Training in progress, step 300, 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 97307544
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f3dd2c7fef6854dded71270b43f0e219e0e385e468e3866c850a76fffaf329d1
|
| 3 |
size 97307544
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 194840426
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eb7650112bd10ab3e207424fb68c73e772429116566efa506ced1f9aa984ef52
|
| 3 |
size 194840426
|
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:dc14897b4c8d12b2fd679afea6e33ff956fb43524baeeaef8422be9b77027753
|
| 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:a965b4f99c848f0b9008cc4d4102df26407e8276df19747f112091a2e2ab78f1
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 1.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 25,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1845,6 +1845,372 @@
|
|
| 1845 |
"eval_samples_per_second": 8.366,
|
| 1846 |
"eval_steps_per_second": 2.844,
|
| 1847 |
"step": 250
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1848 |
}
|
| 1849 |
],
|
| 1850 |
"logging_steps": 1,
|
|
@@ -1873,7 +2239,7 @@
|
|
| 1873 |
"attributes": {}
|
| 1874 |
}
|
| 1875 |
},
|
| 1876 |
-
"total_flos":
|
| 1877 |
"train_batch_size": 3,
|
| 1878 |
"trial_name": null,
|
| 1879 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 1.07566499710083,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-300",
|
| 4 |
+
"epoch": 0.17410481109627995,
|
| 5 |
"eval_steps": 25,
|
| 6 |
+
"global_step": 300,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1845 |
"eval_samples_per_second": 8.366,
|
| 1846 |
"eval_steps_per_second": 2.844,
|
| 1847 |
"step": 250
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 0.14566769195055423,
|
| 1851 |
+
"grad_norm": 0.3264401853084564,
|
| 1852 |
+
"learning_rate": 1.7341257188154625e-05,
|
| 1853 |
+
"loss": 1.2756,
|
| 1854 |
+
"step": 251
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 0.14624804132087516,
|
| 1858 |
+
"grad_norm": 0.3468761146068573,
|
| 1859 |
+
"learning_rate": 1.720347135357937e-05,
|
| 1860 |
+
"loss": 1.2133,
|
| 1861 |
+
"step": 252
|
| 1862 |
+
},
|
| 1863 |
+
{
|
| 1864 |
+
"epoch": 0.1468283906911961,
|
| 1865 |
+
"grad_norm": 0.41729360818862915,
|
| 1866 |
+
"learning_rate": 1.7065987398863007e-05,
|
| 1867 |
+
"loss": 1.3962,
|
| 1868 |
+
"step": 253
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 0.14740874006151702,
|
| 1872 |
+
"grad_norm": 0.40159016847610474,
|
| 1873 |
+
"learning_rate": 1.6928813976854267e-05,
|
| 1874 |
+
"loss": 1.2083,
|
| 1875 |
+
"step": 254
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 0.14798908943183797,
|
| 1879 |
+
"grad_norm": 0.43321022391319275,
|
| 1880 |
+
"learning_rate": 1.6791959720857855e-05,
|
| 1881 |
+
"loss": 1.2977,
|
| 1882 |
+
"step": 255
|
| 1883 |
+
},
|
| 1884 |
+
{
|
| 1885 |
+
"epoch": 0.1485694388021589,
|
| 1886 |
+
"grad_norm": 0.4180416464805603,
|
| 1887 |
+
"learning_rate": 1.665543324409106e-05,
|
| 1888 |
+
"loss": 1.1649,
|
| 1889 |
+
"step": 256
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 0.14914978817247984,
|
| 1893 |
+
"grad_norm": 0.48152637481689453,
|
| 1894 |
+
"learning_rate": 1.651924313914168e-05,
|
| 1895 |
+
"loss": 1.2545,
|
| 1896 |
+
"step": 257
|
| 1897 |
+
},
|
| 1898 |
+
{
|
| 1899 |
+
"epoch": 0.14973013754280076,
|
| 1900 |
+
"grad_norm": 0.4317726194858551,
|
| 1901 |
+
"learning_rate": 1.6383397977427237e-05,
|
| 1902 |
+
"loss": 1.1986,
|
| 1903 |
