Training in progress, step 1500, 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 735217848
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:74572018ac522fe8bf3b91fec5b2e11a917f01268eb7fa79d8c28e82716c3641
|
| 3 |
size 735217848
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1470521978
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aaf6e503e47ad9a45b9fb2dbb896936f1d179d52c20c550add65c77f44a8193c
|
| 3 |
size 1470521978
|
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 sha256:5672d4a2bab2f5ec1b202aa86f336deecf9ade33ecc3e9f1ae101d08c2403c85
|
| 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 1000
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8dfb306257217b253dcb010fa8e7db4904b5a105fb159cc7b2977c1d185fc223
|
| 3 |
size 1000
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 100,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -867,6 +867,436 @@
|
|
| 867 |
"eval_spearman_manhattan": 0.8284194308491212,
|
| 868 |
"eval_steps_per_second": 15.014,
|
| 869 |
"step": 1000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 870 |
}
|
| 871 |
],
|
| 872 |
"logging_steps": 10,
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.7029053420805998,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 1500,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 867 |
"eval_spearman_manhattan": 0.8284194308491212,
|
| 868 |
"eval_steps_per_second": 15.014,
|
| 869 |
"step": 1000
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 0.4732895970009372,
|
| 873 |
+
"grad_norm": 2.064291477203369,
|
| 874 |
+
"learning_rate": 4.7041940018744145e-05,
|
| 875 |
+
"loss": 0.3188,
|
| 876 |
+
"step": 1010
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 0.47797563261480785,
|
| 880 |
+
"grad_norm": 1.338891625404358,
|
| 881 |
+
"learning_rate": 4.7012652296157454e-05,
|
| 882 |
+
"loss": 0.3381,
|
| 883 |
+
"step": 1020
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 0.48266166822867856,
|
| 887 |
+
"grad_norm": 1.4479578733444214,
|
| 888 |
+
"learning_rate": 4.698336457357076e-05,
|
| 889 |
+
"loss": 0.3187,
|
| 890 |
+
"step": 1030
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.4873477038425492,
|
| 894 |
+
"grad_norm": 2.0868189334869385,
|
| 895 |
+
"learning_rate": 4.695407685098407e-05,
|
| 896 |
+
"loss": 0.3697,
|
| 897 |
+
"step": 1040
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 0.49203373945641987,
|
| 901 |
+
"grad_norm": 1.9820175170898438,
|
| 902 |
+
"learning_rate": 4.692478912839738e-05,
|
| 903 |
+
"loss": 0.3079,
|
| 904 |
+
"step": 1050
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 0.4967197750702905,
|
| 908 |
+
"grad_norm": 1.2479910850524902,
|
| 909 |
+
"learning_rate": 4.689550140581069e-05,
|
| 910 |
+
"loss": 0.3129,
|
| 911 |
+
"step": 1060
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 0.5014058106841612,
|
| 915 |
+
"grad_norm": 1.5005191564559937,
|
| 916 |
+
"learning_rate": 4.6866213683223996e-05,
|
| 917 |
+
"loss": 0.3588,
|
| 918 |
+
"step": 1070
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 0.5060918462980318,
|
| 922 |
+
"grad_norm": 1.730153203010559,
|
| 923 |
+
"learning_rate": 4.68369259606373e-05,
|
| 924 |
+
"loss": 0.3511,
|
| 925 |
+
"step": 1080
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 0.5107778819119025,
