Training in progress, step 43500, 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 140637408
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c3ce4d88189913400b49335ca2fb16534ab9d7e1d3780188f1982360fe1a27ce
|
| 3 |
size 140637408
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 281362507
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bf75b25ccd25b73c5615796dea6b86b52a2fa2f5bb852f93aca81fc6aa31725e
|
| 3 |
size 281362507
|
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:a6dbb90f03320796635b26281f9a3921d95277c32e6673e792faf729c478a88b
|
| 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:a23657582995ac85300267e69eaa0ab85c1547387fcc6138cc09514e7f270fd1
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -30882,6 +30882,365 @@
|
|
| 30882 |
"eval_samples_per_second": 33.704,
|
| 30883 |
"eval_steps_per_second": 0.539,
|
| 30884 |
"step": 43000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 30885 |
}
|
| 30886 |
],
|
| 30887 |
"logging_steps": 10,
|
|
@@ -30901,7 +31260,7 @@
|
|
| 30901 |
"attributes": {}
|
| 30902 |
}
|
| 30903 |
},
|
| 30904 |
-
"total_flos": 7.
|
| 30905 |
"train_batch_size": 128,
|
| 30906 |
"trial_name": null,
|
| 30907 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.5701627913073112,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 43500,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 30882 |
"eval_samples_per_second": 33.704,
|
| 30883 |
"eval_steps_per_second": 0.539,
|
| 30884 |
"step": 43000
|
| 30885 |
+
},
|
| 30886 |
+
{
|
| 30887 |
+
"epoch": 0.563740267910976,
|
| 30888 |
+
"grad_norm": 0.6213541626930237,
|
| 30889 |
+
"learning_rate": 0.0001214998414820911,
|
| 30890 |
+
"loss": 3.272,
|
| 30891 |
+
"step": 43010
|
| 30892 |
+
},
|
| 30893 |
+
{
|
| 30894 |
+
"epoch": 0.5638713398170236,
|
| 30895 |
+
"grad_norm": 0.7418097853660583,
|
| 30896 |
+
"learning_rate": 0.00012143880285274015,
|
| 30897 |
+
"loss": 3.2694,
|
| 30898 |
+
"step": 43020
|
| 30899 |
+
},
|
| 30900 |
+
{
|
| 30901 |
+
"epoch": 0.5640024117230713,
|
| 30902 |
+
"grad_norm": 0.6527386903762817,
|
| 30903 |
+
"learning_rate": 0.00012137776913028117,
|
| 30904 |
+
"loss": 3.1938,
|
| 30905 |
+
"step": 43030
|
| 30906 |
+
},
|
| 30907 |
+
{
|
| 30908 |
+
"epoch": 0.564133483629119,
|
| 30909 |
+
"grad_norm": 0.7256929874420166,
|
| 30910 |
+
"learning_rate": 0.0001213167403251998,
|
| 30911 |
+
"loss": 3.2594,
|
| 30912 |
+
"step": 43040
|
| 30913 |
+
},
|
| 30914 |
+
{
|
| 30915 |
+
"epoch": 0.5642645555351666,
|
| 30916 |
+
"grad_norm": 0.6325649619102478,
|
| 30917 |
+
"learning_rate": 0.00012125571644798101,
|
| 30918 |
+
"loss": 3.2726,
|
| 30919 |
+
"step": 43050
|
| 30920 |
+
},
|
| 30921 |
+
{
|
| 30922 |
+
"epoch": 0.5643956274412143,
|
| 30923 |
+
"grad_norm": 0.9402995109558105,
