Training in progress, step 8500, 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 267829484
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ac527dbe9885338fecfab3563e193004fe5015a782e142033b00cd4adacfef33
|
| 3 |
size 267829484
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 535721146
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:00f13c7ab601a5530ba1fb17752c295cedf45760ae9ef8465585890e8a01c1fc
|
| 3 |
size 535721146
|
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 sha256:5a44913291ab392a066273d395cc9d02e12ea5bb71bddbffc3e8561a7aa61aa5
|
| 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 sha256:4e341f500c90835c03128f6fd935e02b54a4d6066b4a71f0997f5acf2e055176
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "./results/checkpoint-
|
| 4 |
-
"epoch": 2.
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -5815,6 +5815,369 @@
|
|
| 5815 |
"eval_samples_per_second": 124.195,
|
| 5816 |
"eval_steps_per_second": 7.771,
|
| 5817 |
"step": 8000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5818 |
}
|
| 5819 |
],
|
| 5820 |
"logging_steps": 10,
|
|
@@ -5843,7 +6206,7 @@
|
|
| 5843 |
"attributes": {}
|
| 5844 |
}
|
| 5845 |
},
|
| 5846 |
-
"total_flos": 1.
|
| 5847 |
"train_batch_size": 16,
|
| 5848 |
"trial_name": null,
|
| 5849 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.9451296249685376,
|
| 3 |
+
"best_model_checkpoint": "./results/checkpoint-8500",
|
| 4 |
+
"epoch": 2.905982905982906,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 8500,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 5815 |
"eval_samples_per_second": 124.195,
|
| 5816 |
"eval_steps_per_second": 7.771,
|
| 5817 |
"step": 8000
|
| 5818 |
+
},
|
| 5819 |
+
{
|
| 5820 |
+
"epoch": 2.7384615384615385,
|
| 5821 |
+
"grad_norm": 15.186147689819336,
|
| 5822 |
+
"learning_rate": 4.3589743589743586e-06,
|
| 5823 |
+
"loss": 0.1738,
|
| 5824 |
+
"step": 8010
|
| 5825 |
+
},
|
| 5826 |
+
{
|
| 5827 |
+
"epoch": 2.741880341880342,
|
| 5828 |
+
"grad_norm": 8.640946388244629,
|
| 5829 |
+
"learning_rate": 4.301994301994302e-06,
|
| 5830 |
+
"loss": 0.1729,
|
| 5831 |
+
"step": 8020
|
| 5832 |
+
},
|
| 5833 |
+
{
|
| 5834 |
+
"epoch": 2.7452991452991453,
|
| 5835 |
+
"grad_norm": 11.053187370300293,
|
| 5836 |
+
"learning_rate": 4.245014245014245e-06,
|
| 5837 |
+
"loss": 0.1322,
|
| 5838 |
+
"step": 8030
|
| 5839 |
+
},
|
| 5840 |
+
{
|
| 5841 |
+
"epoch": 2.7487179487179487,
|
| 5842 |
+
"grad_norm": 5.279446601867676,
|
| 5843 |
+
"learning_rate": 4.188034188034188e-06,
|
| 5844 |
+
"loss": 0.1461,
|
| 5845 |
+
"step": 8040
|
| 5846 |
+
},
|
| 5847 |
+
{
|
| 5848 |
+
"epoch": 2.752136752136752,
|
| 5849 |
+
"grad_norm": 4.323098659515381,
|
| 5850 |
+
"learning_rate": 4.131054131054131e-06,
