End of training
Browse files- all_results.json +13 -0
- eval_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +466 -0
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"eval_accuracy": 0.9933852555211778,
|
| 4 |
+
"eval_loss": 0.020666491240262985,
|
| 5 |
+
"eval_runtime": 84.9137,
|
| 6 |
+
"eval_samples_per_second": 110.383,
|
| 7 |
+
"eval_steps_per_second": 1.731,
|
| 8 |
+
"total_flos": 4.885017433779732e+18,
|
| 9 |
+
"train_loss": 0.39566828084175093,
|
| 10 |
+
"train_runtime": 4610.2266,
|
| 11 |
+
"train_samples_per_second": 32.531,
|
| 12 |
+
"train_steps_per_second": 0.127
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"eval_accuracy": 0.9933852555211778,
|
| 4 |
+
"eval_loss": 0.020666491240262985,
|
| 5 |
+
"eval_runtime": 84.9137,
|
| 6 |
+
"eval_samples_per_second": 110.383,
|
| 7 |
+
"eval_steps_per_second": 1.731
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"total_flos": 4.885017433779732e+18,
|
| 4 |
+
"train_loss": 0.39566828084175093,
|
| 5 |
+
"train_runtime": 4610.2266,
|
| 6 |
+
"train_samples_per_second": 32.531,
|
| 7 |
+
"train_steps_per_second": 0.127
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,466 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9933852555211778,
|
| 3 |
+
"best_model_checkpoint": "PlantDiseaseDetectorSwin/checkpoint-586",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 586,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.034129692832764506,
|
| 13 |
+
"grad_norm": 2.5282273292541504,
|
| 14 |
+
"learning_rate": 8.47457627118644e-06,
|
| 15 |
+
"loss": 3.7855,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.06825938566552901,
|
| 20 |
+
"grad_norm": 2.772298574447632,
|
| 21 |
+
"learning_rate": 1.694915254237288e-05,
|
| 22 |
+
"loss": 3.6821,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.10238907849829351,
|
| 27 |
+
"grad_norm": 4.345815658569336,
|
| 28 |
+
"learning_rate": 2.5423728813559322e-05,
|
| 29 |
+
"loss": 3.4188,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.13651877133105803,
|
| 34 |
+
"grad_norm": 5.607214450836182,
|
| 35 |
+
"learning_rate": 3.389830508474576e-05,
|
| 36 |
+
"loss": 2.8902,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.17064846416382254,
|
| 41 |
+
"grad_norm": 7.322582244873047,
|
| 42 |
+
"learning_rate": 4.2372881355932206e-05,
|
| 43 |
+
"loss": 2.0221,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.20477815699658702,
|
| 48 |
+
"grad_norm": 6.228772163391113,
|
| 49 |
+
"learning_rate": 4.9905123339658444e-05,
|
| 50 |
+
"loss": 1.1851,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.23890784982935154,
|
| 55 |
+
"grad_norm": 6.644685745239258,
|
| 56 |
+
"learning_rate": 4.8956356736242885e-05,
|
| 57 |
+
"loss": 0.7389,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.27303754266211605,
|
| 62 |
+
"grad_norm": 4.455648899078369,
|
| 63 |
+
"learning_rate": 4.8007590132827326e-05,
|
| 64 |
+
"loss": 0.5337,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.30716723549488056,
|
| 69 |
+
"grad_norm": 5.760767459869385,
|
| 70 |
+
"learning_rate": 4.705882352941177e-05,
|
| 71 |
+
"loss": 0.3791,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.3412969283276451,
|
| 76 |
+
"grad_norm": 4.172321796417236,
|
| 77 |
+
"learning_rate": 4.611005692599621e-05,
|
| 78 |
+
"loss": 0.3207,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.37542662116040953,
|
| 83 |
+
"grad_norm": 3.1008100509643555,
|
| 84 |
+
"learning_rate": 4.516129032258064e-05,
|
| 85 |
+
"loss": 0.2725,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.40955631399317405,
|
| 90 |
+
"grad_norm": 3.1988861560821533,
|
| 91 |
+
"learning_rate": 4.421252371916509e-05,
|
| 92 |
+
"loss": 0.2191,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.44368600682593856,