+
"step": 258
|
| 1904 |
+
},
|
| 1905 |
+
{
|
| 1906 |
+
"epoch": 0.1503104869131217,
|
| 1907 |
+
"grad_norm": 0.47157958149909973,
|
| 1908 |
+
"learning_rate": 1.6247906308655506e-05,
|
| 1909 |
+
"loss": 1.3727,
|
| 1910 |
+
"step": 259
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 0.15089083628344263,
|
| 1914 |
+
"grad_norm": 0.4798511862754822,
|
| 1915 |
+
"learning_rate": 1.611277666028641e-05,
|
| 1916 |
+
"loss": 1.2121,
|
| 1917 |
+
"step": 260
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 0.15147118565376358,
|
| 1921 |
+
"grad_norm": 0.5214879512786865,
|
| 1922 |
+
"learning_rate": 1.5978017536995366e-05,
|
| 1923 |
+
"loss": 1.2163,
|
| 1924 |
+
"step": 261
|
| 1925 |
+
},
|
| 1926 |
+
{
|
| 1927 |
+
"epoch": 0.1520515350240845,
|
| 1928 |
+
"grad_norm": 0.5045842528343201,
|
| 1929 |
+
"learning_rate": 1.5843637420137964e-05,
|
| 1930 |
+
"loss": 1.2251,
|
| 1931 |
+
"step": 262
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 0.15263188439440542,
|
| 1935 |
+
"grad_norm": 0.5405669212341309,
|
| 1936 |
+
"learning_rate": 1.5709644767216233e-05,
|
| 1937 |
+
"loss": 1.3358,
|
| 1938 |
+
"step": 263
|
| 1939 |
+
},
|
| 1940 |
+
{
|
| 1941 |
+
"epoch": 0.15321223376472637,
|
| 1942 |
+
"grad_norm": 0.51566481590271,
|
| 1943 |
+
"learning_rate": 1.557604801134631e-05,
|
| 1944 |
+
"loss": 1.2341,
|
| 1945 |
+
"step": 264
|
| 1946 |
+
},
|
| 1947 |
+
{
|
| 1948 |
+
"epoch": 0.1537925831350473,
|
| 1949 |
+
"grad_norm": 0.5638424158096313,
|
| 1950 |
+
"learning_rate": 1.54428555607277e-05,
|
| 1951 |
+
"loss": 1.1778,
|
| 1952 |
+
"step": 265
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 0.15437293250536824,
|
| 1956 |
+
"grad_norm": 0.549906313419342,
|
| 1957 |
+
"learning_rate": 1.5310075798114107e-05,
|
| 1958 |
+
"loss": 1.213,
|
| 1959 |
+
"step": 266
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 0.15495328187568916,
|
| 1963 |
+
"grad_norm": 0.5369941592216492,
|
| 1964 |
+
"learning_rate": 1.5177717080285796e-05,
|
| 1965 |
+
"loss": 1.1816,
|
| 1966 |
+
"step": 267
|
| 1967 |
+
},
|
| 1968 |
+
{
|
| 1969 |
+
"epoch": 0.1555336312460101,
|
| 1970 |
+
"grad_norm": 0.5978056788444519,
|
| 1971 |
+
"learning_rate": 1.5045787737523685e-05,
|
| 1972 |
+
"loss": 1.2219,
|
| 1973 |
+
"step": 268
|
| 1974 |
+
},
|
| 1975 |
+
{
|
| 1976 |
+
"epoch": 0.15611398061633103,
|
| 1977 |
+
"grad_norm": 0.5928854942321777,
|
| 1978 |
+
"learning_rate": 1.491429607308506e-05,
|
| 1979 |
+
"loss": 1.226,
|
| 1980 |
+
"step": 269
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 0.15669432998665198,
|
| 1984 |
+
"grad_norm": 0.6016438603401184,
|
| 1985 |
+
"learning_rate": 1.478325036268096e-05,
|
| 1986 |
+
"loss": 1.2603,
|
| 1987 |
+
"step": 270
|
| 1988 |
+
},
|
| 1989 |
+
{
|
| 1990 |
+
"epoch": 0.1572746793569729,
|
| 1991 |
+
"grad_norm": 0.6282125115394592,
|
| 1992 |
+
"learning_rate": 1.4652658853955349e-05,
|
| 1993 |
+
"loss": 1.2888,
|
| 1994 |
+
"step": 271
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 0.15785502872729384,
|
| 1998 |
+
"grad_norm": 0.6291938424110413,
|
| 1999 |
+
"learning_rate": 1.4522529765966048e-05,
|
| 2000 |
+
"loss": 1.1523,
|
| 2001 |
+
"step": 272
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 0.15843537809761477,
|
| 2005 |
+
"grad_norm": 0.6878090500831604,
|
| 2006 |
+
"learning_rate": 1.4392871288667415e-05,
|
| 2007 |
+
"loss": 1.2104,
|
| 2008 |
+
"step": 273
|
| 2009 |
+
},
|
| 2010 |
+
{
|
| 2011 |