|
| 5851 |
+
"loss": 0.1352,
|
| 5852 |
+
"step": 8050
|
| 5853 |
+
},
|
| 5854 |
+
{
|
| 5855 |
+
"epoch": 2.7555555555555555,
|
| 5856 |
+
"grad_norm": 3.249645233154297,
|
| 5857 |
+
"learning_rate": 4.074074074074075e-06,
|
| 5858 |
+
"loss": 0.1768,
|
| 5859 |
+
"step": 8060
|
| 5860 |
+
},
|
| 5861 |
+
{
|
| 5862 |
+
"epoch": 2.758974358974359,
|
| 5863 |
+
"grad_norm": 1.8464040756225586,
|
| 5864 |
+
"learning_rate": 4.017094017094018e-06,
|
| 5865 |
+
"loss": 0.1526,
|
| 5866 |
+
"step": 8070
|
| 5867 |
+
},
|
| 5868 |
+
{
|
| 5869 |
+
"epoch": 2.7623931623931623,
|
| 5870 |
+
"grad_norm": 6.442093849182129,
|
| 5871 |
+
"learning_rate": 3.960113960113961e-06,
|
| 5872 |
+
"loss": 0.0894,
|
| 5873 |
+
"step": 8080
|
| 5874 |
+
},
|
| 5875 |
+
{
|
| 5876 |
+
"epoch": 2.7658119658119658,
|
| 5877 |
+
"grad_norm": 6.989425182342529,
|
| 5878 |
+
"learning_rate": 3.903133903133904e-06,
|
| 5879 |
+
"loss": 0.185,
|
| 5880 |
+
"step": 8090
|
| 5881 |
+
},
|
| 5882 |
+
{
|
| 5883 |
+
"epoch": 2.769230769230769,
|
| 5884 |
+
"grad_norm": 3.125702142715454,
|
| 5885 |
+
"learning_rate": 3.846153846153847e-06,
|
| 5886 |
+
"loss": 0.1088,
|
| 5887 |
+
"step": 8100
|
| 5888 |
+
},
|
| 5889 |
+
{
|
| 5890 |
+
"epoch": 2.7726495726495726,
|
| 5891 |
+
"grad_norm": 4.960128307342529,
|
| 5892 |
+
"learning_rate": 3.7891737891737897e-06,
|
| 5893 |
+
"loss": 0.0836,
|
| 5894 |
+
"step": 8110
|
| 5895 |
+
},
|
| 5896 |
+
{
|
| 5897 |
+
"epoch": 2.776068376068376,
|
| 5898 |
+
"grad_norm": 1.6848759651184082,
|
| 5899 |
+
"learning_rate": 3.7321937321937327e-06,
|
| 5900 |
+
"loss": 0.1298,
|
| 5901 |
+
"step": 8120
|
| 5902 |
+
},
|
| 5903 |
+
{
|
| 5904 |
+
"epoch": 2.7794871794871794,
|
| 5905 |
+
"grad_norm": 4.422844409942627,
|
| 5906 |
+
"learning_rate": 3.6752136752136756e-06,
|
| 5907 |
+
"loss": 0.2038,
|
| 5908 |
+
"step": 8130
|
| 5909 |
+
},
|
| 5910 |
+
{
|
| 5911 |
+
"epoch": 2.782905982905983,
|
| 5912 |
+
"grad_norm": 1.5880966186523438,
|
| 5913 |
+
"learning_rate": 3.6182336182336186e-06,
|
| 5914 |
+
"loss": 0.1254,
|
| 5915 |
+
"step": 8140
|
| 5916 |
+
},
|
| 5917 |
+
{
|
| 5918 |
+
"epoch": 2.786324786324786,
|
| 5919 |
+
"grad_norm": 7.746439456939697,
|
| 5920 |
+
"learning_rate": 3.5612535612535615e-06,
|
| 5921 |
+
"loss": 0.1215,
|
| 5922 |
+
"step": 8150
|
| 5923 |
+
},
|
| 5924 |
+
{
|
| 5925 |
+
"epoch": 2.7897435897435896,
|
| 5926 |
+
"grad_norm": 0.546491801738739,
|
| 5927 |
+
"learning_rate": 3.5042735042735045e-06,
|
| 5928 |
+
"loss": 0.1141,
|
| 5929 |
+
"step": 8160
|
| 5930 |
+
},
|
| 5931 |
+
{
|
| 5932 |
+
"epoch": 2.793162393162393,
|
| 5933 |
+
"grad_norm": 9.661653518676758,
|
| 5934 |
+
"learning_rate": 3.4472934472934474e-06,
|
| 5935 |
+
"loss": 0.1641,
|
| 5936 |
