End of training
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.9955555555555557,
|
| 3 |
+
"eval_accuracy": 0.9829629629629629,
|
| 4 |
+
"eval_loss": 0.05130303278565407,
|
| 5 |
+
"eval_runtime": 33.0298,
|
| 6 |
+
"eval_samples_per_second": 163.489,
|
| 7 |
+
"eval_steps_per_second": 10.233,
|
| 8 |
+
"total_flos": 1.608641425498964e+18,
|
| 9 |
+
"train_loss": 0.3095316025258054,
|
| 10 |
+
"train_runtime": 1092.5533,
|
| 11 |
+
"train_samples_per_second": 59.311,
|
| 12 |
+
"train_steps_per_second": 0.925
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.9955555555555557,
|
| 3 |
+
"eval_accuracy": 0.9829629629629629,
|
| 4 |
+
"eval_loss": 0.05130303278565407,
|
| 5 |
+
"eval_runtime": 33.0298,
|
| 6 |
+
"eval_samples_per_second": 163.489,
|
| 7 |
+
"eval_steps_per_second": 10.233
|
| 8 |
+
}
|
runs/Jun19_04-00-16_bfb2d875fb25/events.out.tfevents.1718770765.bfb2d875fb25.156.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1897edbc0920ae93027a3a1c6492d82438232d8880142e79305dde4f780a59b1
|
| 3 |
+
size 411
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.9955555555555557,
|
| 3 |
+
"total_flos": 1.608641425498964e+18,
|
| 4 |
+
"train_loss": 0.3095316025258054,
|
| 5 |
+
"train_runtime": 1092.5533,
|
| 6 |
+
"train_samples_per_second": 59.311,
|
| 7 |
+
"train_steps_per_second": 0.925
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,776 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9829629629629629,
|
| 3 |
+
"best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-eurosat/checkpoint-1011",
|
| 4 |
+
"epoch": 2.9955555555555557,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1011,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02962962962962963,
|
| 13 |
+
"grad_norm": 6.962750434875488,
|
| 14 |
+
"learning_rate": 4.901960784313726e-06,
|
| 15 |
+
"loss": 2.2752,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.05925925925925926,
|
| 20 |
+
"grad_norm": 7.147613048553467,
|
| 21 |
+
"learning_rate": 9.803921568627451e-06,
|
| 22 |
+
"loss": 2.1844,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.08888888888888889,
|
| 27 |
+
"grad_norm": 8.060157775878906,
|
| 28 |
+
"learning_rate": 1.4705882352941177e-05,
|
| 29 |
+
"loss": 1.9996,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.11851851851851852,
|
| 34 |
+
"grad_norm": 10.434203147888184,
|
| 35 |
+
"learning_rate": 1.9607843137254903e-05,
|
| 36 |
+
"loss": 1.692,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.14814814814814814,
|
| 41 |
+
"grad_norm": 14.684704780578613,
|
| 42 |
+
"learning_rate": 2.4509803921568626e-05,
|
| 43 |
+
"loss": 1.2463,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.17777777777777778,
|
| 48 |
+
"grad_norm": 15.639288902282715,
|
| 49 |
+
"learning_rate": 2.9411764705882354e-05,
|
| 50 |
+
"loss": 0.8793,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.2074074074074074,
|
| 55 |
+
"grad_norm": 14.043598175048828,
|
| 56 |
+
"learning_rate": 3.431372549019608e-05,
|
| 57 |
+
"loss": 0.6109,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.23703703703703705,
|
| 62 |
+
"grad_norm": 26.815710067749023,
|
| 63 |
+
"learning_rate": 3.9215686274509805e-05,
|
| 64 |
+
"loss": 0.5877,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.26666666666666666,
|
| 69 |
+
"grad_norm": 18.923198699951172,
|
| 70 |
+
"learning_rate": 4.411764705882353e-05,
|
| 71 |
+
"loss": 0.4282,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.2962962962962963,
|
| 76 |
+
"grad_norm": 21.275344848632812,
|
| 77 |
+
"learning_rate": 4.901960784313725e-05,
|
| 78 |
+