|
| 97 |
+
"grad_norm": 3.0658085346221924,
|
| 98 |
+
"learning_rate": 4.3263757115749525e-05,
|
| 99 |
+
"loss": 0.1984,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.4778156996587031,
|
| 104 |
+
"grad_norm": 3.4681520462036133,
|
| 105 |
+
"learning_rate": 4.2314990512333967e-05,
|
| 106 |
+
"loss": 0.185,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.5119453924914675,
|
| 111 |
+
"grad_norm": 3.333508014678955,
|
| 112 |
+
"learning_rate": 4.136622390891841e-05,
|
| 113 |
+
"loss": 0.1879,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.5460750853242321,
|
| 118 |
+
"grad_norm": 4.511691093444824,
|
| 119 |
+
"learning_rate": 4.041745730550285e-05,
|
| 120 |
+
"loss": 0.1665,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.5802047781569966,
|
| 125 |
+
"grad_norm": 4.345585346221924,
|
| 126 |
+
"learning_rate": 3.946869070208729e-05,
|
| 127 |
+
"loss": 0.1439,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.6143344709897611,
|
| 132 |
+
"grad_norm": 4.149386405944824,
|
| 133 |
+
"learning_rate": 3.851992409867173e-05,
|
| 134 |
+
"loss": 0.1249,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.6484641638225256,
|
| 139 |
+
"grad_norm": 2.935561418533325,
|
| 140 |
+
"learning_rate": 3.757115749525617e-05,
|
| 141 |
+
"loss": 0.1228,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.6825938566552902,
|
| 146 |
+
"grad_norm": 2.6845579147338867,
|
| 147 |
+
"learning_rate": 3.662239089184061e-05,
|
| 148 |
+
"loss": 0.1204,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.7167235494880546,
|
| 153 |
+
"grad_norm": 2.9296746253967285,
|
| 154 |
+
"learning_rate": 3.567362428842505e-05,
|
| 155 |
+
"loss": 0.1271,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.7508532423208191,
|
| 160 |
+
"grad_norm": 3.4379637241363525,
|
| 161 |
+
"learning_rate": 3.472485768500949e-05,
|
| 162 |
+
"loss": 0.1311,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.7849829351535836,
|
| 167 |
+
"grad_norm": 2.2085835933685303,
|
| 168 |
+
"learning_rate": 3.377609108159393e-05,
|
| 169 |
+
"loss": 0.0897,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.8191126279863481,
|
| 174 |
+
"grad_norm": 2.58469557762146,
|
| 175 |
+
"learning_rate": 3.2827324478178365e-05,
|
| 176 |
+
"loss": 0.1059,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.8532423208191127,
|
| 181 |
+
"grad_norm": 2.334817886352539,
|
| 182 |
+
"learning_rate": 3.187855787476281e-05,
|
| 183 |
+
"loss": 0.1,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.8873720136518771,
|
| 188 |
+
"grad_norm": 2.652193784713745,
|
| 189 |
+
"learning_rate": 3.0929791271347254e-05,
|
| 190 |
+
"loss": 0.0913,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.9215017064846417,
|
| 195 |
+
"grad_norm": 3.43129301071167,
|
| 196 |
+
"learning_rate": 2.998102466793169e-05,
|
| 197 |
+
"loss": 0.0979,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.9556313993174061,
|
| 202 |
+
"grad_norm": 3.3414859771728516,
|
| 203 |
+
"learning_rate": 2.9032258064516133e-05,
|
| 204 |
+
"loss": 0.0903,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.9897610921501706,
|
| 209 |
+
"grad_norm": 2.1187517642974854,
|
| 210 |
+
"learning_rate": 2.808349146110057e-05,
|
| 211 |
+
"loss": 0.0717,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.0,
|
| 216 |
+
"eval_accuracy": 0.9863437533340446,
|
| 217 |
+
"eval_loss": 0.039264366030693054,
|
| 218 |
+
"eval_runtime": 85.1157,
|
| 219 |
+
"eval_samples_per_second": 110.121,
|
| 220 |
+
"eval_steps_per_second": 1.727,
|
| 221 |
+
"step": 293
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 1.023890784982935,
|
| 225 |
+
"grad_norm": 2.212913751602173,
|
| 226 |
+
"learning_rate": 2.7134724857685012e-05,
|
| 227 |
+
"loss": 0.0797,
|
| 228 |
+
"step": 300
