End of training
Browse files- all_results.json +13 -0
- eval_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +655 -0
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.99,
|
| 3 |
+
"eval_accuracy": 0.9796296296296296,
|
| 4 |
+
"eval_loss": 0.1896590143442154,
|
| 5 |
+
"eval_runtime": 69.4801,
|
| 6 |
+
"eval_samples_per_second": 77.72,
|
| 7 |
+
"eval_steps_per_second": 2.432,
|
| 8 |
+
"total_flos": 4.999521681367106e+18,
|
| 9 |
+
"train_loss": 0.5908920062439782,
|
| 10 |
+
"train_runtime": 1706.6198,
|
| 11 |
+
"train_samples_per_second": 37.97,
|
| 12 |
+
"train_steps_per_second": 0.295
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.99,
|
| 3 |
+
"eval_accuracy": 0.9796296296296296,
|
| 4 |
+
"eval_loss": 0.1896590143442154,
|
| 5 |
+
"eval_runtime": 69.4801,
|
| 6 |
+
"eval_samples_per_second": 77.72,
|
| 7 |
+
"eval_steps_per_second": 2.432
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.99,
|
| 3 |
+
"total_flos": 4.999521681367106e+18,
|
| 4 |
+
"train_loss": 0.5908920062439782,
|
| 5 |
+
"train_runtime": 1706.6198,
|
| 6 |
+
"train_samples_per_second": 37.97,
|
| 7 |
+
"train_steps_per_second": 0.295
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,655 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9796296296296296,
|
| 3 |
+
"best_model_checkpoint": "bottom-layers-finetuned-eurosat/checkpoint-504",
|
| 4 |
+
"epoch": 2.986666666666667,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 504,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.03,
|
| 13 |
+
"learning_rate": 4.901960784313726e-06,
|
| 14 |
+
"loss": 2.2965,
|
| 15 |
+
"step": 5
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.06,
|
| 19 |
+
"learning_rate": 9.803921568627451e-06,
|
| 20 |
+
"loss": 2.2868,
|
| 21 |
+
"step": 10
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.09,
|
| 25 |
+
"learning_rate": 1.4705882352941177e-05,
|
| 26 |
+
"loss": 2.2711,
|
| 27 |
+
"step": 15
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.12,
|
| 31 |
+
"learning_rate": 1.9607843137254903e-05,
|
| 32 |
+
"loss": 2.239,
|
| 33 |
+
"step": 20
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.15,
|
| 37 |
+
"learning_rate": 2.4509803921568626e-05,
|
| 38 |
+
"loss": 2.1905,
|
| 39 |
+
"step": 25
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.18,
|
| 43 |
+
"learning_rate": 2.9411764705882354e-05,
|
| 44 |
+
"loss": 2.1375,
|
| 45 |
+
"step": 30
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.21,
|
| 49 |
+
"learning_rate": 3.431372549019608e-05,
|
| 50 |
+
"loss": 2.0589,
|
| 51 |
+
"step": 35
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.24,
|
| 55 |
+
"learning_rate": 3.9215686274509805e-05,
|
| 56 |
+
"loss": 1.9787,
|
| 57 |
+
"step": 40
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.27,
|
| 61 |
+
"learning_rate": 4.411764705882353e-05,
|
| 62 |
+
"loss": 1.8736,
|
| 63 |
+
"step": 45
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.3,
|
| 67 |
+
"learning_rate": 4.901960784313725e-05,
|
| 68 |
+
"loss": 1.7237,
|
| 69 |
+
"step": 50
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.33,
|
| 73 |
+
"learning_rate": 4.955849889624724e-05,
|
| 74 |
+
"loss": 1.6018,
|
| 75 |
+
"step": 55
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.36,
|
| 79 |
+
"learning_rate": 4.900662251655629e-05,
|
| 80 |
+
"loss": 1.4656,
|
| 81 |
+
"step": 60
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.39,
|
| 85 |
+
"learning_rate": 4.845474613686534e-05,
|
| 86 |
+
"loss": 1.3407,
|
| 87 |
+
"step": 65
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.41,
|
| 91 |
+
"learning_rate": 4.7902869757174393e-05,
|
| 92 |
+
"loss": 1.2151,
|
| 93 |
+
"step": 70
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.44,
|
| 97 |
+
"learning_rate": 4.735099337748345e-05,
|
| 98 |
+
"loss": 1.1103,
|
| 99 |
+
"step": 75