+
"step": 8170
|
| 5937 |
+
},
|
| 5938 |
+
{
|
| 5939 |
+
"epoch": 2.7965811965811964,
|
| 5940 |
+
"grad_norm": 15.40501880645752,
|
| 5941 |
+
"learning_rate": 3.3903133903133904e-06,
|
| 5942 |
+
"loss": 0.1494,
|
| 5943 |
+
"step": 8180
|
| 5944 |
+
},
|
| 5945 |
+
{
|
| 5946 |
+
"epoch": 2.8,
|
| 5947 |
+
"grad_norm": 10.524840354919434,
|
| 5948 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 5949 |
+
"loss": 0.1519,
|
| 5950 |
+
"step": 8190
|
| 5951 |
+
},
|
| 5952 |
+
{
|
| 5953 |
+
"epoch": 2.8034188034188032,
|
| 5954 |
+
"grad_norm": 1.179671049118042,
|
| 5955 |
+
"learning_rate": 3.2763532763532763e-06,
|
| 5956 |
+
"loss": 0.1024,
|
| 5957 |
+
"step": 8200
|
| 5958 |
+
},
|
| 5959 |
+
{
|
| 5960 |
+
"epoch": 2.8068376068376066,
|
| 5961 |
+
"grad_norm": 9.402092933654785,
|
| 5962 |
+
"learning_rate": 3.2193732193732192e-06,
|
| 5963 |
+
"loss": 0.1104,
|
| 5964 |
+
"step": 8210
|
| 5965 |
+
},
|
| 5966 |
+
{
|
| 5967 |
+
"epoch": 2.81025641025641,
|
| 5968 |
+
"grad_norm": 1.9074152708053589,
|
| 5969 |
+
"learning_rate": 3.1623931623931626e-06,
|
| 5970 |
+
"loss": 0.103,
|
| 5971 |
+
"step": 8220
|
| 5972 |
+
},
|
| 5973 |
+
{
|
| 5974 |
+
"epoch": 2.8136752136752134,
|
| 5975 |
+
"grad_norm": 4.427234172821045,
|
| 5976 |
+
"learning_rate": 3.1054131054131055e-06,
|
| 5977 |
+
"loss": 0.1374,
|
| 5978 |
+
"step": 8230
|
| 5979 |
+
},
|
| 5980 |
+
{
|
| 5981 |
+
"epoch": 2.817094017094017,
|
| 5982 |
+
"grad_norm": 1.9632576704025269,
|
| 5983 |
+
"learning_rate": 3.048433048433049e-06,
|
| 5984 |
+
"loss": 0.1426,
|
| 5985 |
+
"step": 8240
|
| 5986 |
+
},
|
| 5987 |
+
{
|
| 5988 |
+
"epoch": 2.8205128205128203,
|
| 5989 |
+
"grad_norm": 21.75174331665039,
|
| 5990 |
+
"learning_rate": 2.991452991452992e-06,
|
| 5991 |
+
"loss": 0.1714,
|
| 5992 |
+
"step": 8250
|
| 5993 |
+
},
|
| 5994 |
+
{
|
| 5995 |
+
"epoch": 2.8239316239316237,
|
| 5996 |
+
"grad_norm": 7.342422008514404,
|
| 5997 |
+
"learning_rate": 2.934472934472935e-06,
|
| 5998 |
+
"loss": 0.1286,
|
| 5999 |
+
"step": 8260
|
| 6000 |
+
},
|
| 6001 |
+
{
|
| 6002 |
+
"epoch": 2.827350427350427,
|
| 6003 |
+
"grad_norm": 0.8742788434028625,
|
| 6004 |
+
"learning_rate": 2.8774928774928778e-06,
|
| 6005 |
+
"loss": 0.1025,
|
| 6006 |
+
"step": 8270
|
| 6007 |
+
},
|
| 6008 |
+
{
|
| 6009 |
+
"epoch": 2.830769230769231,
|
| 6010 |
+
"grad_norm": 27.062646865844727,
|
| 6011 |
+
"learning_rate": 2.8205128205128207e-06,
|
| 6012 |
+
"loss": 0.157,
|
| 6013 |
+
"step": 8280
|
| 6014 |
+
},
|
| 6015 |
+
{
|
| 6016 |
+
"epoch": 2.8341880341880343,
|
| 6017 |
+
"grad_norm": 2.987420082092285,
|
| 6018 |
+
"learning_rate": 2.7635327635327636e-06,
|
| 6019 |
+
"loss": 0.0961,
|
| 6020 |
+
"step": 8290
|
| 6021 |
+
},
|
| 6022 |
+