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 1.0580204778156996,
|
| 232 |
+
"grad_norm": 1.8552402257919312,
|
| 233 |
+
"learning_rate": 2.618595825426945e-05,
|
| 234 |
+
"loss": 0.0649,
|
| 235 |
+
"step": 310
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 1.0921501706484642,
|
| 239 |
+
"grad_norm": 3.584589958190918,
|
| 240 |
+
"learning_rate": 2.523719165085389e-05,
|
| 241 |
+
"loss": 0.068,
|
| 242 |
+
"step": 320
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.1262798634812285,
|
| 246 |
+
"grad_norm": 1.6320329904556274,
|
| 247 |
+
"learning_rate": 2.4288425047438332e-05,
|
| 248 |
+
"loss": 0.0687,
|
| 249 |
+
"step": 330
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 1.1604095563139931,
|
| 253 |
+
"grad_norm": 2.0501434803009033,
|
| 254 |
+
"learning_rate": 2.3339658444022773e-05,
|
| 255 |
+
"loss": 0.0726,
|
| 256 |
+
"step": 340
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 1.1945392491467577,
|
| 260 |
+
"grad_norm": 2.5777804851531982,
|
| 261 |
+
"learning_rate": 2.239089184060721e-05,
|
| 262 |
+
"loss": 0.0742,
|
| 263 |
+
"step": 350
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 1.2286689419795223,
|
| 267 |
+
"grad_norm": 1.8839725255966187,
|
| 268 |
+
"learning_rate": 2.1442125237191652e-05,
|
| 269 |
+
"loss": 0.0719,
|
| 270 |
+
"step": 360
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 1.2627986348122868,
|
| 274 |
+
"grad_norm": 2.721660614013672,
|
| 275 |
+
"learning_rate": 2.049335863377609e-05,
|
| 276 |
+
"loss": 0.0635,
|
| 277 |
+
"step": 370
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 1.2969283276450512,
|
| 281 |
+
"grad_norm": 1.7737756967544556,
|
| 282 |
+
"learning_rate": 1.954459203036053e-05,
|
| 283 |
+
"loss": 0.0634,
|
| 284 |
+
"step": 380
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.3310580204778157,
|
| 288 |
+
"grad_norm": 1.465468168258667,
|
| 289 |
+
"learning_rate": 1.8595825426944972e-05,
|
| 290 |
+
"loss": 0.0626,
|
| 291 |
+
"step": 390
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 1.36518771331058,
|
| 295 |
+
"grad_norm": 1.9786465167999268,
|
| 296 |
+
"learning_rate": 1.7647058823529414e-05,
|
| 297 |
+
"loss": 0.0663,
|
| 298 |
+
"step": 400
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 1.3993174061433447,
|
| 302 |
+
"grad_norm": 1.6383670568466187,
|
| 303 |
+
"learning_rate": 1.6698292220113855e-05,
|
| 304 |
+
"loss": 0.054,
|
| 305 |
+
"step": 410
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.4334470989761092,
|
| 309 |
+
"grad_norm": 3.260576009750366,
|
| 310 |
+
"learning_rate": 1.5749525616698293e-05,
|
| 311 |
+
"loss": 0.0526,
|
| 312 |
+
"step": 420
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 1.4675767918088738,
|
| 316 |
+
"grad_norm": 1.9332594871520996,
|
| 317 |
+
"learning_rate": 1.4800759013282734e-05,
|
| 318 |
+
"loss": 0.0446,
|
| 319 |
+
"step": 430
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 1.5017064846416384,
|
| 323 |
+
"grad_norm": 1.9378633499145508,
|
| 324 |
+
"learning_rate": 1.3851992409867173e-05,
|
| 325 |
+
"loss": 0.0492,
|
| 326 |
+
"step": 440
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.5358361774744027,
|
| 330 |
+
"grad_norm": 2.8215017318725586,
|
| 331 |
+
"learning_rate": 1.2903225806451613e-05,
|
| 332 |
+
"loss": 0.0568,
|
| 333 |
+
"step": 450
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 1.5699658703071673,
|
| 337 |
+
"grad_norm": 1.6412670612335205,
|
| 338 |
+
"learning_rate": 1.1954459203036052e-05,
|
| 339 |
+
"loss": 0.0429,
|
| 340 |
+
"step": 460
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 1.6040955631399316,
|
| 344 |
+
"grad_norm": 1.4835631847381592,
|
| 345 |
+
"learning_rate": 1.1005692599620495e-05,
|
| 346 |
+
"loss": 0.0408,
|
| 347 |
+
