Training in progress, step 8208, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 290403936
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:51c63f2ddd7a6e10a97d28920b370b15420c9cbf7029f8fda3d06ba1218f04c1
|
| 3 |
size 290403936
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 574811979
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:57d4591400ba60c7024d3e4f2557c0eec73519bb30ceefe230c80e60c2fcbe08
|
| 3 |
size 574811979
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14645
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c80acdc2532ec7797b634e601103d59024528d38f5012f3d98746fee880405d0
|
| 3 |
size 14645
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9fac151c2fc0bed5786af9225b2875725cef39eda30f3cfd01728ae1cd6a89cd
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric":
|
| 3 |
-
"best_model_checkpoint": "./whissm-plus/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 2052,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1756,6 +1756,589 @@
|
|
| 1756 |
"eval_steps_per_second": 0.338,
|
| 1757 |
"eval_wer": 16.973751274209988,
|
| 1758 |
"step": 6156
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1759 |
}
|
| 1760 |
],
|
| 1761 |
"logging_steps": 25,
|
|
@@ -1770,12 +2353,12 @@
|
|
| 1770 |
"should_evaluate": false,
|
| 1771 |
"should_log": false,
|
| 1772 |
"should_save": true,
|
| 1773 |
-
"should_training_stop":
|
| 1774 |
},
|
| 1775 |
"attributes": {}
|
| 1776 |
}
|
| 1777 |
},
|
| 1778 |
-
"total_flos":
|
| 1779 |
"train_batch_size": 4,
|
| 1780 |
"trial_name": null,
|
| 1781 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 14.794215086646279,
|
| 3 |
+
"best_model_checkpoint": "./whissm-plus/checkpoint-8208",
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
"eval_steps": 2052,
|
| 6 |
+
"global_step": 8208,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1756 |
"eval_steps_per_second": 0.338,
|
| 1757 |
"eval_wer": 16.973751274209988,
|
| 1758 |
"step": 6156
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 0.7523148148148148,
|
| 1762 |
+
"grad_norm": 6.905988693237305,
|
| 1763 |
+
"learning_rate": 2.64270887389725e-06,
|
| 1764 |
+
"loss": 0.0611,
|
| 1765 |
+
"step": 6175
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"epoch": 0.7553606237816765,
|
| 1769 |
+
"grad_norm": 5.302619457244873,
|
| 1770 |
+
"learning_rate": 2.610275038920602e-06,
|
| 1771 |
+
"loss": 0.0487,
|
| 1772 |
+
"step": 6200
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 0.758406432748538,
|
| 1776 |
+
"grad_norm": 2.7596232891082764,
|
| 1777 |
+
"learning_rate": 2.5778412039439545e-06,
|
| 1778 |
+
"loss": 0.0431,
|
| 1779 |
+
"step": 6225
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 0.7614522417153996,
|
| 1783 |
+
"grad_norm": 2.1329431533813477,
|
| 1784 |
+
"learning_rate": 2.545407368967307e-06,
|
| 1785 |
+
"loss": 0.0511,
|
| 1786 |
+
"step": 6250
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"epoch": 0.7644980506822612,
|
| 1790 |
+
"grad_norm": 5.41237211227417,
|
| 1791 |
+
"learning_rate": 2.512973533990659e-06,