|
| 929 |
+
"grad_norm": 1.8256272077560425,
|
| 930 |
+
"learning_rate": 4.680763823805061e-05,
|
| 931 |
+
"loss": 0.3483,
|
| 932 |
+
"step": 1090
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 0.5154639175257731,
|
| 936 |
+
"grad_norm": 1.8275713920593262,
|
| 937 |
+
"learning_rate": 4.677835051546392e-05,
|
| 938 |
+
"loss": 0.3301,
|
| 939 |
+
"step": 1100
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 0.5154639175257731,
|
| 943 |
+
"eval_loss": 0.06174962595105171,
|
| 944 |
+
"eval_pearson_cosine": 0.8116036079365685,
|
| 945 |
+
"eval_pearson_dot": 0.7637833872485942,
|
| 946 |
+
"eval_pearson_euclidean": 0.8160585823410784,
|
| 947 |
+
"eval_pearson_manhattan": 0.8176746469698344,
|
| 948 |
+
"eval_runtime": 6.1383,
|
| 949 |
+
"eval_samples_per_second": 244.369,
|
| 950 |
+
"eval_spearman_cosine": 0.8149952741898824,
|
| 951 |
+
"eval_spearman_dot": 0.759784369983796,
|
| 952 |
+
"eval_spearman_euclidean": 0.8211588786730816,
|
| 953 |
+
"eval_spearman_manhattan": 0.8228445193252625,
|
| 954 |
+
"eval_steps_per_second": 15.314,
|
| 955 |
+
"step": 1100
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 0.5201499531396439,
|
| 959 |
+
"grad_norm": 1.7808656692504883,
|
| 960 |
+
"learning_rate": 4.674906279287723e-05,
|
| 961 |
+
"loss": 0.3378,
|
| 962 |
+
"step": 1110
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.5248359887535146,
|
| 966 |
+
"grad_norm": 1.3912303447723389,
|
| 967 |
+
"learning_rate": 4.671977507029054e-05,
|
| 968 |
+
"loss": 0.3247,
|
| 969 |
+
"step": 1120
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.5295220243673852,
|
| 973 |
+
"grad_norm": 1.619547724723816,
|
| 974 |
+
"learning_rate": 4.669048734770384e-05,
|
| 975 |
+
"loss": 0.3548,
|
| 976 |
+
"step": 1130
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 0.5342080599812559,
|
| 980 |
+
"grad_norm": 1.6785143613815308,
|
| 981 |
+
"learning_rate": 4.666119962511715e-05,
|
| 982 |
+
"loss": 0.4056,
|
| 983 |
+
"step": 1140
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 0.5388940955951266,
|
| 987 |
+
"grad_norm": 1.4282417297363281,
|
| 988 |
+
"learning_rate": 4.6631911902530465e-05,
|
| 989 |
+
"loss": 0.3136,
|
| 990 |
+
"step": 1150
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 0.5435801312089972,
|
| 994 |
+
"grad_norm": 1.5950373411178589,
|
| 995 |
+
"learning_rate": 4.660262417994377e-05,
|
| 996 |
+
"loss": 0.3094,
|
| 997 |
+
"step": 1160
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 0.5482661668228679,
|
| 1001 |
+
"grad_norm": 1.9235565662384033,
|
| 1002 |
+
"learning_rate": 4.657333645735708e-05,
|
| 1003 |
+
"loss": 0.3409,
|
| 1004 |
+
"step": 1170
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.5529522024367385,
|
| 1008 |
+
"grad_norm": 1.2192574739456177,
|
| 1009 |
+
"learning_rate": 4.6544048734770383e-05,
|
| 1010 |
+
"loss": 0.3387,
|
| 1011 |
+
"step": 1180
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 0.5576382380506092,
|
| 1015 |
+
"grad_norm": 1.5550990104675293,
|
| 1016 |
+
"learning_rate": 4.651476101218369e-05,
|
| 1017 |
+
"loss": 0.3184,
|
| 1018 |
+
"step": 1190
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 0.5623242736644799,
|
| 1022 |
+
"grad_norm": 1.8576079607009888,
|
| 1023 |
+
"learning_rate": 4.6485473289597e-05,