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.47,
|
| 103 |
+
"learning_rate": 4.67991169977925e-05,
|
| 104 |
+
"loss": 0.9809,
|
| 105 |
+
"step": 80
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.5,
|
| 109 |
+
"learning_rate": 4.624724061810155e-05,
|
| 110 |
+
"loss": 0.9362,
|
| 111 |
+
"step": 85
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.53,
|
| 115 |
+
"learning_rate": 4.56953642384106e-05,
|
| 116 |
+
"loss": 0.8689,
|
| 117 |
+
"step": 90
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.56,
|
| 121 |
+
"learning_rate": 4.514348785871965e-05,
|
| 122 |
+
"loss": 0.8,
|
| 123 |
+
"step": 95
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.59,
|
| 127 |
+
"learning_rate": 4.45916114790287e-05,
|
| 128 |
+
"loss": 0.7548,
|
| 129 |
+
"step": 100
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.62,
|
| 133 |
+
"learning_rate": 4.403973509933775e-05,
|
| 134 |
+
"loss": 0.7139,
|
| 135 |
+
"step": 105
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.65,
|
| 139 |
+
"learning_rate": 4.34878587196468e-05,
|
| 140 |
+
"loss": 0.6564,
|
| 141 |
+
"step": 110
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.68,
|
| 145 |
+
"learning_rate": 4.293598233995585e-05,
|
| 146 |
+
"loss": 0.629,
|
| 147 |
+
"step": 115
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.71,
|
| 151 |
+
"learning_rate": 4.23841059602649e-05,
|
| 152 |
+
"loss": 0.5827,
|
| 153 |
+
"step": 120
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.74,
|
| 157 |
+
"learning_rate": 4.1832229580573955e-05,
|
| 158 |
+
"loss": 0.5532,
|
| 159 |
+
"step": 125
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.77,
|
| 163 |
+
"learning_rate": 4.1280353200883e-05,
|
| 164 |
+
"loss": 0.529,
|
| 165 |
+
"step": 130
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.8,
|
| 169 |
+
"learning_rate": 4.0728476821192055e-05,
|
| 170 |
+
"loss": 0.5351,
|
| 171 |
+
"step": 135
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.83,
|
| 175 |
+
"learning_rate": 4.017660044150111e-05,
|
| 176 |
+
"loss": 0.5099,
|
| 177 |
+
"step": 140
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.86,
|
| 181 |
+
"learning_rate": 3.962472406181016e-05,
|
| 182 |
+
"loss": 0.4486,
|
| 183 |
+
"step": 145
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.89,
|
| 187 |
+
"learning_rate": 3.907284768211921e-05,
|
| 188 |
+
"loss": 0.4603,
|
| 189 |
+
"step": 150
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.92,
|
| 193 |
+
"learning_rate": 3.852097130242826e-05,
|
| 194 |
+
"loss": 0.4566,
|
| 195 |
+
"step": 155
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.95,
|
| 199 |
+
"learning_rate": 3.796909492273731e-05,
|
| 200 |
+
"loss": 0.4401,
|
| 201 |
+
"step": 160
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.98,
|
| 205 |
+
"learning_rate": 3.741721854304636e-05,
|
| 206 |
+
"loss": 0.4395,
|
| 207 |
+
"step": 165
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 1.0,
|
| 211 |
+
"eval_accuracy": 0.9648148148148148,
|
| 212 |
+
"eval_loss": 0.3520647883415222,
|
| 213 |
+
"eval_runtime": 68.6805,
|
| 214 |
+
"eval_samples_per_second": 78.625,
|
| 215 |
+
"eval_steps_per_second": 2.461,
|
| 216 |
+
"step": 168
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 1.01,
|
| 220 |
+
"learning_rate": 3.686534216335541e-05,
|
| 221 |
+
"loss": 0.423,
|
| 222 |
+
"step": 170
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 1.04,
|
| 226 |
+
"learning_rate": 3.631346578366446e-05,
|
| 227 |
+
"loss": 0.3978,
|
| 228 |
+
"step": 175
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 1.07,
|
| 232 |
+
"learning_rate": 3.576158940397351e-05,
|
| 233 |
+
"loss": 0.3975,
|
| 234 |
+
"step": 180
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 1.1,
|
| 238 |
+
"learning_rate": 3.520971302428256e-05,
|
| 239 |
+
"loss": 0.3968,
|
| 240 |
+
"step": 185