|
| 1792 |
+
"loss": 0.059,
|
| 1793 |
+
"step": 6275
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 0.7675438596491229,
|
| 1797 |
+
"grad_norm": 4.981463432312012,
|
| 1798 |
+
"learning_rate": 2.4805396990140115e-06,
|
| 1799 |
+
"loss": 0.0545,
|
| 1800 |
+
"step": 6300
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 0.7705896686159844,
|
| 1804 |
+
"grad_norm": 2.5017101764678955,
|
| 1805 |
+
"learning_rate": 2.448105864037364e-06,
|
| 1806 |
+
"loss": 0.0515,
|
| 1807 |
+
"step": 6325
|
| 1808 |
+
},
|
| 1809 |
+
{
|
| 1810 |
+
"epoch": 0.773635477582846,
|
| 1811 |
+
"grad_norm": 2.7259082794189453,
|
| 1812 |
+
"learning_rate": 2.4156720290607163e-06,
|
| 1813 |
+
"loss": 0.0458,
|
| 1814 |
+
"step": 6350
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 0.7766812865497076,
|
| 1818 |
+
"grad_norm": 5.928169250488281,
|
| 1819 |
+
"learning_rate": 2.3832381940840686e-06,
|
| 1820 |
+
"loss": 0.0556,
|
| 1821 |
+
"step": 6375
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 0.7797270955165692,
|
| 1825 |
+
"grad_norm": 4.780055999755859,
|
| 1826 |
+
"learning_rate": 2.350804359107421e-06,
|
| 1827 |
+
"loss": 0.0474,
|
| 1828 |
+
"step": 6400
|
| 1829 |
+
},
|
| 1830 |
+
{
|
| 1831 |
+
"epoch": 0.7827729044834308,
|
| 1832 |
+
"grad_norm": 6.110754013061523,
|
| 1833 |
+
"learning_rate": 2.3183705241307733e-06,
|
| 1834 |
+
"loss": 0.0495,
|
| 1835 |
+
"step": 6425
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 0.7858187134502924,
|
| 1839 |
+
"grad_norm": 5.388917922973633,
|
| 1840 |
+
"learning_rate": 2.2859366891541257e-06,
|
| 1841 |
+
"loss": 0.0483,
|
| 1842 |
+
"step": 6450
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 0.788864522417154,
|
| 1846 |
+
"grad_norm": 3.1466193199157715,
|
| 1847 |
+
"learning_rate": 2.253502854177478e-06,
|
| 1848 |
+
"loss": 0.0435,
|
| 1849 |
+
"step": 6475
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 0.7919103313840156,
|
| 1853 |
+
"grad_norm": 2.081556797027588,
|
| 1854 |
+
"learning_rate": 2.2210690192008304e-06,
|
| 1855 |
+
"loss": 0.0644,
|
| 1856 |
+
"step": 6500
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 0.7949561403508771,
|
| 1860 |
+
"grad_norm": 4.22133731842041,
|
| 1861 |
+
"learning_rate": 2.1886351842241828e-06,
|
| 1862 |
+
"loss": 0.0466,
|
| 1863 |
+
"step": 6525
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 0.7980019493177388,
|
| 1867 |
+
"grad_norm": 2.9023261070251465,
|
| 1868 |
+
"learning_rate": 2.156201349247535e-06,
|
| 1869 |
+
"loss": 0.0486,
|
| 1870 |
+
"step": 6550
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 0.8010477582846004,
|
| 1874 |
+
"grad_norm": 5.409677982330322,
|
| 1875 |
+
"learning_rate": 2.1237675142708875e-06,
|
| 1876 |
+
"loss": 0.0402,
|
| 1877 |
+
"step": 6575
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 0.804093567251462,
|
| 1881 |
+
"grad_norm": 4.244589328765869,
|
| 1882 |
+
"learning_rate": 2.09133367929424e-06,
|
| 1883 |
+
"loss": 0.0557,
|
| 1884 |
+
"step": 6600