|
| 1024 |
+
"loss": 0.3637,
|
| 1025 |
+
"step": 1200
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 0.5623242736644799,
|
| 1029 |
+
"eval_loss": 0.05324321612715721,
|
| 1030 |
+
"eval_pearson_cosine": 0.8107929850304814,
|
| 1031 |
+
"eval_pearson_dot": 0.768063847349957,
|
| 1032 |
+
"eval_pearson_euclidean": 0.8155502077488883,
|
| 1033 |
+
"eval_pearson_manhattan": 0.8174981555238503,
|
| 1034 |
+
"eval_runtime": 5.9644,
|
| 1035 |
+
"eval_samples_per_second": 251.492,
|
| 1036 |
+
"eval_spearman_cosine": 0.8145222586962418,
|
| 1037 |
+
"eval_spearman_dot": 0.7642997219390888,
|
| 1038 |
+
"eval_spearman_euclidean": 0.8201735536723759,
|
| 1039 |
+
"eval_spearman_manhattan": 0.8222186632592043,
|
| 1040 |
+
"eval_steps_per_second": 15.76,
|
| 1041 |
+
"step": 1200
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 0.5670103092783505,
|
| 1045 |
+
"grad_norm": 1.41835618019104,
|
| 1046 |
+
"learning_rate": 4.6456185567010316e-05,
|
| 1047 |
+
"loss": 0.3039,
|
| 1048 |
+
"step": 1210
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 0.5716963448922212,
|
| 1052 |
+
"grad_norm": 1.3850994110107422,
|
| 1053 |
+
"learning_rate": 4.6426897844423624e-05,
|
| 1054 |
+
"loss": 0.3187,
|
| 1055 |
+
"step": 1220
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 0.5763823805060918,
|
| 1059 |
+
"grad_norm": 1.437373399734497,
|
| 1060 |
+
"learning_rate": 4.6397610121836926e-05,
|
| 1061 |
+
"loss": 0.3337,
|
| 1062 |
+
"step": 1230
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 0.5810684161199625,
|
| 1066 |
+
"grad_norm": 1.2328146696090698,
|
| 1067 |
+
"learning_rate": 4.6368322399250235e-05,
|
| 1068 |
+
"loss": 0.2975,
|
| 1069 |
+
"step": 1240
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 0.5857544517338332,
|
| 1073 |
+
"grad_norm": 1.6191329956054688,
|
| 1074 |
+
"learning_rate": 4.633903467666354e-05,
|
| 1075 |
+
"loss": 0.3275,
|
| 1076 |
+
"step": 1250
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.5904404873477038,
|
| 1080 |
+
"grad_norm": 1.695470929145813,
|
| 1081 |
+
"learning_rate": 4.630974695407685e-05,
|
| 1082 |
+
"loss": 0.3485,
|
| 1083 |
+
"step": 1260
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 0.5951265229615745,
|
| 1087 |
+
"grad_norm": 1.6120591163635254,
|
| 1088 |
+
"learning_rate": 4.628045923149017e-05,
|
| 1089 |
+
"loss": 0.3515,
|
| 1090 |
+
"step": 1270
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 0.5998125585754451,
|
| 1094 |
+
"grad_norm": 1.6157792806625366,
|
| 1095 |
+
"learning_rate": 4.625117150890347e-05,
|
| 1096 |
+
"loss": 0.3043,
|
| 1097 |
+
"step": 1280
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 0.6044985941893158,
|
| 1101 |
+
"grad_norm": 1.1465294361114502,
|
| 1102 |
+
"learning_rate": 4.622188378631678e-05,
|
| 1103 |
+
"loss": 0.2884,
|
| 1104 |
+
"step": 1290
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 0.6091846298031866,
|
| 1108 |
+
"grad_norm": 1.583688497543335,
|
| 1109 |
+
"learning_rate": 4.6192596063730086e-05,
|
| 1110 |
+
"loss": 0.2885,
|
| 1111 |
+
"step": 1300
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 0.6091846298031866,
|
| 1115 |
+
"eval_loss": 0.04511857405304909,
|
| 1116 |
+
"eval_pearson_cosine": 0.8272008065998051,
|
| 1117 |
+
"eval_pearson_dot": 0.7924789666171037,