"loss": 0.4887,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.32592592592592595,
|
| 83 |
+
"grad_norm": 58.36842346191406,
|
| 84 |
+
"learning_rate": 4.955995599559956e-05,
|
| 85 |
+
"loss": 0.3896,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.35555555555555557,
|
| 90 |
+
"grad_norm": 18.006397247314453,
|
| 91 |
+
"learning_rate": 4.9009900990099014e-05,
|
| 92 |
+
"loss": 0.4162,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.3851851851851852,
|
| 97 |
+
"grad_norm": 13.68799114227295,
|
| 98 |
+
"learning_rate": 4.845984598459846e-05,
|
| 99 |
+
"loss": 0.3605,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.4148148148148148,
|
| 104 |
+
"grad_norm": 15.165467262268066,
|
| 105 |
+
"learning_rate": 4.790979097909791e-05,
|
| 106 |
+
"loss": 0.4391,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.4444444444444444,
|
| 111 |
+
"grad_norm": 10.881718635559082,
|
| 112 |
+
"learning_rate": 4.7359735973597365e-05,
|
| 113 |
+
"loss": 0.3706,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.4740740740740741,
|
| 118 |
+
"grad_norm": 15.909368515014648,
|
| 119 |
+
"learning_rate": 4.680968096809682e-05,
|
| 120 |
+
"loss": 0.3148,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.5037037037037037,
|
| 125 |
+
"grad_norm": 25.3117733001709,
|
| 126 |
+
"learning_rate": 4.6259625962596256e-05,
|
| 127 |
+
"loss": 0.3611,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.5333333333333333,
|
| 132 |
+
"grad_norm": 10.446706771850586,
|
| 133 |
+
"learning_rate": 4.570957095709571e-05,
|
| 134 |
+
"loss": 0.345,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.562962962962963,
|
| 139 |
+
"grad_norm": 15.340864181518555,
|
| 140 |
+
"learning_rate": 4.515951595159516e-05,
|
| 141 |
+
"loss": 0.3645,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.5925925925925926,
|
| 146 |
+
"grad_norm": 21.318401336669922,
|
| 147 |
+
"learning_rate": 4.4609460946094614e-05,
|
| 148 |
+
"loss": 0.3063,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.6222222222222222,
|
| 153 |
+
"grad_norm": 13.705307006835938,
|
| 154 |
+
"learning_rate": 4.405940594059406e-05,
|
| 155 |
+
"loss": 0.2603,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.6518518518518519,
|
| 160 |
+
"grad_norm": 14.29479694366455,
|
| 161 |
+
"learning_rate": 4.350935093509351e-05,
|
| 162 |
+
"loss": 0.2617,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.6814814814814815,
|
| 167 |
+
"grad_norm": 15.511056900024414,
|
| 168 |
+
"learning_rate": 4.2959295929592964e-05,
|
| 169 |
+
"loss": 0.2583,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.7111111111111111,
|
| 174 |
+
"grad_norm": 42.817989349365234,
|
| 175 |
+
"learning_rate": 4.240924092409242e-05,
|
| 176 |
+
"loss": 0.3146,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.7407407407407407,
|
| 181 |
+
"grad_norm": 11.812172889709473,
|
| 182 |
+
"learning_rate": 4.1859185918591856e-05,
|
| 183 |
+
"loss": 0.2315,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.7703703703703704,
|
| 188 |
+
"grad_norm": 11.193636894226074,
|
| 189 |
+
"learning_rate": 4.130913091309131e-05,
|
| 190 |
+
"loss": 0.329,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.8,
|
| 195 |
+
"grad_norm": 12.144834518432617,
|
| 196 |
+
"learning_rate": 4.075907590759076e-05,
|
| 197 |
+
"loss": 0.3005,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.8296296296296296,
|
| 202 |
+
"grad_norm": 10.566153526306152,
|
| 203 |
+
"learning_rate": 4.020902090209021e-05,
|
| 204 |
+
"loss": 0.2443,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.8592592592592593,
|
| 209 |
+