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 0.8071393762183235,
|
| 1888 |
+
"grad_norm": 8.207276344299316,
|
| 1889 |
+
"learning_rate": 2.058899844317592e-06,
|
| 1890 |
+
"loss": 0.0537,
|
| 1891 |
+
"step": 6625
|
| 1892 |
+
},
|
| 1893 |
+
{
|
| 1894 |
+
"epoch": 0.8101851851851852,
|
| 1895 |
+
"grad_norm": 3.6732988357543945,
|
| 1896 |
+
"learning_rate": 2.0264660093409446e-06,
|
| 1897 |
+
"loss": 0.0452,
|
| 1898 |
+
"step": 6650
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 0.8132309941520468,
|
| 1902 |
+
"grad_norm": 3.1428933143615723,
|
| 1903 |
+
"learning_rate": 1.994032174364297e-06,
|
| 1904 |
+
"loss": 0.0465,
|
| 1905 |
+
"step": 6675
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 0.8162768031189084,
|
| 1909 |
+
"grad_norm": 3.1963727474212646,
|
| 1910 |
+
"learning_rate": 1.9615983393876493e-06,
|
| 1911 |
+
"loss": 0.0479,
|
| 1912 |
+
"step": 6700
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 0.81932261208577,
|
| 1916 |
+
"grad_norm": 7.395984649658203,
|
| 1917 |
+
"learning_rate": 1.9291645044110016e-06,
|
| 1918 |
+
"loss": 0.0489,
|
| 1919 |
+
"step": 6725
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 0.8223684210526315,
|
| 1923 |
+
"grad_norm": 2.5216851234436035,
|
| 1924 |
+
"learning_rate": 1.896730669434354e-06,
|
| 1925 |
+
"loss": 0.044,
|
| 1926 |
+
"step": 6750
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 0.8254142300194932,
|
| 1930 |
+
"grad_norm": 2.955092668533325,
|
| 1931 |
+
"learning_rate": 1.8642968344577064e-06,
|
| 1932 |
+
"loss": 0.0483,
|
| 1933 |
+
"step": 6775
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"epoch": 0.8284600389863548,
|
| 1937 |
+
"grad_norm": 6.115810871124268,
|
| 1938 |
+
"learning_rate": 1.831862999481059e-06,
|
| 1939 |
+
"loss": 0.0512,
|
| 1940 |
+
"step": 6800
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 0.8315058479532164,
|
| 1944 |
+
"grad_norm": 3.0200977325439453,
|
| 1945 |
+
"learning_rate": 1.7994291645044113e-06,
|
| 1946 |
+
"loss": 0.0509,
|
| 1947 |
+
"step": 6825
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 0.8345516569200779,
|
| 1951 |
+
"grad_norm": 3.954040288925171,
|
| 1952 |
+
"learning_rate": 1.7669953295277634e-06,
|
| 1953 |
+
"loss": 0.0474,
|
| 1954 |
+
"step": 6850
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 0.8375974658869396,
|
| 1958 |
+
"grad_norm": 4.320163249969482,
|
| 1959 |
+
"learning_rate": 1.7345614945511158e-06,
|
| 1960 |
+
"loss": 0.0429,
|
| 1961 |
+
"step": 6875
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 0.8406432748538012,
|
| 1965 |
+
"grad_norm": 2.413540840148926,
|
| 1966 |
+
"learning_rate": 1.7021276595744682e-06,
|
| 1967 |
+
"loss": 0.0428,
|
| 1968 |
+
"step": 6900
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 0.8436890838206628,
|
| 1972 |
+
"grad_norm": 4.413974761962891,
|
| 1973 |
+
"learning_rate": 1.6696938245978205e-06,
|
| 1974 |
+
"loss": 0.0439,
|
| 1975 |
+
"step": 6925
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 0.8467348927875243,
|
| 1979 |
+
"grad_norm": 6.30503511428833,