|
| 1118 |
+
"eval_pearson_euclidean": 0.8268065548805623,
|
| 1119 |
+
"eval_pearson_manhattan": 0.827500349653536,
|
| 1120 |
+
"eval_runtime": 6.0244,
|
| 1121 |
+
"eval_samples_per_second": 248.988,
|
| 1122 |
+
"eval_spearman_cosine": 0.8278338594350843,
|
| 1123 |
+
"eval_spearman_dot": 0.7887595412839734,
|
| 1124 |
+
"eval_spearman_euclidean": 0.8317669408319824,
|
| 1125 |
+
"eval_spearman_manhattan": 0.8323949761116776,
|
| 1126 |
+
"eval_steps_per_second": 15.603,
|
| 1127 |
+
"step": 1300
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 0.6138706654170571,
|
| 1131 |
+
"grad_norm": 1.081640362739563,
|
| 1132 |
+
"learning_rate": 4.6163308341143395e-05,
|
| 1133 |
+
"loss": 0.3526,
|
| 1134 |
+
"step": 1310
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 0.6185567010309279,
|
| 1138 |
+
"grad_norm": 1.760512351989746,
|
| 1139 |
+
"learning_rate": 4.61340206185567e-05,
|
| 1140 |
+
"loss": 0.3113,
|
| 1141 |
+
"step": 1320
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 0.6232427366447985,
|
| 1145 |
+
"grad_norm": 1.106444239616394,
|
| 1146 |
+
"learning_rate": 4.610473289597001e-05,
|
| 1147 |
+
"loss": 0.3126,
|
| 1148 |
+
"step": 1330
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.6279287722586692,
|
| 1152 |
+
"grad_norm": 1.3500837087631226,
|
| 1153 |
+
"learning_rate": 4.607544517338332e-05,
|
| 1154 |
+
"loss": 0.3094,
|
| 1155 |
+
"step": 1340
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 0.6326148078725399,
|
| 1159 |
+
"grad_norm": 1.727953553199768,
|
| 1160 |
+
"learning_rate": 4.604615745079663e-05,
|
| 1161 |
+
"loss": 0.3304,
|
| 1162 |
+
"step": 1350
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 0.6373008434864105,
|
| 1166 |
+
"grad_norm": 1.4341022968292236,
|
| 1167 |
+
"learning_rate": 4.601686972820994e-05,
|
| 1168 |
+
"loss": 0.2804,
|
| 1169 |
+
"step": 1360
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 0.6419868791002812,
|
| 1173 |
+
"grad_norm": 1.4479708671569824,
|
| 1174 |
+
"learning_rate": 4.5987582005623246e-05,
|
| 1175 |
+
"loss": 0.31,
|
| 1176 |
+
"step": 1370
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 0.6466729147141518,
|
| 1180 |
+
"grad_norm": 1.5667890310287476,
|
| 1181 |
+
"learning_rate": 4.5958294283036554e-05,
|
| 1182 |
+
"loss": 0.3149,
|
| 1183 |
+
"step": 1380
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 0.6513589503280225,
|
| 1187 |
+
"grad_norm": 1.7333146333694458,
|
| 1188 |
+
"learning_rate": 4.592900656044986e-05,
|
| 1189 |
+
"loss": 0.3247,
|
| 1190 |
+
"step": 1390
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.6560449859418932,
|
| 1194 |
+
"grad_norm": 1.914392113685608,
|
| 1195 |
+
"learning_rate": 4.589971883786317e-05,
|
| 1196 |
+
"loss": 0.2852,
|
| 1197 |
+
"step": 1400
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 0.6560449859418932,
|
| 1201 |
+
"eval_loss": 0.04731455817818642,
|
| 1202 |
+
"eval_pearson_cosine": 0.8245641713392331,
|
| 1203 |
+
"eval_pearson_dot": 0.7893189374890994,
|
| 1204 |
+
"eval_pearson_euclidean": 0.8220644314223797,
|
| 1205 |
+
"eval_pearson_manhattan": 0.8227839674683928,
|
| 1206 |
+
"eval_runtime": 6.0521,
|
| 1207 |
+
"eval_samples_per_second": 247.846,
|
| 1208 |
+
"eval_spearman_cosine": 0.8264178003782281,
|
| 1209 |