|
| 30924 |
+
"learning_rate": 0.00012119469750910893,
|
| 30925 |
+
"loss": 3.2868,
|
| 30926 |
+
"step": 43060
|
| 30927 |
+
},
|
| 30928 |
+
{
|
| 30929 |
+
"epoch": 0.564526699347262,
|
| 30930 |
+
"grad_norm": 0.67952960729599,
|
| 30931 |
+
"learning_rate": 0.0001211336835190667,
|
| 30932 |
+
"loss": 3.235,
|
| 30933 |
+
"step": 43070
|
| 30934 |
+
},
|
| 30935 |
+
{
|
| 30936 |
+
"epoch": 0.5646577712533095,
|
| 30937 |
+
"grad_norm": 0.6465554237365723,
|
| 30938 |
+
"learning_rate": 0.00012107267448833675,
|
| 30939 |
+
"loss": 3.2616,
|
| 30940 |
+
"step": 43080
|
| 30941 |
+
},
|
| 30942 |
+
{
|
| 30943 |
+
"epoch": 0.5647888431593572,
|
| 30944 |
+
"grad_norm": 1.083726406097412,
|
| 30945 |
+
"learning_rate": 0.00012101167042740051,
|
| 30946 |
+
"loss": 3.2264,
|
| 30947 |
+
"step": 43090
|
| 30948 |
+
},
|
| 30949 |
+
{
|
| 30950 |
+
"epoch": 0.5649199150654048,
|
| 30951 |
+
"grad_norm": 0.5495240092277527,
|
| 30952 |
+
"learning_rate": 0.00012095067134673875,
|
| 30953 |
+
"loss": 3.2628,
|
| 30954 |
+
"step": 43100
|
| 30955 |
+
},
|
| 30956 |
+
{
|
| 30957 |
+
"epoch": 0.5650509869714525,
|
| 30958 |
+
"grad_norm": 0.7268524169921875,
|
| 30959 |
+
"learning_rate": 0.0001208896772568312,
|
| 30960 |
+
"loss": 3.2529,
|
| 30961 |
+
"step": 43110
|
| 30962 |
+
},
|
| 30963 |
+
{
|
| 30964 |
+
"epoch": 0.5651820588775002,
|
| 30965 |
+
"grad_norm": 0.6181415319442749,
|
| 30966 |
+
"learning_rate": 0.00012082868816815682,
|
| 30967 |
+
"loss": 3.3322,
|
| 30968 |
+
"step": 43120
|
| 30969 |
+
},
|
| 30970 |
+
{
|
| 30971 |
+
"epoch": 0.5653131307835478,
|
| 30972 |
+
"grad_norm": 0.8651639223098755,
|
| 30973 |
+
"learning_rate": 0.00012076770409119375,
|
| 30974 |
+
"loss": 3.2786,
|
| 30975 |
+
"step": 43130
|
| 30976 |
+
},
|
| 30977 |
+
{
|
| 30978 |
+
"epoch": 0.5654442026895955,
|
| 30979 |
+
"grad_norm": 0.6164896488189697,
|
| 30980 |
+
"learning_rate": 0.00012070672503641923,
|
| 30981 |
+
"loss": 3.277,
|
| 30982 |
+
"step": 43140
|
| 30983 |
+
},
|
| 30984 |
+
{
|
| 30985 |
+
"epoch": 0.5655752745956432,
|
| 30986 |
+
"grad_norm": 0.7646924257278442,
|
| 30987 |
+
"learning_rate": 0.00012064575101430953,
|
| 30988 |
+
"loss": 3.2511,
|
| 30989 |
+
"step": 43150
|
| 30990 |
+
},
|
| 30991 |
+
{
|
| 30992 |
+
"epoch": 0.5657063465016908,
|
| 30993 |
+
"grad_norm": 0.5828595757484436,
|
| 30994 |
+
"learning_rate": 0.00012058478203534028,
|
| 30995 |
+
"loss": 3.2324,
|
| 30996 |
+
"step": 43160
|
| 30997 |
+
},
|
| 30998 |
+
{
|
| 30999 |
+
"epoch": 0.5658374184077385,
|
| 31000 |
+
"grad_norm": 0.6382690668106079,
|
| 31001 |
+
"learning_rate": 0.000120523818109986,
|
| 31002 |
+
"loss": 3.2458,
|
| 31003 |
+
"step": 43170
|
| 31004 |
+
},
|
| 31005 |
+
{
|
| 31006 |
+
"epoch": 0.5659684903137862,