"grad_norm": 10.785486221313477,
|
| 210 |
+
"learning_rate": 3.965896589658966e-05,
|
| 211 |
+
"loss": 0.242,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.8888888888888888,
|
| 216 |
+
"grad_norm": 14.177382469177246,
|
| 217 |
+
"learning_rate": 3.910891089108911e-05,
|
| 218 |
+
"loss": 0.2615,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.9185185185185185,
|
| 223 |
+
"grad_norm": 16.649311065673828,
|
| 224 |
+
"learning_rate": 3.8558855885588564e-05,
|
| 225 |
+
"loss": 0.257,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.9481481481481482,
|
| 230 |
+
"grad_norm": 11.974807739257812,
|
| 231 |
+
"learning_rate": 3.800880088008801e-05,
|
| 232 |
+
"loss": 0.2559,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.9777777777777777,
|
| 237 |
+
"grad_norm": 9.136080741882324,
|
| 238 |
+
"learning_rate": 3.745874587458746e-05,
|
| 239 |
+
"loss": 0.2381,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.9985185185185185,
|
| 244 |
+
"eval_accuracy": 0.9575925925925926,
|
| 245 |
+
"eval_loss": 0.1251903772354126,
|
| 246 |
+
"eval_runtime": 33.135,
|
| 247 |
+
"eval_samples_per_second": 162.97,
|
| 248 |
+
"eval_steps_per_second": 10.201,
|
| 249 |
+
"step": 337
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 1.0074074074074073,
|
| 253 |
+
"grad_norm": 23.605918884277344,
|
| 254 |
+
"learning_rate": 3.690869086908691e-05,
|
| 255 |
+
"loss": 0.2405,
|
| 256 |
+
"step": 340
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 1.037037037037037,
|
| 260 |
+
"grad_norm": 9.887730598449707,
|
| 261 |
+
"learning_rate": 3.635863586358636e-05,
|
| 262 |
+
"loss": 0.2412,
|
| 263 |
+
"step": 350
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 1.0666666666666667,
|
| 267 |
+
"grad_norm": 12.197110176086426,
|
| 268 |
+
"learning_rate": 3.580858085808581e-05,
|
| 269 |
+
"loss": 0.2298,
|
| 270 |
+
"step": 360
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 1.0962962962962963,
|
| 274 |
+
"grad_norm": 12.699515342712402,
|
| 275 |
+
"learning_rate": 3.525852585258526e-05,
|
| 276 |
+
"loss": 0.2001,
|
| 277 |
+
"step": 370
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 1.125925925925926,
|
| 281 |
+
"grad_norm": 13.966781616210938,
|
| 282 |
+
"learning_rate": 3.470847084708471e-05,
|
| 283 |
+
"loss": 0.2106,
|
| 284 |
+
"step": 380
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.1555555555555554,
|
| 288 |
+
"grad_norm": 20.575868606567383,
|
| 289 |
+
"learning_rate": 3.415841584158416e-05,
|
| 290 |
+
"loss": 0.2002,
|
| 291 |
+
"step": 390
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 1.1851851851851851,
|
| 295 |
+
"grad_norm": 15.80282974243164,
|
| 296 |
+
"learning_rate": 3.360836083608361e-05,
|
| 297 |
+
"loss": 0.2532,
|
| 298 |
+
"step": 400
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 1.2148148148148148,
|
| 302 |
+
"grad_norm": 18.04647445678711,
|
| 303 |
+
"learning_rate": 3.305830583058306e-05,
|
| 304 |
+
"loss": 0.2609,
|
| 305 |
+
"step": 410
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.2444444444444445,
|
| 309 |
+
"grad_norm": 8.6199312210083,
|
| 310 |
+
"learning_rate": 3.250825082508251e-05,
|
| 311 |
+
"loss": 0.253,
|
| 312 |
+
"step": 420
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 1.2740740740740741,
|
| 316 |
+
"grad_norm": 7.412827968597412,
|
| 317 |
+
"learning_rate": 3.195819581958196e-05,
|
| 318 |
+
"loss": 0.213,
|
| 319 |
+
"step": 430
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 1.3037037037037038,
|
| 323 |
+
"grad_norm": 7.904267311096191,
|
| 324 |
+
"learning_rate": 3.140814081408141e-05,