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.13,
|
| 244 |
+
"learning_rate": 3.465783664459161e-05,
|
| 245 |
+
"loss": 0.3974,
|
| 246 |
+
"step": 190
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 1.16,
|
| 250 |
+
"learning_rate": 3.410596026490066e-05,
|
| 251 |
+
"loss": 0.3757,
|
| 252 |
+
"step": 195
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 1.19,
|
| 256 |
+
"learning_rate": 3.3554083885209716e-05,
|
| 257 |
+
"loss": 0.3432,
|
| 258 |
+
"step": 200
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 1.21,
|
| 262 |
+
"learning_rate": 3.300220750551877e-05,
|
| 263 |
+
"loss": 0.3592,
|
| 264 |
+
"step": 205
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 1.24,
|
| 268 |
+
"learning_rate": 3.2450331125827816e-05,
|
| 269 |
+
"loss": 0.3528,
|
| 270 |
+
"step": 210
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 1.27,
|
| 274 |
+
"learning_rate": 3.189845474613687e-05,
|
| 275 |
+
"loss": 0.3405,
|
| 276 |
+
"step": 215
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 1.3,
|
| 280 |
+
"learning_rate": 3.134657836644592e-05,
|
| 281 |
+
"loss": 0.3274,
|
| 282 |
+
"step": 220
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 1.33,
|
| 286 |
+
"learning_rate": 3.079470198675497e-05,
|
| 287 |
+
"loss": 0.3841,
|
| 288 |
+
"step": 225
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 1.36,
|
| 292 |
+
"learning_rate": 3.0242825607064017e-05,
|
| 293 |
+
"loss": 0.3571,
|
| 294 |
+
"step": 230
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 1.39,
|
| 298 |
+
"learning_rate": 2.969094922737307e-05,
|
| 299 |
+
"loss": 0.3171,
|
| 300 |
+
"step": 235
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 1.42,
|
| 304 |
+
"learning_rate": 2.913907284768212e-05,
|
| 305 |
+
"loss": 0.3445,
|
| 306 |
+
"step": 240
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 1.45,
|
| 310 |
+
"learning_rate": 2.8587196467991174e-05,
|
| 311 |
+
"loss": 0.3717,
|
| 312 |
+
"step": 245
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 1.48,
|
| 316 |
+
"learning_rate": 2.803532008830022e-05,
|
| 317 |
+
"loss": 0.3768,
|
| 318 |
+
"step": 250
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 1.51,
|
| 322 |
+
"learning_rate": 2.7483443708609274e-05,
|
| 323 |
+
"loss": 0.3281,
|
| 324 |
+
"step": 255
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 1.54,
|
| 328 |
+
"learning_rate": 2.693156732891832e-05,
|
| 329 |
+
"loss": 0.3437,
|
| 330 |
+
"step": 260
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 1.57,
|
| 334 |
+
"learning_rate": 2.6379690949227374e-05,
|
| 335 |
+
"loss": 0.3113,
|
| 336 |
+
"step": 265
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 1.6,
|
| 340 |
+
"learning_rate": 2.5827814569536424e-05,
|
| 341 |
+
"loss": 0.3119,
|
| 342 |
+
"step": 270
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 1.63,
|
| 346 |
+
"learning_rate": 2.5275938189845478e-05,
|
| 347 |
+
"loss": 0.2737,
|
| 348 |
+
"step": 275
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"epoch": 1.66,
|
| 352 |
+
"learning_rate": 2.4724061810154528e-05,
|
| 353 |
+
"loss": 0.3649,
|
| 354 |
+
"step": 280
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.69,
|
| 358 |
+
"learning_rate": 2.4172185430463578e-05,
|
| 359 |
+
"loss": 0.328,
|
| 360 |
+
"step": 285
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 1.72,
|
| 364 |
+
"learning_rate": 2.3620309050772628e-05,
|
| 365 |
+
"loss": 0.3084,
|
| 366 |
+
"step": 290
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 1.75,
|
| 370 |
+
"learning_rate": 2.306843267108168e-05,
|
| 371 |
+
"loss": 0.3011,
|
| 372 |
+
"step": 295
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 1.78,
|
| 376 |
+
"learning_rate": 2.2516556291390732e-05,
|
| 377 |
+
"loss": 0.3178,
|
| 378 |
+
"step": 300
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 1.81,
|
| 382 |
+