|
| 31007 |
+
"grad_norm": 0.6081956624984741,
|
| 31008 |
+
"learning_rate": 0.00012046285924872058,
|
| 31009 |
+
"loss": 3.2345,
|
| 31010 |
+
"step": 43180
|
| 31011 |
+
},
|
| 31012 |
+
{
|
| 31013 |
+
"epoch": 0.5660995622198338,
|
| 31014 |
+
"grad_norm": 0.716952383518219,
|
| 31015 |
+
"learning_rate": 0.0001204019054620168,
|
| 31016 |
+
"loss": 3.2913,
|
| 31017 |
+
"step": 43190
|
| 31018 |
+
},
|
| 31019 |
+
{
|
| 31020 |
+
"epoch": 0.5662306341258815,
|
| 31021 |
+
"grad_norm": 0.6723486185073853,
|
| 31022 |
+
"learning_rate": 0.00012034095676034675,
|
| 31023 |
+
"loss": 3.2534,
|
| 31024 |
+
"step": 43200
|
| 31025 |
+
},
|
| 31026 |
+
{
|
| 31027 |
+
"epoch": 0.5663617060319291,
|
| 31028 |
+
"grad_norm": 0.5785861015319824,
|
| 31029 |
+
"learning_rate": 0.00012028001315418155,
|
| 31030 |
+
"loss": 3.2725,
|
| 31031 |
+
"step": 43210
|
| 31032 |
+
},
|
| 31033 |
+
{
|
| 31034 |
+
"epoch": 0.5664927779379768,
|
| 31035 |
+
"grad_norm": 0.6231510639190674,
|
| 31036 |
+
"learning_rate": 0.00012021907465399159,
|
| 31037 |
+
"loss": 3.2744,
|
| 31038 |
+
"step": 43220
|
| 31039 |
+
},
|
| 31040 |
+
{
|
| 31041 |
+
"epoch": 0.5666238498440245,
|
| 31042 |
+
"grad_norm": 0.6003397703170776,
|
| 31043 |
+
"learning_rate": 0.00012015814127024612,
|
| 31044 |
+
"loss": 3.2267,
|
| 31045 |
+
"step": 43230
|
| 31046 |
+
},
|
| 31047 |
+
{
|
| 31048 |
+
"epoch": 0.5667549217500721,
|
| 31049 |
+
"grad_norm": 0.6370406746864319,
|
| 31050 |
+
"learning_rate": 0.00012009721301341375,
|
| 31051 |
+
"loss": 3.2198,
|
| 31052 |
+
"step": 43240
|
| 31053 |
+
},
|
| 31054 |
+
{
|
| 31055 |
+
"epoch": 0.5668859936561198,
|
| 31056 |
+
"grad_norm": 0.5604175329208374,
|
| 31057 |
+
"learning_rate": 0.00012003628989396211,
|
| 31058 |
+
"loss": 3.2311,
|
| 31059 |
+
"step": 43250
|
| 31060 |
+
},
|
| 31061 |
+
{
|
| 31062 |
+
"epoch": 0.5670170655621674,
|
| 31063 |
+
"grad_norm": 0.7046313285827637,
|
| 31064 |
+
"learning_rate": 0.00011997537192235794,
|
| 31065 |
+
"loss": 3.3137,
|
| 31066 |
+
"step": 43260
|
| 31067 |
+
},
|
| 31068 |
+
{
|
| 31069 |
+
"epoch": 0.567148137468215,
|
| 31070 |
+
"grad_norm": 0.6208400130271912,
|
| 31071 |
+
"learning_rate": 0.00011991445910906709,
|
| 31072 |
+
"loss": 3.2613,
|
| 31073 |
+
"step": 43270
|
| 31074 |
+
},
|
| 31075 |
+
{
|
| 31076 |
+
"epoch": 0.5672792093742627,
|
| 31077 |
+
"grad_norm": 0.6924196481704712,
|
| 31078 |
+
"learning_rate": 0.00011985355146455461,
|
| 31079 |
+
"loss": 3.2797,
|
| 31080 |
+
"step": 43280
|
| 31081 |
+
},
|
| 31082 |
+
{
|
| 31083 |
+
"epoch": 0.5674102812803103,
|
| 31084 |
+
"grad_norm": 0.6309049129486084,
|
| 31085 |
+
"learning_rate": 0.00011979264899928459,
|
| 31086 |
+
"loss": 3.214,
|
| 31087 |
+
"step": 43290
|
| 31088 |