{
|
| 6023 |
+
"epoch": 2.8376068376068377,
|
| 6024 |
+
"grad_norm": 7.459050178527832,
|
| 6025 |
+
"learning_rate": 2.7065527065527066e-06,
|
| 6026 |
+
"loss": 0.1054,
|
| 6027 |
+
"step": 8300
|
| 6028 |
+
},
|
| 6029 |
+
{
|
| 6030 |
+
"epoch": 2.841025641025641,
|
| 6031 |
+
"grad_norm": 6.120343208312988,
|
| 6032 |
+
"learning_rate": 2.6495726495726495e-06,
|
| 6033 |
+
"loss": 0.1354,
|
| 6034 |
+
"step": 8310
|
| 6035 |
+
},
|
| 6036 |
+
{
|
| 6037 |
+
"epoch": 2.8444444444444446,
|
| 6038 |
+
"grad_norm": 8.228652000427246,
|
| 6039 |
+
"learning_rate": 2.5925925925925925e-06,
|
| 6040 |
+
"loss": 0.104,
|
| 6041 |
+
"step": 8320
|
| 6042 |
+
},
|
| 6043 |
+
{
|
| 6044 |
+
"epoch": 2.847863247863248,
|
| 6045 |
+
"grad_norm": 9.32483196258545,
|
| 6046 |
+
"learning_rate": 2.5356125356125354e-06,
|
| 6047 |
+
"loss": 0.0644,
|
| 6048 |
+
"step": 8330
|
| 6049 |
+
},
|
| 6050 |
+
{
|
| 6051 |
+
"epoch": 2.8512820512820514,
|
| 6052 |
+
"grad_norm": 0.9455430507659912,
|
| 6053 |
+
"learning_rate": 2.478632478632479e-06,
|
| 6054 |
+
"loss": 0.0869,
|
| 6055 |
+
"step": 8340
|
| 6056 |
+
},
|
| 6057 |
+
{
|
| 6058 |
+
"epoch": 2.8547008547008548,
|
| 6059 |
+
"grad_norm": 1.2311069965362549,
|
| 6060 |
+
"learning_rate": 2.4216524216524218e-06,
|
| 6061 |
+
"loss": 0.1625,
|
| 6062 |
+
"step": 8350
|
| 6063 |
+
},
|
| 6064 |
+
{
|
| 6065 |
+
"epoch": 2.858119658119658,
|
| 6066 |
+
"grad_norm": 2.0796940326690674,
|
| 6067 |
+
"learning_rate": 2.364672364672365e-06,
|
| 6068 |
+
"loss": 0.1166,
|
| 6069 |
+
"step": 8360
|
| 6070 |
+
},
|
| 6071 |
+
{
|
| 6072 |
+
"epoch": 2.8615384615384616,
|
| 6073 |
+
"grad_norm": 2.2284929752349854,
|
| 6074 |
+
"learning_rate": 2.307692307692308e-06,
|
| 6075 |
+
"loss": 0.0761,
|
| 6076 |
+
"step": 8370
|
| 6077 |
+
},
|
| 6078 |
+
{
|
| 6079 |
+
"epoch": 2.864957264957265,
|
| 6080 |
+
"grad_norm": 8.17489242553711,
|
| 6081 |
+
"learning_rate": 2.250712250712251e-06,
|
| 6082 |
+
"loss": 0.1465,
|
| 6083 |
+
"step": 8380
|
| 6084 |
+
},
|
| 6085 |
+
{
|
| 6086 |
+
"epoch": 2.8683760683760684,
|
| 6087 |
+
"grad_norm": 12.591669082641602,
|
| 6088 |
+
"learning_rate": 2.193732193732194e-06,
|
| 6089 |
+
"loss": 0.1878,
|
| 6090 |
+
"step": 8390
|
| 6091 |
+
},
|
| 6092 |
+
{
|
| 6093 |
+
"epoch": 2.871794871794872,
|
| 6094 |
+
"grad_norm": 12.328243255615234,
|
| 6095 |
+
"learning_rate": 2.136752136752137e-06,
|
| 6096 |
+
"loss": 0.1238,
|
| 6097 |
+
"step": 8400
|
| 6098 |
+
},
|
| 6099 |
+
{
|
| 6100 |
+
"epoch": 2.875213675213675,
|
| 6101 |
+
"grad_norm": 8.8943510055542,
|
| 6102 |
+
"learning_rate": 2.07977207977208e-06,
|
| 6103 |
+
"loss": 0.1561,
|
| 6104 |
+
"step": 8410
|
| 6105 |
+
},
|
| 6106 |
+
{
|
| 6107 |
+
"epoch": 2.8786324786324786,