|
| 1980 |
+
"learning_rate": 1.637259989621173e-06,
|
| 1981 |
+
"loss": 0.043,
|
| 1982 |
+
"step": 6950
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 0.8497807017543859,
|
| 1986 |
+
"grad_norm": 3.963675022125244,
|
| 1987 |
+
"learning_rate": 1.6048261546445255e-06,
|
| 1988 |
+
"loss": 0.0413,
|
| 1989 |
+
"step": 6975
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 0.8528265107212476,
|
| 1993 |
+
"grad_norm": 3.7029097080230713,
|
| 1994 |
+
"learning_rate": 1.5723923196678776e-06,
|
| 1995 |
+
"loss": 0.0387,
|
| 1996 |
+
"step": 7000
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 0.8558723196881092,
|
| 2000 |
+
"grad_norm": 4.904066562652588,
|
| 2001 |
+
"learning_rate": 1.53995848469123e-06,
|
| 2002 |
+
"loss": 0.046,
|
| 2003 |
+
"step": 7025
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 0.8589181286549707,
|
| 2007 |
+
"grad_norm": 2.4294283390045166,
|
| 2008 |
+
"learning_rate": 1.5075246497145823e-06,
|
| 2009 |
+
"loss": 0.0542,
|
| 2010 |
+
"step": 7050
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 0.8619639376218323,
|
| 2014 |
+
"grad_norm": 3.33664870262146,
|
| 2015 |
+
"learning_rate": 1.4750908147379347e-06,
|
| 2016 |
+
"loss": 0.0453,
|
| 2017 |
+
"step": 7075
|
| 2018 |
+
},
|
| 2019 |
+
{
|
| 2020 |
+
"epoch": 0.865009746588694,
|
| 2021 |
+
"grad_norm": 4.458865165710449,
|
| 2022 |
+
"learning_rate": 1.4426569797612872e-06,
|
| 2023 |
+
"loss": 0.0492,
|
| 2024 |
+
"step": 7100
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 0.8680555555555556,
|
| 2028 |
+
"grad_norm": 6.270244121551514,
|
| 2029 |
+
"learning_rate": 1.4102231447846396e-06,
|
| 2030 |
+
"loss": 0.0455,
|
| 2031 |
+
"step": 7125
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 0.8711013645224172,
|
| 2035 |
+
"grad_norm": 4.222137451171875,
|
| 2036 |
+
"learning_rate": 1.3777893098079918e-06,
|
| 2037 |
+
"loss": 0.0445,
|
| 2038 |
+
"step": 7150
|
| 2039 |
+
},
|
| 2040 |
+
{
|
| 2041 |
+
"epoch": 0.8741471734892787,
|
| 2042 |
+
"grad_norm": 3.1992828845977783,
|
| 2043 |
+
"learning_rate": 1.3453554748313441e-06,
|
| 2044 |
+
"loss": 0.0471,
|
| 2045 |
+
"step": 7175
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 0.8771929824561403,
|
| 2049 |
+
"grad_norm": 4.043613433837891,
|
| 2050 |
+
"learning_rate": 1.3129216398546965e-06,
|
| 2051 |
+
"loss": 0.0386,
|
| 2052 |
+
"step": 7200
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 0.880238791423002,
|
| 2056 |
+
"grad_norm": 2.8727216720581055,
|
| 2057 |
+
"learning_rate": 1.2804878048780488e-06,
|
| 2058 |
+
"loss": 0.0417,
|
| 2059 |
+
"step": 7225
|
| 2060 |
+
},
|
| 2061 |
+
{
|
| 2062 |
+
"epoch": 0.8832846003898636,
|
| 2063 |
+
"grad_norm": 4.605580806732178,
|
| 2064 |
+
"learning_rate": 1.2480539699014012e-06,
|
| 2065 |
+
"loss": 0.0485,
|
| 2066 |
+
"step": 7250
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 0.8863304093567251,
|
| 2070 |
+
"grad_norm": 6.22200345993042,
|
| 2071 |
+
"learning_rate": 1.2156201349247536e-06,
|
| 2072 |
+
"loss": 0.0424,
|
| 2073 |