|
| 325 |
+
"loss": 0.2584,
|
| 326 |
+
"step": 440
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.3333333333333333,
|
| 330 |
+
"grad_norm": 18.86865234375,
|
| 331 |
+
"learning_rate": 3.0858085808580864e-05,
|
| 332 |
+
"loss": 0.227,
|
| 333 |
+
"step": 450
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 1.362962962962963,
|
| 337 |
+
"grad_norm": 5.769145965576172,
|
| 338 |
+
"learning_rate": 3.0308030803080313e-05,
|
| 339 |
+
"loss": 0.2016,
|
| 340 |
+
"step": 460
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 1.3925925925925926,
|
| 344 |
+
"grad_norm": 11.96108341217041,
|
| 345 |
+
"learning_rate": 2.9757975797579756e-05,
|
| 346 |
+
"loss": 0.1635,
|
| 347 |
+
"step": 470
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.4222222222222223,
|
| 351 |
+
"grad_norm": 13.738970756530762,
|
| 352 |
+
"learning_rate": 2.9207920792079208e-05,
|
| 353 |
+
"loss": 0.1717,
|
| 354 |
+
"step": 480
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.4518518518518517,
|
| 358 |
+
"grad_norm": 11.988739967346191,
|
| 359 |
+
"learning_rate": 2.8657865786578657e-05,
|
| 360 |
+
"loss": 0.2126,
|
| 361 |
+
"step": 490
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 1.4814814814814814,
|
| 365 |
+
"grad_norm": 11.078845977783203,
|
| 366 |
+
"learning_rate": 2.810781078107811e-05,
|
| 367 |
+
"loss": 0.2282,
|
| 368 |
+
"step": 500
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.511111111111111,
|
| 372 |
+
"grad_norm": 10.5954008102417,
|
| 373 |
+
"learning_rate": 2.755775577557756e-05,
|
| 374 |
+
"loss": 0.1351,
|
| 375 |
+
"step": 510
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 1.5407407407407407,
|
| 379 |
+
"grad_norm": 15.503268241882324,
|
| 380 |
+
"learning_rate": 2.700770077007701e-05,
|
| 381 |
+
"loss": 0.2507,
|
| 382 |
+
"step": 520
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 1.5703703703703704,
|
| 386 |
+
"grad_norm": 15.796453475952148,
|
| 387 |
+
"learning_rate": 2.645764576457646e-05,
|
| 388 |
+
"loss": 0.185,
|
| 389 |
+
"step": 530
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.6,
|
| 393 |
+
"grad_norm": 11.540029525756836,
|
| 394 |
+
"learning_rate": 2.5907590759075913e-05,
|
| 395 |
+
"loss": 0.1605,
|
| 396 |
+
"step": 540
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 1.6296296296296298,
|
| 400 |
+
"grad_norm": 13.062150001525879,
|
| 401 |
+
"learning_rate": 2.5357535753575355e-05,
|
| 402 |
+
"loss": 0.2464,
|
| 403 |
+
"step": 550
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 1.6592592592592592,
|
| 407 |
+
"grad_norm": 11.136322021484375,
|
| 408 |
+
"learning_rate": 2.480748074807481e-05,
|
| 409 |
+
"loss": 0.2067,
|
| 410 |
+
"step": 560
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 1.6888888888888889,
|
| 414 |
+
"grad_norm": 8.773016929626465,
|
| 415 |
+
"learning_rate": 2.4257425742574257e-05,
|
| 416 |
+
"loss": 0.229,
|
| 417 |
+
"step": 570
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 1.7185185185185186,
|
| 421 |
+
"grad_norm": 8.625940322875977,
|
| 422 |
+
"learning_rate": 2.370737073707371e-05,
|
| 423 |
+
"loss": 0.1535,
|
| 424 |
+
"step": 580
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 1.748148148148148,
|
| 428 |
+
"grad_norm": 6.024807929992676,
|
| 429 |
+
"learning_rate": 2.3157315731573158e-05,
|
| 430 |
+
"loss": 0.1205,
|
| 431 |
+
"step": 590
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 1.7777777777777777,
|
| 435 |
+
"grad_norm": 9.04062557220459,
|
| 436 |
+
"learning_rate": 2.2607260726072607e-05,
|
| 437 |
+
"loss": 0.1724,
|
| 438 |
+
"step": 600