|
| 6108 |
+
"grad_norm": 1.040313482284546,
|
| 6109 |
+
"learning_rate": 2.022792022792023e-06,
|
| 6110 |
+
"loss": 0.1668,
|
| 6111 |
+
"step": 8420
|
| 6112 |
+
},
|
| 6113 |
+
{
|
| 6114 |
+
"epoch": 2.882051282051282,
|
| 6115 |
+
"grad_norm": 15.345014572143555,
|
| 6116 |
+
"learning_rate": 1.9658119658119658e-06,
|
| 6117 |
+
"loss": 0.1305,
|
| 6118 |
+
"step": 8430
|
| 6119 |
+
},
|
| 6120 |
+
{
|
| 6121 |
+
"epoch": 2.8854700854700854,
|
| 6122 |
+
"grad_norm": 9.492894172668457,
|
| 6123 |
+
"learning_rate": 1.9088319088319087e-06,
|
| 6124 |
+
"loss": 0.1852,
|
| 6125 |
+
"step": 8440
|
| 6126 |
+
},
|
| 6127 |
+
{
|
| 6128 |
+
"epoch": 2.888888888888889,
|
| 6129 |
+
"grad_norm": 6.429811477661133,
|
| 6130 |
+
"learning_rate": 1.8518518518518519e-06,
|
| 6131 |
+
"loss": 0.1891,
|
| 6132 |
+
"step": 8450
|
| 6133 |
+
},
|
| 6134 |
+
{
|
| 6135 |
+
"epoch": 2.8923076923076922,
|
| 6136 |
+
"grad_norm": 8.479264259338379,
|
| 6137 |
+
"learning_rate": 1.7948717948717948e-06,
|
| 6138 |
+
"loss": 0.2209,
|
| 6139 |
+
"step": 8460
|
| 6140 |
+
},
|
| 6141 |
+
{
|
| 6142 |
+
"epoch": 2.8957264957264957,
|
| 6143 |
+
"grad_norm": 0.2948354482650757,
|
| 6144 |
+
"learning_rate": 1.7378917378917378e-06,
|
| 6145 |
+
"loss": 0.0827,
|
| 6146 |
+
"step": 8470
|
| 6147 |
+
},
|
| 6148 |
+
{
|
| 6149 |
+
"epoch": 2.899145299145299,
|
| 6150 |
+
"grad_norm": 5.249404430389404,
|
| 6151 |
+
"learning_rate": 1.6809116809116811e-06,
|
| 6152 |
+
"loss": 0.1036,
|
| 6153 |
+
"step": 8480
|
| 6154 |
+
},
|
| 6155 |
+
{
|
| 6156 |
+
"epoch": 2.9025641025641025,
|
| 6157 |
+
"grad_norm": 32.02479553222656,
|
| 6158 |
+
"learning_rate": 1.623931623931624e-06,
|
| 6159 |
+
"loss": 0.2021,
|
| 6160 |
+
"step": 8490
|
| 6161 |
+
},
|
| 6162 |
+
{
|
| 6163 |
+
"epoch": 2.905982905982906,
|
| 6164 |
+
"grad_norm": 1.3994070291519165,
|
| 6165 |
+
"learning_rate": 1.566951566951567e-06,
|
| 6166 |
+
"loss": 0.1744,
|
| 6167 |
+
"step": 8500
|
| 6168 |
+
},
|
| 6169 |
+
{
|
| 6170 |
+
"epoch": 2.905982905982906,
|
| 6171 |
+
"eval_accuracy": 0.9627318574237115,
|
| 6172 |
+
"eval_f1": 0.9451296249685376,
|
| 6173 |
+
"eval_loss": 0.10316536575555801,
|
| 6174 |
+
"eval_precision": 0.9432303441346396,
|
| 6175 |
+
"eval_recall": 0.9470365699873896,
|
| 6176 |
+
"eval_roc_auc": 0.9964824096022509,
|
| 6177 |
+
"eval_runtime": 93.0154,
|
| 6178 |
+
"eval_samples_per_second": 125.775,
|
| 6179 |
+
"eval_steps_per_second": 7.87,
|
| 6180 |
+
"step": 8500
|
| 6181 |
}
|
| 6182 |
],
|
| 6183 |
"logging_steps": 10,
|
|
|
|
| 6206 |
"attributes": {}
|
| 6207 |
}
|
| 6208 |
},
|
| 6209 |
+
"total_flos": 1.3510042818494976e+16,
|
| 6210 |
"train_batch_size": 16,
|
| 6211 |
"trial_name": null,
|
| 6212 |
"trial_params": null
|