+
"epoch": 0.15901572746793569,
|
| 2012 |
+
"grad_norm": 0.7197153568267822,
|
| 2013 |
+
"learning_rate": 1.4263691582394912e-05,
|
| 2014 |
+
"loss": 1.2435,
|
| 2015 |
+
"step": 274
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"epoch": 0.15959607683825663,
|
| 2019 |
+
"grad_norm": 0.7492052912712097,
|
| 2020 |
+
"learning_rate": 1.4134998777351533e-05,
|
| 2021 |
+
"loss": 1.1018,
|
| 2022 |
+
"step": 275
|
| 2023 |
+
},
|
| 2024 |
+
{
|
| 2025 |
+
"epoch": 0.15959607683825663,
|
| 2026 |
+
"eval_loss": 1.087742567062378,
|
| 2027 |
+
"eval_runtime": 5.9755,
|
| 2028 |
+
"eval_samples_per_second": 8.368,
|
| 2029 |
+
"eval_steps_per_second": 2.845,
|
| 2030 |
+
"step": 275
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 0.16017642620857755,
|
| 2034 |
+
"grad_norm": 0.7975369095802307,
|
| 2035 |
+
"learning_rate": 1.4006800973096066e-05,
|
| 2036 |
+
"loss": 1.124,
|
| 2037 |
+
"step": 276
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 0.1607567755788985,
|
| 2041 |
+
"grad_norm": 0.8354909420013428,
|
| 2042 |
+
"learning_rate": 1.3879106238033392e-05,
|
| 2043 |
+
"loss": 1.2082,
|
| 2044 |
+
"step": 277
|
| 2045 |
+
},
|
| 2046 |
+
{
|
| 2047 |
+
"epoch": 0.16133712494921942,
|
| 2048 |
+
"grad_norm": 0.7988197803497314,
|
| 2049 |
+
"learning_rate": 1.3751922608906614e-05,
|
| 2050 |
+
"loss": 1.0521,
|
| 2051 |
+
"step": 278
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"epoch": 0.16191747431954037,
|
| 2055 |
+
"grad_norm": 1.0019845962524414,
|
| 2056 |
+
"learning_rate": 1.3625258090291312e-05,
|
| 2057 |
+
"loss": 1.2305,
|
| 2058 |
+
"step": 279
|
| 2059 |
+
},
|
| 2060 |
+
{
|
| 2061 |
+
"epoch": 0.1624978236898613,
|
| 2062 |
+
"grad_norm": 0.8843457102775574,
|
| 2063 |
+
"learning_rate": 1.3499120654091716e-05,
|
| 2064 |
+
"loss": 1.1107,
|
| 2065 |
+
"step": 280
|
| 2066 |
+
},
|
| 2067 |
+
{
|
| 2068 |
+
"epoch": 0.16307817306018224,
|
| 2069 |
+
"grad_norm": 1.0474600791931152,
|
| 2070 |
+
"learning_rate": 1.3373518239038985e-05,
|
| 2071 |
+
"loss": 0.8914,
|
| 2072 |
+
"step": 281
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 0.16365852243050316,
|
| 2076 |
+
"grad_norm": 1.0813342332839966,
|
| 2077 |
+
"learning_rate": 1.3248458750191566e-05,
|
| 2078 |
+
"loss": 1.265,
|
| 2079 |
+
"step": 282
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 0.1642388718008241,
|
| 2083 |
+
"grad_norm": 1.2061476707458496,
|
| 2084 |
+
"learning_rate": 1.3123950058437696e-05,
|
| 2085 |
+
"loss": 1.1564,
|
| 2086 |
+
"step": 283
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 0.16481922117114503,
|
| 2090 |
+
"grad_norm": 1.3668428659439087,
|
| 2091 |
+
"learning_rate": 1.3000000000000006e-05,
|
| 2092 |
+
"loss": 0.9441,
|
| 2093 |
+
"step": 284
|
| 2094 |
+
},
|
| 2095 |
+
{
|
| 2096 |
+
"epoch": 0.16539957054146595,
|
| 2097 |
+
"grad_norm": 1.4264423847198486,
|
| 2098 |
+
"learning_rate": 1.2876616375942311e-05,
|
| 2099 |
+
"loss": 1.0423,
|
| 2100 |
+
"step": 285
|
| 2101 |
+
},
|
| 2102 |
+
{
|
| 2103 |
+
"epoch": 0.1659799199117869,
|
| 2104 |
+
"grad_norm": 1.2884799242019653,
|
| 2105 |
+
"learning_rate": 1.2753806951678694e-05,
|
| 2106 |
+
"loss": 1.0806,
|
| 2107 |
+
"step": 286
|
| 2108 |
+
},
|
| 2109 |
+
{
|
| 2110 |
+
"epoch": 0.16656026928210782,
|
| 2111 |
+
"grad_norm": 1.1869494915008545,
|
| 2112 |
+
"learning_rate": 1.2631579456484704e-05,
|
| 2113 |
+
"loss": 1.0565,
|
| 2114 |
+
"step": 287