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 1.8074074074074074,
|
| 442 |
+
"grad_norm": 20.10211753845215,
|
| 443 |
+
"learning_rate": 2.2057205720572056e-05,
|
| 444 |
+
"loss": 0.1504,
|
| 445 |
+
"step": 610
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 1.837037037037037,
|
| 449 |
+
"grad_norm": 9.104300498962402,
|
| 450 |
+
"learning_rate": 2.150715071507151e-05,
|
| 451 |
+
"loss": 0.157,
|
| 452 |
+
"step": 620
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 1.8666666666666667,
|
| 456 |
+
"grad_norm": 8.573586463928223,
|
| 457 |
+
"learning_rate": 2.0957095709570958e-05,
|
| 458 |
+
"loss": 0.2248,
|
| 459 |
+
"step": 630
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 1.8962962962962964,
|
| 463 |
+
"grad_norm": 7.876155853271484,
|
| 464 |
+
"learning_rate": 2.0407040704070407e-05,
|
| 465 |
+
"loss": 0.1622,
|
| 466 |
+
"step": 640
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 1.925925925925926,
|
| 470 |
+
"grad_norm": 6.861870288848877,
|
| 471 |
+
"learning_rate": 1.9856985698569856e-05,
|
| 472 |
+
"loss": 0.1846,
|
| 473 |
+
"step": 650
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.9555555555555557,
|
| 477 |
+
"grad_norm": 8.942523956298828,
|
| 478 |
+
"learning_rate": 1.930693069306931e-05,
|
| 479 |
+
"loss": 0.1585,
|
| 480 |
+
"step": 660
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 1.9851851851851852,
|
| 484 |
+
"grad_norm": 12.770540237426758,
|
| 485 |
+
"learning_rate": 1.8756875687568758e-05,
|
| 486 |
+
"loss": 0.1729,
|
| 487 |
+
"step": 670
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 2.0,
|
| 491 |
+
"eval_accuracy": 0.9775925925925926,
|
| 492 |
+
"eval_loss": 0.06562940031290054,
|
| 493 |
+
"eval_runtime": 32.5381,
|
| 494 |
+
"eval_samples_per_second": 165.959,
|
| 495 |
+
"eval_steps_per_second": 10.388,
|
| 496 |
+
"step": 675
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 2.0148148148148146,
|
| 500 |
+
"grad_norm": 8.93161678314209,
|
| 501 |
+
"learning_rate": 1.8206820682068207e-05,
|
| 502 |
+
"loss": 0.1736,
|
| 503 |
+
"step": 680
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 2.0444444444444443,
|
| 507 |
+
"grad_norm": 8.148313522338867,
|
| 508 |
+
"learning_rate": 1.765676567656766e-05,
|
| 509 |
+
"loss": 0.1518,
|
| 510 |
+
"step": 690
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 2.074074074074074,
|
| 514 |
+
"grad_norm": 9.238503456115723,
|
| 515 |
+
"learning_rate": 1.7106710671067108e-05,
|
| 516 |
+
"loss": 0.181,
|
| 517 |
+
"step": 700
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 2.1037037037037036,
|
| 521 |
+
"grad_norm": 7.2876081466674805,
|
| 522 |
+
"learning_rate": 1.6556655665566557e-05,
|
| 523 |
+
"loss": 0.1277,
|
| 524 |
+
"step": 710
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.1333333333333333,
|
| 528 |
+
"grad_norm": 14.298846244812012,
|
| 529 |
+
"learning_rate": 1.6006600660066006e-05,
|
| 530 |
+
"loss": 0.1439,
|
| 531 |
+
"step": 720
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 2.162962962962963,
|
| 535 |
+
"grad_norm": 19.348224639892578,
|
| 536 |
+
"learning_rate": 1.545654565456546e-05,
|
| 537 |
+
"loss": 0.202,
|
| 538 |
+
"step": 730
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 2.1925925925925926,
|
| 542 |
+
"grad_norm": 8.49962043762207,
|
| 543 |
+
"learning_rate": 1.4906490649064908e-05,
|
| 544 |
+
"loss": 0.1552,
|
| 545 |
+
"step": 740
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 2.2222222222222223,
|
| 549 |
+
"grad_norm": 6.094121932983398,
|
| 550 |
+
"learning_rate": 1.4356435643564355e-05,
|