"step": 470
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.6382252559726962,
|
| 351 |
+
"grad_norm": 3.138010263442993,
|
| 352 |
+
"learning_rate": 1.0056925996204935e-05,
|
| 353 |
+
"loss": 0.0552,
|
| 354 |
+
"step": 480
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.6723549488054608,
|
| 358 |
+
"grad_norm": 2.551870346069336,
|
| 359 |
+
"learning_rate": 9.108159392789374e-06,
|
| 360 |
+
"loss": 0.0442,
|
| 361 |
+
"step": 490
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 1.7064846416382253,
|
| 365 |
+
"grad_norm": 2.015259027481079,
|
| 366 |
+
"learning_rate": 8.159392789373814e-06,
|
| 367 |
+
"loss": 0.0459,
|
| 368 |
+
"step": 500
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.74061433447099,
|
| 372 |
+
"grad_norm": 1.7179967164993286,
|
| 373 |
+
"learning_rate": 7.210626185958256e-06,
|
| 374 |
+
"loss": 0.037,
|
| 375 |
+
"step": 510
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 1.7747440273037542,
|
| 379 |
+
"grad_norm": 1.3625215291976929,
|
| 380 |
+
"learning_rate": 6.261859582542695e-06,
|
| 381 |
+
"loss": 0.0399,
|
| 382 |
+
"step": 520
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 1.8088737201365188,
|
| 386 |
+
"grad_norm": 1.3097296953201294,
|
| 387 |
+
"learning_rate": 5.313092979127135e-06,
|
| 388 |
+
"loss": 0.0327,
|
| 389 |
+
"step": 530
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.8430034129692832,
|
| 393 |
+
"grad_norm": 1.0734279155731201,
|
| 394 |
+
"learning_rate": 4.364326375711575e-06,
|
| 395 |
+
"loss": 0.0419,
|
| 396 |
+
"step": 540
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 1.8771331058020477,
|
| 400 |
+
"grad_norm": 2.2438840866088867,
|
| 401 |
+
"learning_rate": 3.4155597722960153e-06,
|
| 402 |
+
"loss": 0.0437,
|
| 403 |
+
"step": 550
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 1.9112627986348123,
|
| 407 |
+
"grad_norm": 2.878282308578491,
|
| 408 |
+
"learning_rate": 2.4667931688804556e-06,
|
| 409 |
+
"loss": 0.0439,
|
| 410 |
+
"step": 560
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 1.9453924914675769,
|
| 414 |
+
"grad_norm": 1.2499854564666748,
|
| 415 |
+
"learning_rate": 1.5180265654648958e-06,
|
| 416 |
+
"loss": 0.0398,
|
| 417 |
+
"step": 570
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 1.9795221843003414,
|
| 421 |
+
"grad_norm": 1.1977864503860474,
|
| 422 |
+
"learning_rate": 5.692599620493359e-07,
|
| 423 |
+
"loss": 0.0384,
|
| 424 |
+
"step": 580
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 2.0,
|
| 428 |
+
"eval_accuracy": 0.9933852555211778,
|
| 429 |
+
"eval_loss": 0.020666491240262985,
|
| 430 |
+
"eval_runtime": 83.7475,
|
| 431 |
+
"eval_samples_per_second": 111.92,
|
| 432 |
+
"eval_steps_per_second": 1.755,
|
| 433 |
+
"step": 586
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 2.0,
|
| 437 |
+
"step": 586,
|
| 438 |
+
"total_flos": 4.885017433779732e+18,
|
| 439 |
+
"train_loss": 0.39566828084175093,
|
| 440 |
+
"train_runtime": 4610.2266,
|
| 441 |
+
"train_samples_per_second": 32.531,
|
| 442 |
+
"train_steps_per_second": 0.127
|
| 443 |
+
}
|
| 444 |
+
],
|
| 445 |
+
"logging_steps": 10,
|
| 446 |
+
"max_steps": 586,
|
| 447 |
+
"num_input_tokens_seen": 0,
|
| 448 |
+
"num_train_epochs": 2,
|
| 449 |
+
"save_steps": 500,
|
| 450 |
+
"stateful_callbacks": {
|
| 451 |
+
"TrainerControl": {
|
| 452 |
+
"args": {
|
| 453 |
+
"should_epoch_stop": false,
|
| 454 |
+
"should_evaluate": false,
|
| 455 |
+
"should_log": false,
|
| 456 |
+
"should_save": true,
|
| 457 |
+
"should_training_stop": true
|
| 458 |
+
},
|
| 459 |
+
"attributes": {}
|
| 460 |
+
}
|
| 461 |
+
},
|
| 462 |
+
"total_flos": 4.885017433779732e+18,
|
| 463 |
+
"train_batch_size": 64,
|
| 464 |
+
"trial_name": null,
|
| 465 |
+
"trial_params": null
|
| 466 |
+
}
|