+
},
|
| 31089 |
+
{
|
| 31090 |
+
"epoch": 0.567541353186358,
|
| 31091 |
+
"grad_norm": 0.6157112717628479,
|
| 31092 |
+
"learning_rate": 0.00011973175172372019,
|
| 31093 |
+
"loss": 3.2656,
|
| 31094 |
+
"step": 43300
|
| 31095 |
+
},
|
| 31096 |
+
{
|
| 31097 |
+
"epoch": 0.5676724250924057,
|
| 31098 |
+
"grad_norm": 0.6638853549957275,
|
| 31099 |
+
"learning_rate": 0.00011967085964832376,
|
| 31100 |
+
"loss": 3.1937,
|
| 31101 |
+
"step": 43310
|
| 31102 |
+
},
|
| 31103 |
+
{
|
| 31104 |
+
"epoch": 0.5678034969984533,
|
| 31105 |
+
"grad_norm": 0.5795360803604126,
|
| 31106 |
+
"learning_rate": 0.00011960997278355671,
|
| 31107 |
+
"loss": 3.2498,
|
| 31108 |
+
"step": 43320
|
| 31109 |
+
},
|
| 31110 |
+
{
|
| 31111 |
+
"epoch": 0.567934568904501,
|
| 31112 |
+
"grad_norm": 1.2784022092819214,
|
| 31113 |
+
"learning_rate": 0.00011954909113987965,
|
| 31114 |
+
"loss": 3.3026,
|
| 31115 |
+
"step": 43330
|
| 31116 |
+
},
|
| 31117 |
+
{
|
| 31118 |
+
"epoch": 0.5680656408105487,
|
| 31119 |
+
"grad_norm": 0.6048783659934998,
|
| 31120 |
+
"learning_rate": 0.00011948821472775213,
|
| 31121 |
+
"loss": 3.293,
|
| 31122 |
+
"step": 43340
|
| 31123 |
+
},
|
| 31124 |
+
{
|
| 31125 |
+
"epoch": 0.5681967127165963,
|
| 31126 |
+
"grad_norm": 0.6757849454879761,
|
| 31127 |
+
"learning_rate": 0.00011942734355763287,
|
| 31128 |
+
"loss": 3.2655,
|
| 31129 |
+
"step": 43350
|
| 31130 |
+
},
|
| 31131 |
+
{
|
| 31132 |
+
"epoch": 0.568327784622644,
|
| 31133 |
+
"grad_norm": 0.6158847808837891,
|
| 31134 |
+
"learning_rate": 0.0001193664776399798,
|
| 31135 |
+
"loss": 3.255,
|
| 31136 |
+
"step": 43360
|
| 31137 |
+
},
|
| 31138 |
+
{
|
| 31139 |
+
"epoch": 0.5684588565286917,
|
| 31140 |
+
"grad_norm": 0.7038004398345947,
|
| 31141 |
+
"learning_rate": 0.00011930561698524971,
|
| 31142 |
+
"loss": 3.2474,
|
| 31143 |
+
"step": 43370
|
| 31144 |
+
},
|
| 31145 |
+
{
|
| 31146 |
+
"epoch": 0.5685899284347393,
|
| 31147 |
+
"grad_norm": 0.6013663411140442,
|
| 31148 |
+
"learning_rate": 0.00011924476160389872,
|
| 31149 |
+
"loss": 3.238,
|
| 31150 |
+
"step": 43380
|
| 31151 |
+
},
|
| 31152 |
+
{
|
| 31153 |
+
"epoch": 0.568721000340787,
|
| 31154 |
+
"grad_norm": 0.620312511920929,
|
| 31155 |
+
"learning_rate": 0.00011918391150638194,
|
| 31156 |
+
"loss": 3.192,
|
| 31157 |
+
"step": 43390
|
| 31158 |
+
},
|
| 31159 |
+
{
|
| 31160 |
+
"epoch": 0.5688520722468347,
|
| 31161 |
+
"grad_norm": 0.5915484428405762,
|
| 31162 |
+
"learning_rate": 0.00011912306670315365,
|
| 31163 |
+
"loss": 3.2049,
|
| 31164 |
+
"step": 43400
|
| 31165 |
+
},
|
| 31166 |
+
{
|
| 31167 |
+
"epoch": 0.5689831441528823,
|
| 31168 |
+
"grad_norm": 0.6437747478485107,
|
| 31169 |
+
"learning_rate": 0.00011906222720466699,
|
| 31170 |
+