"learning_rate": 2.1964679911699782e-05,
|
| 383 |
+
"loss": 0.2653,
|
| 384 |
+
"step": 305
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 1.84,
|
| 388 |
+
"learning_rate": 2.1412803532008832e-05,
|
| 389 |
+
"loss": 0.3,
|
| 390 |
+
"step": 310
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 1.87,
|
| 394 |
+
"learning_rate": 2.0860927152317882e-05,
|
| 395 |
+
"loss": 0.3085,
|
| 396 |
+
"step": 315
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 1.9,
|
| 400 |
+
"learning_rate": 2.0309050772626932e-05,
|
| 401 |
+
"loss": 0.2785,
|
| 402 |
+
"step": 320
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 1.93,
|
| 406 |
+
"learning_rate": 1.9757174392935986e-05,
|
| 407 |
+
"loss": 0.2812,
|
| 408 |
+
"step": 325
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 1.96,
|
| 412 |
+
"learning_rate": 1.9205298013245036e-05,
|
| 413 |
+
"loss": 0.2709,
|
| 414 |
+
"step": 330
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 1.99,
|
| 418 |
+
"learning_rate": 1.8653421633554086e-05,
|
| 419 |
+
"loss": 0.3093,
|
| 420 |
+
"step": 335
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 2.0,
|
| 424 |
+
"eval_accuracy": 0.9764814814814815,
|
| 425 |
+
"eval_loss": 0.216535285115242,
|
| 426 |
+
"eval_runtime": 69.25,
|
| 427 |
+
"eval_samples_per_second": 77.978,
|
| 428 |
+
"eval_steps_per_second": 2.44,
|
| 429 |
+
"step": 337
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 2.01,
|
| 433 |
+
"learning_rate": 1.8101545253863136e-05,
|
| 434 |
+
"loss": 0.2985,
|
| 435 |
+
"step": 340
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 2.04,
|
| 439 |
+
"learning_rate": 1.7549668874172186e-05,
|
| 440 |
+
"loss": 0.2863,
|
| 441 |
+
"step": 345
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 2.07,
|
| 445 |
+
"learning_rate": 1.699779249448124e-05,
|
| 446 |
+
"loss": 0.2912,
|
| 447 |
+
"step": 350
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 2.1,
|
| 451 |
+
"learning_rate": 1.644591611479029e-05,
|
| 452 |
+
"loss": 0.2363,
|
| 453 |
+
"step": 355
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 2.13,
|
| 457 |
+
"learning_rate": 1.589403973509934e-05,
|
| 458 |
+
"loss": 0.2927,
|
| 459 |
+
"step": 360
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 2.16,
|
| 463 |
+
"learning_rate": 1.534216335540839e-05,
|
| 464 |
+
"loss": 0.2855,
|
| 465 |
+
"step": 365
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 2.19,
|
| 469 |
+
"learning_rate": 1.479028697571744e-05,
|
| 470 |
+
"loss": 0.2891,
|
| 471 |
+
"step": 370
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 2.22,
|
| 475 |
+
"learning_rate": 1.4238410596026492e-05,
|
| 476 |
+
"loss": 0.2918,
|
| 477 |
+
"step": 375
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 2.25,
|
| 481 |
+
"learning_rate": 1.3686534216335542e-05,
|
| 482 |
+
"loss": 0.2656,
|
| 483 |
+
"step": 380
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 2.28,
|
| 487 |
+
"learning_rate": 1.3134657836644593e-05,
|
| 488 |
+
"loss": 0.2515,
|
| 489 |
+
"step": 385
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 2.31,
|
| 493 |
+
"learning_rate": 1.2582781456953644e-05,
|
| 494 |
+
"loss": 0.2983,
|
| 495 |
+
"step": 390
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 2.34,
|
| 499 |
+
"learning_rate": 1.2030905077262694e-05,
|
| 500 |
+
"loss": 0.2691,
|
| 501 |
+
"step": 395
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 2.37,
|
| 505 |
+
"learning_rate": 1.1479028697571745e-05,
|
| 506 |
+
"loss": 0.2863,
|
| 507 |
+
"step": 400
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 2.4,
|
| 511 |
+
"learning_rate": 1.0927152317880796e-05,
|
| 512 |
+
"loss": 0.2205,
|
| 513 |
+
"step": 405
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 2.43,
|
| 517 |
+
"learning_rate": 1.0375275938189846e-05,
|
| 518 |
+
"loss": 0.261,
|
| 519 |
+
"step": 410