+
"eval_spearman_dot": 0.7874134051082518,
|
| 1210 |
+
"eval_spearman_euclidean": 0.8274821508565314,
|
| 1211 |
+
"eval_spearman_manhattan": 0.8280999297389011,
|
| 1212 |
+
"eval_steps_per_second": 15.532,
|
| 1213 |
+
"step": 1400
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.6607310215557638,
|
| 1217 |
+
"grad_norm": 1.2458995580673218,
|
| 1218 |
+
"learning_rate": 4.587043111527648e-05,
|
| 1219 |
+
"loss": 0.3068,
|
| 1220 |
+
"step": 1410
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.6654170571696345,
|
| 1224 |
+
"grad_norm": 1.6540151834487915,
|
| 1225 |
+
"learning_rate": 4.584114339268979e-05,
|
| 1226 |
+
"loss": 0.3034,
|
| 1227 |
+
"step": 1420
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.6701030927835051,
|
| 1231 |
+
"grad_norm": 1.2585715055465698,
|
| 1232 |
+
"learning_rate": 4.581185567010309e-05,
|
| 1233 |
+
"loss": 0.3297,
|
| 1234 |
+
"step": 1430
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.6747891283973758,
|
| 1238 |
+
"grad_norm": 1.5088609457015991,
|
| 1239 |
+
"learning_rate": 4.5782567947516406e-05,
|
| 1240 |
+
"loss": 0.3444,
|
| 1241 |
+
"step": 1440
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.6794751640112465,
|
| 1245 |
+
"grad_norm": 1.3120390176773071,
|
| 1246 |
+
"learning_rate": 4.5753280224929714e-05,
|
| 1247 |
+
"loss": 0.2882,
|
| 1248 |
+
"step": 1450
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.6841611996251171,
|
| 1252 |
+
"grad_norm": 1.1074262857437134,
|
| 1253 |
+
"learning_rate": 4.572399250234302e-05,
|
| 1254 |
+
"loss": 0.287,
|
| 1255 |
+
"step": 1460
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.6888472352389878,
|
| 1259 |
+
"grad_norm": 1.5284086465835571,
|
| 1260 |
+
"learning_rate": 4.569470477975633e-05,
|
| 1261 |
+
"loss": 0.3175,
|
| 1262 |
+
"step": 1470
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.6935332708528584,
|
| 1266 |
+
"grad_norm": 1.5610471963882446,
|
| 1267 |
+
"learning_rate": 4.566541705716963e-05,
|
| 1268 |
+
"loss": 0.3033,
|
| 1269 |
+
"step": 1480
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.6982193064667291,
|
| 1273 |
+
"grad_norm": 1.1839112043380737,
|
| 1274 |
+
"learning_rate": 4.563612933458294e-05,
|
| 1275 |
+
"loss": 0.2917,
|
| 1276 |
+
"step": 1490
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.7029053420805998,
|
| 1280 |
+
"grad_norm": 1.2611138820648193,
|
| 1281 |
+
"learning_rate": 4.560684161199626e-05,
|
| 1282 |
+
"loss": 0.3225,
|
| 1283 |
+
"step": 1500
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.7029053420805998,
|
| 1287 |
+
"eval_loss": 0.05073302239179611,
|
| 1288 |
+
"eval_pearson_cosine": 0.8258767050389224,
|
| 1289 |
+
"eval_pearson_dot": 0.7737186817222579,
|
| 1290 |
+
"eval_pearson_euclidean": 0.826299226589029,
|
| 1291 |
+
"eval_pearson_manhattan": 0.8274116157485736,
|
| 1292 |
+
"eval_runtime": 6.1756,
|
| 1293 |
+
"eval_samples_per_second": 242.892,
|
| 1294 |
+
"eval_spearman_cosine": 0.828410027637777,
|
| 1295 |
+
"eval_spearman_dot": 0.7707599871747091,
|
| 1296 |
+
"eval_spearman_euclidean": 0.8324611004376368,
|
| 1297 |
+
"eval_spearman_manhattan": 0.8335300441487923,
|
| 1298 |
+
"eval_steps_per_second": 15.221,
|
| 1299 |
+
"step": 1500
|
| 1300 |
}
|
| 1301 |
],
|
| 1302 |
"logging_steps": 10,
|