"loss": 3.2986,
|
| 31171 |
+
"step": 43410
|
| 31172 |
+
},
|
| 31173 |
+
{
|
| 31174 |
+
"epoch": 0.56911421605893,
|
| 31175 |
+
"grad_norm": 0.6576946377754211,
|
| 31176 |
+
"learning_rate": 0.00011900139302137447,
|
| 31177 |
+
"loss": 3.2759,
|
| 31178 |
+
"step": 43420
|
| 31179 |
+
},
|
| 31180 |
+
{
|
| 31181 |
+
"epoch": 0.5692452879649776,
|
| 31182 |
+
"grad_norm": 0.6345543265342712,
|
| 31183 |
+
"learning_rate": 0.00011894056416372759,
|
| 31184 |
+
"loss": 3.2864,
|
| 31185 |
+
"step": 43430
|
| 31186 |
+
},
|
| 31187 |
+
{
|
| 31188 |
+
"epoch": 0.5693763598710252,
|
| 31189 |
+
"grad_norm": 0.6375799775123596,
|
| 31190 |
+
"learning_rate": 0.00011887974064217683,
|
| 31191 |
+
"loss": 3.2344,
|
| 31192 |
+
"step": 43440
|
| 31193 |
+
},
|
| 31194 |
+
{
|
| 31195 |
+
"epoch": 0.5695074317770729,
|
| 31196 |
+
"grad_norm": 0.6860445141792297,
|
| 31197 |
+
"learning_rate": 0.00011881892246717192,
|
| 31198 |
+
"loss": 3.2574,
|
| 31199 |
+
"step": 43450
|
| 31200 |
+
},
|
| 31201 |
+
{
|
| 31202 |
+
"epoch": 0.5696385036831205,
|
| 31203 |
+
"grad_norm": 1.2995883226394653,
|
| 31204 |
+
"learning_rate": 0.0001187581096491615,
|
| 31205 |
+
"loss": 3.2888,
|
| 31206 |
+
"step": 43460
|
| 31207 |
+
},
|
| 31208 |
+
{
|
| 31209 |
+
"epoch": 0.5697695755891682,
|
| 31210 |
+
"grad_norm": 0.6449354887008667,
|
| 31211 |
+
"learning_rate": 0.00011869730219859348,
|
| 31212 |
+
"loss": 3.2381,
|
| 31213 |
+
"step": 43470
|
| 31214 |
+
},
|
| 31215 |
+
{
|
| 31216 |
+
"epoch": 0.5699006474952159,
|
| 31217 |
+
"grad_norm": 0.6677840352058411,
|
| 31218 |
+
"learning_rate": 0.00011863650012591463,
|
| 31219 |
+
"loss": 3.2961,
|
| 31220 |
+
"step": 43480
|
| 31221 |
+
},
|
| 31222 |
+
{
|
| 31223 |
+
"epoch": 0.5700317194012635,
|
| 31224 |
+
"grad_norm": 0.6202585101127625,
|
| 31225 |
+
"learning_rate": 0.00011857570344157099,
|
| 31226 |
+
"loss": 3.2326,
|
| 31227 |
+
"step": 43490
|
| 31228 |
+
},
|
| 31229 |
+
{
|
| 31230 |
+
"epoch": 0.5701627913073112,
|
| 31231 |
+
"grad_norm": 0.6145714521408081,
|
| 31232 |
+
"learning_rate": 0.00011851491215600758,
|
| 31233 |
+
"loss": 3.2613,
|
| 31234 |
+
"step": 43500
|
| 31235 |
+
},
|
| 31236 |
+
{
|
| 31237 |
+
"epoch": 0.5701627913073112,
|
| 31238 |
+
"eval_accuracy": 0.3975381231671554,
|
| 31239 |
+
"eval_loss": 3.243159294128418,
|
| 31240 |
+
"eval_runtime": 61.2596,
|
| 31241 |
+
"eval_samples_per_second": 32.648,
|
| 31242 |
+
"eval_steps_per_second": 0.522,
|
| 31243 |
+
"step": 43500
|
| 31244 |
}
|
| 31245 |
],
|
| 31246 |
"logging_steps": 10,
|
|
|
|
| 31260 |
"attributes": {}
|
| 31261 |
}
|
| 31262 |
},
|
| 31263 |
+
"total_flos": 7.82295292182528e+17,
|
| 31264 |
"train_batch_size": 128,
|
| 31265 |
"trial_name": null,
|
| 31266 |
"trial_params": null
|