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 2.46,
|
| 523 |
+
"learning_rate": 9.823399558498897e-06,
|
| 524 |
+
"loss": 0.2179,
|
| 525 |
+
"step": 415
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 2.49,
|
| 529 |
+
"learning_rate": 9.271523178807948e-06,
|
| 530 |
+
"loss": 0.2446,
|
| 531 |
+
"step": 420
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 2.52,
|
| 535 |
+
"learning_rate": 8.719646799116998e-06,
|
| 536 |
+
"loss": 0.2695,
|
| 537 |
+
"step": 425
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 2.55,
|
| 541 |
+
"learning_rate": 8.16777041942605e-06,
|
| 542 |
+
"loss": 0.2919,
|
| 543 |
+
"step": 430
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 2.58,
|
| 547 |
+
"learning_rate": 7.6158940397350995e-06,
|
| 548 |
+
"loss": 0.2925,
|
| 549 |
+
"step": 435
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 2.61,
|
| 553 |
+
"learning_rate": 7.0640176600441505e-06,
|
| 554 |
+
"loss": 0.2524,
|
| 555 |
+
"step": 440
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 2.64,
|
| 559 |
+
"learning_rate": 6.512141280353201e-06,
|
| 560 |
+
"loss": 0.2584,
|
| 561 |
+
"step": 445
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 2.67,
|
| 565 |
+
"learning_rate": 5.9602649006622515e-06,
|
| 566 |
+
"loss": 0.2678,
|
| 567 |
+
"step": 450
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 2.7,
|
| 571 |
+
"learning_rate": 5.4083885209713025e-06,
|
| 572 |
+
"loss": 0.2625,
|
| 573 |
+
"step": 455
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 2.73,
|
| 577 |
+
"learning_rate": 4.856512141280353e-06,
|
| 578 |
+
"loss": 0.2577,
|
| 579 |
+
"step": 460
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 2.76,
|
| 583 |
+
"learning_rate": 4.304635761589404e-06,
|
| 584 |
+
"loss": 0.2243,
|
| 585 |
+
"step": 465
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 2.79,
|
| 589 |
+
"learning_rate": 3.752759381898455e-06,
|
| 590 |
+
"loss": 0.267,
|
| 591 |
+
"step": 470
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 2.81,
|
| 595 |
+
"learning_rate": 3.2008830022075054e-06,
|
| 596 |
+
"loss": 0.2599,
|
| 597 |
+
"step": 475
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 2.84,
|
| 601 |
+
"learning_rate": 2.6490066225165563e-06,
|
| 602 |
+
"loss": 0.2431,
|
| 603 |
+
"step": 480
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 2.87,
|
| 607 |
+
"learning_rate": 2.0971302428256073e-06,
|
| 608 |
+
"loss": 0.2688,
|
| 609 |
+
"step": 485
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 2.9,
|
| 613 |
+
"learning_rate": 1.545253863134658e-06,
|
| 614 |
+
"loss": 0.2516,
|
| 615 |
+
"step": 490
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 2.93,
|
| 619 |
+
"learning_rate": 9.933774834437087e-07,
|
| 620 |
+
"loss": 0.2653,
|
| 621 |
+
"step": 495
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 2.96,
|
| 625 |
+
"learning_rate": 4.415011037527594e-07,
|
| 626 |
+
"loss": 0.2801,
|
| 627 |
+
"step": 500
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 2.99,
|
| 631 |
+
"eval_accuracy": 0.9796296296296296,
|
| 632 |
+
"eval_loss": 0.1896590143442154,
|
| 633 |
+
"eval_runtime": 68.8802,
|
| 634 |
+
"eval_samples_per_second": 78.397,
|
| 635 |
+
"eval_steps_per_second": 2.454,
|
| 636 |
+
"step": 504
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 2.99,
|
| 640 |
+
"step": 504,
|
| 641 |
+
"total_flos": 4.999521681367106e+18,
|
| 642 |
+
"train_loss": 0.5908920062439782,
|
| 643 |
+
"train_runtime": 1706.6198,
|
| 644 |
+
"train_samples_per_second": 37.97,
|
| 645 |
+
"train_steps_per_second": 0.295
|
| 646 |
+
}
|
| 647 |
+
],
|
| 648 |
+
"logging_steps": 5,
|
| 649 |
+
"max_steps": 504,
|
| 650 |
+
"num_train_epochs": 3,
|
| 651 |
+
"save_steps": 500,
|
| 652 |
+
"total_flos": 4.999521681367106e+18,
|
| 653 |
+
"trial_name": null,
|
| 654 |
+
"trial_params": null
|
| 655 |
+
}
|