+
"step": 7275
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 0.8893762183235867,
|
| 2077 |
+
"grad_norm": 3.0598998069763184,
|
| 2078 |
+
"learning_rate": 1.183186299948106e-06,
|
| 2079 |
+
"loss": 0.0534,
|
| 2080 |
+
"step": 7300
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 0.8924220272904484,
|
| 2084 |
+
"grad_norm": 4.829048156738281,
|
| 2085 |
+
"learning_rate": 1.1507524649714583e-06,
|
| 2086 |
+
"loss": 0.0456,
|
| 2087 |
+
"step": 7325
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 0.89546783625731,
|
| 2091 |
+
"grad_norm": 5.607174873352051,
|
| 2092 |
+
"learning_rate": 1.1183186299948106e-06,
|
| 2093 |
+
"loss": 0.0394,
|
| 2094 |
+
"step": 7350
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 0.8985136452241715,
|
| 2098 |
+
"grad_norm": 2.056725025177002,
|
| 2099 |
+
"learning_rate": 1.085884795018163e-06,
|
| 2100 |
+
"loss": 0.0423,
|
| 2101 |
+
"step": 7375
|
| 2102 |
+
},
|
| 2103 |
+
{
|
| 2104 |
+
"epoch": 0.9015594541910331,
|
| 2105 |
+
"grad_norm": 1.4574615955352783,
|
| 2106 |
+
"learning_rate": 1.0534509600415153e-06,
|
| 2107 |
+
"loss": 0.0432,
|
| 2108 |
+
"step": 7400
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 0.9046052631578947,
|
| 2112 |
+
"grad_norm": 2.392552137374878,
|
| 2113 |
+
"learning_rate": 1.0210171250648677e-06,
|
| 2114 |
+
"loss": 0.0331,
|
| 2115 |
+
"step": 7425
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 0.9076510721247564,
|
| 2119 |
+
"grad_norm": 3.653775215148926,
|
| 2120 |
+
"learning_rate": 9.8858329008822e-07,
|
| 2121 |
+
"loss": 0.0469,
|
| 2122 |
+
"step": 7450
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 0.910696881091618,
|
| 2126 |
+
"grad_norm": 3.350886344909668,
|
| 2127 |
+
"learning_rate": 9.561494551115724e-07,
|
| 2128 |
+
"loss": 0.044,
|
| 2129 |
+
"step": 7475
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 0.9137426900584795,
|
| 2133 |
+
"grad_norm": 6.157811164855957,
|
| 2134 |
+
"learning_rate": 9.237156201349248e-07,
|
| 2135 |
+
"loss": 0.044,
|
| 2136 |
+
"step": 7500
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 0.9167884990253411,
|
| 2140 |
+
"grad_norm": 4.67850399017334,
|
| 2141 |
+
"learning_rate": 8.912817851582773e-07,
|
| 2142 |
+
"loss": 0.0545,
|
| 2143 |
+
"step": 7525
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"epoch": 0.9198343079922028,
|
| 2147 |
+
"grad_norm": 2.9768331050872803,
|
| 2148 |
+
"learning_rate": 8.588479501816295e-07,
|
| 2149 |
+
"loss": 0.0469,
|
| 2150 |
+
"step": 7550
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 0.9228801169590644,
|
| 2154 |
+
"grad_norm": 2.5755043029785156,
|
| 2155 |
+
"learning_rate": 8.264141152049819e-07,
|
| 2156 |
+
"loss": 0.0436,
|
| 2157 |
+
"step": 7575
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 0.9259259259259259,
|
| 2161 |
+
"grad_norm": 5.110294342041016,
|
| 2162 |
+
"learning_rate": 7.939802802283343e-07,
|
| 2163 |
+
"loss": 0.0461,
|
| 2164 |
+
"step": 7600
|
| 2165 |
+
},
|
| 2166 |
+
{
|
| 2167 |
+
"epoch": 0.9289717348927875,
|