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 0.16714061865242877,
|
| 2118 |
+
"grad_norm": 1.4182007312774658,
|
| 2119 |
+
"learning_rate": 1.2509941583010959e-05,
|
| 2120 |
+
"loss": 0.9715,
|
| 2121 |
+
"step": 288
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 0.1677209680227497,
|
| 2125 |
+
"grad_norm": 1.5443942546844482,
|
| 2126 |
+
"learning_rate": 1.2388900986798953e-05,
|
| 2127 |
+
"loss": 1.056,
|
| 2128 |
+
"step": 289
|
| 2129 |
+
},
|
| 2130 |
+
{
|
| 2131 |
+
"epoch": 0.16830131739307064,
|
| 2132 |
+
"grad_norm": 1.4688053131103516,
|
| 2133 |
+
"learning_rate": 1.226846528579925e-05,
|
| 2134 |
+
"loss": 1.064,
|
| 2135 |
+
"step": 290
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 0.16888166676339156,
|
| 2139 |
+
"grad_norm": 1.9034485816955566,
|
| 2140 |
+
"learning_rate": 1.2148642059892022e-05,
|
| 2141 |
+
"loss": 1.4109,
|
| 2142 |
+
"step": 291
|
| 2143 |
+
},
|
| 2144 |
+
{
|
| 2145 |
+
"epoch": 0.1694620161337125,
|
| 2146 |
+
"grad_norm": 1.4713436365127563,
|
| 2147 |
+
"learning_rate": 1.2029438850410018e-05,
|
| 2148 |
+
"loss": 0.7642,
|
| 2149 |
+
"step": 292
|
| 2150 |
+
},
|
| 2151 |
+
{
|
| 2152 |
+
"epoch": 0.17004236550403343,
|
| 2153 |
+
"grad_norm": 1.6940679550170898,
|
| 2154 |
+
"learning_rate": 1.1910863159663908e-05,
|
| 2155 |
+
"loss": 1.0706,
|
| 2156 |
+
"step": 293
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 0.17062271487435435,
|
| 2160 |
+
"grad_norm": 1.5762420892715454,
|
| 2161 |
+
"learning_rate": 1.179292245047013e-05,
|
| 2162 |
+
"loss": 0.8392,
|
| 2163 |
+
"step": 294
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 0.1712030642446753,
|
| 2167 |
+
"grad_norm": 2.418851613998413,
|
| 2168 |
+
"learning_rate": 1.1675624145681177e-05,
|
| 2169 |
+
"loss": 0.8652,
|
| 2170 |
+
"step": 295
|
| 2171 |
+
},
|
| 2172 |
+
{
|
| 2173 |
+
"epoch": 0.17178341361499622,
|
| 2174 |
+
"grad_norm": 2.076580762863159,
|
| 2175 |
+
"learning_rate": 1.1558975627718437e-05,
|
| 2176 |
+
"loss": 1.0094,
|
| 2177 |
+
"step": 296
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 0.17236376298531716,
|
| 2181 |
+
"grad_norm": 2.1403160095214844,
|
| 2182 |
+
"learning_rate": 1.1442984238107577e-05,
|
| 2183 |
+
"loss": 0.8598,
|
| 2184 |
+
"step": 297
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 0.17294411235563809,
|
| 2188 |
+
"grad_norm": 2.948251962661743,
|
| 2189 |
+
"learning_rate": 1.1327657277016475e-05,
|
| 2190 |
+
"loss": 1.1042,
|
| 2191 |
+
"step": 298
|
| 2192 |
+
},
|
| 2193 |
+
{
|
| 2194 |
+
"epoch": 0.17352446172595903,
|
| 2195 |
+
"grad_norm": 4.150631904602051,
|
| 2196 |
+
"learning_rate": 1.1213002002795757e-05,
|
| 2197 |
+
"loss": 1.4053,
|
| 2198 |
+
"step": 299
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 0.17410481109627995,
|
| 2202 |
+
"grad_norm": 6.424019813537598,
|
| 2203 |
+
"learning_rate": 1.1099025631522004e-05,
|
| 2204 |
+
"loss": 1.8729,
|
| 2205 |
+
"step": 300
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 0.17410481109627995,
|
| 2209 |
+
"eval_loss": 1.07566499710083,
|
| 2210 |
+
"eval_runtime": 5.9749,
|
| 2211 |
+
"eval_samples_per_second": 8.368,
|
| 2212 |
+
"eval_steps_per_second": 2.845,
|
| 2213 |
+
"step": 300
|
| 2214 |
}
|
| 2215 |
],
|
| 2216 |
"logging_steps": 1,
|
|
|
|
| 2239 |
"attributes": {}
|
| 2240 |
}
|
| 2241 |
},
|
| 2242 |
+
"total_flos": 3.14419829538816e+17,
|
| 2243 |
"train_batch_size": 3,
|
| 2244 |
"trial_name": null,
|
| 2245 |
"trial_params": null
|