| 551 |
+
"loss": 0.177,
|
| 552 |
+
"step": 750
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 2.251851851851852,
|
| 556 |
+
"grad_norm": 10.521448135375977,
|
| 557 |
+
"learning_rate": 1.3806380638063806e-05,
|
| 558 |
+
"loss": 0.1667,
|
| 559 |
+
"step": 760
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 2.2814814814814817,
|
| 563 |
+
"grad_norm": 15.489738464355469,
|
| 564 |
+
"learning_rate": 1.3256325632563257e-05,
|
| 565 |
+
"loss": 0.1298,
|
| 566 |
+
"step": 770
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 2.311111111111111,
|
| 570 |
+
"grad_norm": 11.301315307617188,
|
| 571 |
+
"learning_rate": 1.2706270627062708e-05,
|
| 572 |
+
"loss": 0.1701,
|
| 573 |
+
"step": 780
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 2.3407407407407406,
|
| 577 |
+
"grad_norm": 8.533815383911133,
|
| 578 |
+
"learning_rate": 1.2156215621562157e-05,
|
| 579 |
+
"loss": 0.1708,
|
| 580 |
+
"step": 790
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 2.3703703703703702,
|
| 584 |
+
"grad_norm": 13.924063682556152,
|
| 585 |
+
"learning_rate": 1.1606160616061606e-05,
|
| 586 |
+
"loss": 0.1394,
|
| 587 |
+
"step": 800
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 2.4,
|
| 591 |
+
"grad_norm": 2.552699089050293,
|
| 592 |
+
"learning_rate": 1.1056105610561057e-05,
|
| 593 |
+
"loss": 0.1553,
|
| 594 |
+
"step": 810
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 2.4296296296296296,
|
| 598 |
+
"grad_norm": 11.489421844482422,
|
| 599 |
+
"learning_rate": 1.0506050605060507e-05,
|
| 600 |
+
"loss": 0.1479,
|
| 601 |
+
"step": 820
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 2.4592592592592593,
|
| 605 |
+
"grad_norm": 11.2034330368042,
|
| 606 |
+
"learning_rate": 9.955995599559956e-06,
|
| 607 |
+
"loss": 0.1715,
|
| 608 |
+
"step": 830
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 2.488888888888889,
|
| 612 |
+
"grad_norm": 6.7186737060546875,
|
| 613 |
+
"learning_rate": 9.405940594059407e-06,
|
| 614 |
+
"loss": 0.1441,
|
| 615 |
+
"step": 840
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 2.5185185185185186,
|
| 619 |
+
"grad_norm": 7.652073383331299,
|
| 620 |
+
"learning_rate": 8.855885588558856e-06,
|
| 621 |
+
"loss": 0.1058,
|
| 622 |
+
"step": 850
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 2.5481481481481483,
|
| 626 |
+
"grad_norm": 10.549448013305664,
|
| 627 |
+
"learning_rate": 8.305830583058307e-06,
|
| 628 |
+
"loss": 0.1485,
|
| 629 |
+
"step": 860
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 2.5777777777777775,
|
| 633 |
+
"grad_norm": 7.53530740737915,
|
| 634 |
+
"learning_rate": 7.755775577557756e-06,
|
| 635 |
+
"loss": 0.1403,
|
| 636 |
+
"step": 870
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 2.6074074074074076,
|
| 640 |
+
"grad_norm": 12.338042259216309,
|
| 641 |
+
"learning_rate": 7.205720572057206e-06,
|
| 642 |
+
"loss": 0.1746,
|
| 643 |
+
"step": 880
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 2.637037037037037,
|
| 647 |
+
"grad_norm": 18.06306266784668,
|
| 648 |
+
"learning_rate": 6.655665566556656e-06,
|
| 649 |
+
"loss": 0.1506,
|
| 650 |
+
"step": 890
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 2.6666666666666665,
|
| 654 |
+
"grad_norm": 8.664934158325195,
|
| 655 |
+
"learning_rate": 6.105610561056106e-06,
|
| 656 |
+
"loss": 0.1258,
|
| 657 |
+
"step": 900
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 2.696296296296296,
|
| 661 |
+
"grad_norm": 6.592322826385498,
|
| 662 |
+
"learning_rate": 5.555555555555556e-06,
|
| 663 |
+
"loss": 0.1438,
|
| 664 |
+
"step": 910