| 2168 |
+
"grad_norm": 3.600740432739258,
|
| 2169 |
+
"learning_rate": 7.615464452516866e-07,
|
| 2170 |
+
"loss": 0.0378,
|
| 2171 |
+
"step": 7625
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 0.9320175438596491,
|
| 2175 |
+
"grad_norm": 6.710562705993652,
|
| 2176 |
+
"learning_rate": 7.291126102750389e-07,
|
| 2177 |
+
"loss": 0.0461,
|
| 2178 |
+
"step": 7650
|
| 2179 |
+
},
|
| 2180 |
+
{
|
| 2181 |
+
"epoch": 0.9350633528265108,
|
| 2182 |
+
"grad_norm": 6.342244625091553,
|
| 2183 |
+
"learning_rate": 6.966787752983914e-07,
|
| 2184 |
+
"loss": 0.0478,
|
| 2185 |
+
"step": 7675
|
| 2186 |
+
},
|
| 2187 |
+
{
|
| 2188 |
+
"epoch": 0.9381091617933723,
|
| 2189 |
+
"grad_norm": 5.005722999572754,
|
| 2190 |
+
"learning_rate": 6.642449403217437e-07,
|
| 2191 |
+
"loss": 0.0455,
|
| 2192 |
+
"step": 7700
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 0.9411549707602339,
|
| 2196 |
+
"grad_norm": 4.427570343017578,
|
| 2197 |
+
"learning_rate": 6.31811105345096e-07,
|
| 2198 |
+
"loss": 0.0388,
|
| 2199 |
+
"step": 7725
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 0.9442007797270955,
|
| 2203 |
+
"grad_norm": 3.7985281944274902,
|
| 2204 |
+
"learning_rate": 5.993772703684484e-07,
|
| 2205 |
+
"loss": 0.0488,
|
| 2206 |
+
"step": 7750
|
| 2207 |
+
},
|
| 2208 |
+
{
|
| 2209 |
+
"epoch": 0.9472465886939572,
|
| 2210 |
+
"grad_norm": 6.278728008270264,
|
| 2211 |
+
"learning_rate": 5.669434353918007e-07,
|
| 2212 |
+
"loss": 0.0414,
|
| 2213 |
+
"step": 7775
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 0.9502923976608187,
|
| 2217 |
+
"grad_norm": 4.945337772369385,
|
| 2218 |
+
"learning_rate": 5.345096004151531e-07,
|
| 2219 |
+
"loss": 0.0425,
|
| 2220 |
+
"step": 7800
|
| 2221 |
+
},
|
| 2222 |
+
{
|
| 2223 |
+
"epoch": 0.9533382066276803,
|
| 2224 |
+
"grad_norm": 3.97603440284729,
|
| 2225 |
+
"learning_rate": 5.020757654385055e-07,
|
| 2226 |
+
"loss": 0.0443,
|
| 2227 |
+
"step": 7825
|
| 2228 |
+
},
|
| 2229 |
+
{
|
| 2230 |
+
"epoch": 0.9563840155945419,
|
| 2231 |
+
"grad_norm": 5.937435150146484,
|
| 2232 |
+
"learning_rate": 4.696419304618578e-07,
|
| 2233 |
+
"loss": 0.0546,
|
| 2234 |
+
"step": 7850
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 0.9594298245614035,
|
| 2238 |
+
"grad_norm": 2.2830593585968018,
|
| 2239 |
+
"learning_rate": 4.3720809548521023e-07,
|
| 2240 |
+
"loss": 0.0461,
|
| 2241 |
+
"step": 7875
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 0.9624756335282652,
|
| 2245 |
+
"grad_norm": 1.9695100784301758,
|
| 2246 |
+
"learning_rate": 4.047742605085626e-07,
|
| 2247 |
+
"loss": 0.0398,
|
| 2248 |
+
"step": 7900
|
| 2249 |
+
},
|
| 2250 |
+
{
|
| 2251 |
+
"epoch": 0.9655214424951267,
|
| 2252 |
+
"grad_norm": 3.1578965187072754,
|
| 2253 |
+
"learning_rate": 3.723404255319149e-07,
|
| 2254 |
+
"loss": 0.043,
|
| 2255 |
+
"step": 7925
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 0.9685672514619883,
|
| 2259 |
+
"grad_norm": 3.1749889850616455,
|
| 2260 |
+
"learning_rate": 3.399065905552673e-07,
|
