1.4405712028346407e-07,
|
| 2061 |
+
"loss": 10.3312,
|
| 2062 |
+
"step": 14550
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 9.870700583115017,
|
| 2066 |
+
"grad_norm": 0.05972735211253166,
|
| 2067 |
+
"learning_rate": 9.049052914030886e-08,
|
| 2068 |
+
"loss": 10.328,
|
| 2069 |
+
"step": 14600
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 9.904504352235275,
|
| 2073 |
+
"grad_norm": 0.06279122829437256,
|
| 2074 |
+
"learning_rate": 4.93176179901767e-08,
|
| 2075 |
+
"loss": 10.3326,
|
| 2076 |
+
"step": 14650
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 9.938308121355531,
|
| 2080 |
+
"grad_norm": 0.05874359980225563,
|
| 2081 |
+
"learning_rate": 2.054309383605618e-08,
|
| 2082 |
+
"loss": 10.3212,
|
| 2083 |
+
"step": 14700
|
| 2084 |
+
},
|
| 2085 |
+
{
|
| 2086 |
+
"epoch": 9.972111890475787,
|
| 2087 |
+
"grad_norm": 0.060345474630594254,
|
| 2088 |
+
"learning_rate": 4.170246262474031e-09,
|
| 2089 |
+
"loss": 10.3283,
|
| 2090 |
+
"step": 14750
|
| 2091 |
}
|
| 2092 |
],
|
| 2093 |
"logging_steps": 50,
|
|
|
|
| 2111 |
"should_evaluate": false,
|
| 2112 |
"should_log": false,
|
| 2113 |
"should_save": true,
|
| 2114 |
+
"should_training_stop": true
|
| 2115 |
},
|
| 2116 |
"attributes": {}
|
| 2117 |
}
|
| 2118 |
},
|
| 2119 |
+
"total_flos": 2158745813876736.0,
|
| 2120 |
"train_batch_size": 4,
|
| 2121 |
"trial_name": null,
|
| 2122 |
"trial_params": null
|