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 2.725925925925926,
|
| 668 |
+
"grad_norm": 4.439072608947754,
|
| 669 |
+
"learning_rate": 5.005500550055006e-06,
|
| 670 |
+
"loss": 0.1049,
|
| 671 |
+
"step": 920
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 2.7555555555555555,
|
| 675 |
+
"grad_norm": 4.729221343994141,
|
| 676 |
+
"learning_rate": 4.455445544554456e-06,
|
| 677 |
+
"loss": 0.1445,
|
| 678 |
+
"step": 930
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 2.785185185185185,
|
| 682 |
+
"grad_norm": 6.323641300201416,
|
| 683 |
+
"learning_rate": 3.905390539053906e-06,
|
| 684 |
+
"loss": 0.1269,
|
| 685 |
+
"step": 940
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 2.814814814814815,
|
| 689 |
+
"grad_norm": 8.931975364685059,
|
| 690 |
+
"learning_rate": 3.3553355335533555e-06,
|
| 691 |
+
"loss": 0.1046,
|
| 692 |
+
"step": 950
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 2.8444444444444446,
|
| 696 |
+
"grad_norm": 25.645519256591797,
|
| 697 |
+
"learning_rate": 2.8052805280528055e-06,
|
| 698 |
+
"loss": 0.166,
|
| 699 |
+
"step": 960
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 2.8740740740740742,
|
| 703 |
+
"grad_norm": 5.7749128341674805,
|
| 704 |
+
"learning_rate": 2.2552255225522554e-06,
|
| 705 |
+
"loss": 0.1325,
|
| 706 |
+
"step": 970
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 2.9037037037037035,
|
| 710 |
+
"grad_norm": 3.4846718311309814,
|
| 711 |
+
"learning_rate": 1.7051705170517053e-06,
|
| 712 |
+
"loss": 0.1116,
|
| 713 |
+
"step": 980
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 2.9333333333333336,
|
| 717 |
+
"grad_norm": 9.06553840637207,
|
| 718 |
+
"learning_rate": 1.155115511551155e-06,
|
| 719 |
+
"loss": 0.1218,
|
| 720 |
+
"step": 990
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 2.962962962962963,
|
| 724 |
+
"grad_norm": 4.044698238372803,
|
| 725 |
+
"learning_rate": 6.050605060506051e-07,
|
| 726 |
+
"loss": 0.0964,
|
| 727 |
+
"step": 1000
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 2.9925925925925925,
|
| 731 |
+
"grad_norm": 13.73366928100586,
|
| 732 |
+
"learning_rate": 5.5005500550055006e-08,
|
| 733 |
+
"loss": 0.1188,
|
| 734 |
+
"step": 1010
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 2.9955555555555557,
|
| 738 |
+
"eval_accuracy": 0.9829629629629629,
|
| 739 |
+
"eval_loss": 0.05130303278565407,
|
| 740 |
+
"eval_runtime": 32.1348,
|
| 741 |
+
"eval_samples_per_second": 168.042,
|
| 742 |
+
"eval_steps_per_second": 10.518,
|
| 743 |
+
"step": 1011
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 2.9955555555555557,
|
| 747 |
+
"step": 1011,
|
| 748 |
+
"total_flos": 1.608641425498964e+18,
|
| 749 |
+
"train_loss": 0.3095316025258054,
|
| 750 |
+
"train_runtime": 1092.5533,
|
| 751 |
+
"train_samples_per_second": 59.311,
|
| 752 |
+
"train_steps_per_second": 0.925
|
| 753 |
+
}
|
| 754 |
+
],
|
| 755 |
+
"logging_steps": 10,
|
| 756 |
+
"max_steps": 1011,
|
| 757 |
+
"num_input_tokens_seen": 0,
|
| 758 |
+
"num_train_epochs": 3,
|
| 759 |
+
"save_steps": 500,
|
| 760 |
+
"stateful_callbacks": {
|
| 761 |
+
"TrainerControl": {
|
| 762 |
+
"args": {
|
| 763 |
+
"should_epoch_stop": false,
|
| 764 |
+
"should_evaluate": false,
|
| 765 |
+
"should_log": false,
|
| 766 |
+
"should_save": true,
|
| 767 |
+
"should_training_stop": true
|
| 768 |
+
},
|
| 769 |
+
"attributes": {}
|
| 770 |
+
}
|
| 771 |
+
},
|
| 772 |
+
"total_flos": 1.608641425498964e+18,
|
| 773 |
+
"train_batch_size": 16,
|
| 774 |
+
"trial_name": null,
|
| 775 |
+
"trial_params": null
|
| 776 |
+
}
|