| 2261 |
+
"loss": 0.0496,
|
| 2262 |
+
"step": 7950
|
| 2263 |
+
},
|
| 2264 |
+
{
|
| 2265 |
+
"epoch": 0.9716130604288499,
|
| 2266 |
+
"grad_norm": 3.351229190826416,
|
| 2267 |
+
"learning_rate": 3.0747275557861967e-07,
|
| 2268 |
+
"loss": 0.0433,
|
| 2269 |
+
"step": 7975
|
| 2270 |
+
},
|
| 2271 |
+
{
|
| 2272 |
+
"epoch": 0.9746588693957114,
|
| 2273 |
+
"grad_norm": 3.811175584793091,
|
| 2274 |
+
"learning_rate": 2.75038920601972e-07,
|
| 2275 |
+
"loss": 0.0348,
|
| 2276 |
+
"step": 8000
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 0.9777046783625731,
|
| 2280 |
+
"grad_norm": 1.274682879447937,
|
| 2281 |
+
"learning_rate": 2.4260508562532433e-07,
|
| 2282 |
+
"loss": 0.046,
|
| 2283 |
+
"step": 8025
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 0.9807504873294347,
|
| 2287 |
+
"grad_norm": 3.756114959716797,
|
| 2288 |
+
"learning_rate": 2.1017125064867672e-07,
|
| 2289 |
+
"loss": 0.043,
|
| 2290 |
+
"step": 8050
|
| 2291 |
+
},
|
| 2292 |
+
{
|
| 2293 |
+
"epoch": 0.9837962962962963,
|
| 2294 |
+
"grad_norm": 4.1706037521362305,
|
| 2295 |
+
"learning_rate": 1.7773741567202908e-07,
|
| 2296 |
+
"loss": 0.0359,
|
| 2297 |
+
"step": 8075
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 0.9868421052631579,
|
| 2301 |
+
"grad_norm": 4.425180435180664,
|
| 2302 |
+
"learning_rate": 1.4530358069538144e-07,
|
| 2303 |
+
"loss": 0.039,
|
| 2304 |
+
"step": 8100
|
| 2305 |
+
},
|
| 2306 |
+
{
|
| 2307 |
+
"epoch": 0.9898879142300195,
|
| 2308 |
+
"grad_norm": 3.7593772411346436,
|
| 2309 |
+
"learning_rate": 1.128697457187338e-07,
|
| 2310 |
+
"loss": 0.0458,
|
| 2311 |
+
"step": 8125
|
| 2312 |
+
},
|
| 2313 |
+
{
|
| 2314 |
+
"epoch": 0.9929337231968811,
|
| 2315 |
+
"grad_norm": 4.999473571777344,
|
| 2316 |
+
"learning_rate": 8.043591074208614e-08,
|
| 2317 |
+
"loss": 0.0493,
|
| 2318 |
+
"step": 8150
|
| 2319 |
+
},
|
| 2320 |
+
{
|
| 2321 |
+
"epoch": 0.9959795321637427,
|
| 2322 |
+
"grad_norm": 1.8714449405670166,
|
| 2323 |
+
"learning_rate": 4.800207576543851e-08,
|
| 2324 |
+
"loss": 0.0433,
|
| 2325 |
+
"step": 8175
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 0.9990253411306043,
|
| 2329 |
+
"grad_norm": 1.6236313581466675,
|
| 2330 |
+
"learning_rate": 1.5568240788790867e-08,
|
| 2331 |
+
"loss": 0.0387,
|
| 2332 |
+
"step": 8200
|
| 2333 |
+
},
|
| 2334 |
+
{
|
| 2335 |
+
"epoch": 1.0,
|
| 2336 |
+
"eval_loss": 0.04804893210530281,
|
| 2337 |
+
"eval_runtime": 6129.2535,
|
| 2338 |
+
"eval_samples_per_second": 1.34,
|
| 2339 |
+
"eval_steps_per_second": 0.335,
|
| 2340 |
+
"eval_wer": 14.794215086646279,
|
| 2341 |
+
"step": 8208
|
| 2342 |
}
|
| 2343 |
],
|
| 2344 |
"logging_steps": 25,
|
|
|
|
| 2353 |
"should_evaluate": false,
|
| 2354 |
"should_log": false,
|
| 2355 |
"should_save": true,
|
| 2356 |
+
"should_training_stop": true
|
| 2357 |
},
|
| 2358 |
"attributes": {}
|
| 2359 |
}
|
| 2360 |
},
|
| 2361 |
+
"total_flos": 2.12948339392512e+18,
|
| 2362 |
"train_batch_size": 4,
|
| 2363 |
"trial_name": null,
|
| 2364 |
"trial_params": null
|