Commit
·
d036598
1
Parent(s):
f6bed27
Upload trainer_state.json
Browse files- trainer_state.json +674 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,674 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6546083688735962,
|
| 3 |
+
"best_model_checkpoint": "models_gitignored/distilbert-base-uncased-finetuned-sentence-classification/checkpoint-12626",
|
| 4 |
+
"epoch": 4.0,
|
| 5 |
+
"global_step": 50504,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.04,
|
| 12 |
+
"learning_rate": 1.984159670521147e-05,
|
| 13 |
+
"loss": 1.177,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.08,
|
| 18 |
+
"learning_rate": 1.968319341042294e-05,
|
| 19 |
+
"loss": 0.9186,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.12,
|
| 24 |
+
"learning_rate": 1.9524790115634406e-05,
|
| 25 |
+
"loss": 0.8261,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.16,
|
| 30 |
+
"learning_rate": 1.9366386820845877e-05,
|
| 31 |
+
"loss": 0.8177,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.2,
|
| 36 |
+
"learning_rate": 1.9207983526057344e-05,
|
| 37 |
+
"loss": 0.7788,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.24,
|
| 42 |
+
"learning_rate": 1.904958023126881e-05,
|
| 43 |
+
"loss": 0.7519,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.28,
|
| 48 |
+
"learning_rate": 1.889117693648028e-05,
|
| 49 |
+
"loss": 0.7243,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.32,
|
| 54 |
+
"learning_rate": 1.8732773641691748e-05,
|
| 55 |
+
"loss": 0.7431,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.36,
|
| 60 |
+
"learning_rate": 1.8574370346903215e-05,
|
| 61 |
+
"loss": 0.7297,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.4,
|
| 66 |
+
"learning_rate": 1.8415967052114686e-05,
|
| 67 |
+
"loss": 0.7061,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.44,
|
| 72 |
+
"learning_rate": 1.8257563757326153e-05,
|
| 73 |
+
"loss": 0.698,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.48,
|
| 78 |
+
"learning_rate": 1.8099160462537623e-05,
|
| 79 |
+
"loss": 0.7074,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.51,
|
| 84 |
+
"learning_rate": 1.794075716774909e-05,
|
| 85 |
+
"loss": 0.6975,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.55,
|
| 90 |
+
"learning_rate": 1.7782353872960557e-05,
|
| 91 |
+
"loss": 0.6943,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.59,
|
| 96 |
+
"learning_rate": 1.7623950578172028e-05,
|
| 97 |
+
"loss": 0.7045,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.63,
|
| 102 |
+
"learning_rate": 1.7465547283383495e-05,
|
| 103 |
+
"loss": 0.6764,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.67,
|
| 108 |
+
"learning_rate": 1.730714398859496e-05,
|
| 109 |
+
"loss": 0.6722,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.71,
|
| 114 |
+
"learning_rate": 1.7148740693806432e-05,
|
| 115 |
+
"loss": 0.6884,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.75,
|
| 120 |
+
"learning_rate": 1.69903373990179e-05,
|
| 121 |
+
"loss": 0.6903,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.79,
|
| 126 |
+
"learning_rate": 1.683193410422937e-05,
|
| 127 |
+
"loss": 0.6854,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.83,
|
| 132 |
+
"learning_rate": 1.667353080944084e-05,
|
| 133 |
+
"loss": 0.6861,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.87,
|
| 138 |
+
"learning_rate": 1.6515127514652307e-05,
|
| 139 |
+
"loss": 0.6894,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.91,
|
| 144 |
+
"learning_rate": 1.6356724219863774e-05,
|
| 145 |
+
"loss": 0.6754,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.95,
|
| 150 |
+
"learning_rate": 1.6198320925075244e-05,
|
| 151 |
+
"loss": 0.6918,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.99,
|
| 156 |
+
"learning_rate": 1.603991763028671e-05,
|
| 157 |
+
"loss": 0.6746,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 1.0,
|
| 162 |
+
"eval_accuracy": 0.778547431856752,
|
| 163 |
+
"eval_f1": 0.7757136063142371,
|
| 164 |
+
"eval_kappa": 0.7014269317598121,
|
| 165 |
+
"eval_loss": 0.6546083688735962,
|
| 166 |
+
"eval_precision": 0.7762465189891471,
|
| 167 |
+
"eval_recall": 0.778547431856752,
|
| 168 |
+
"eval_runtime": 150.616,
|
| 169 |
+
"eval_samples_per_second": 191.699,
|
| 170 |
+
"eval_steps_per_second": 23.968,
|
| 171 |
+
"step": 12626
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 1.03,
|
| 175 |
+
"learning_rate": 1.5881514335498182e-05,
|
| 176 |
+
"loss": 0.5694,
|
| 177 |
+
"step": 13000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 1.07,
|
| 181 |
+
"learning_rate": 1.572311104070965e-05,
|
| 182 |
+
"loss": 0.5509,
|
| 183 |
+
"step": 13500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 1.11,
|
| 187 |
+
"learning_rate": 1.5564707745921116e-05,
|
| 188 |
+
"loss": 0.5466,
|
| 189 |
+
"step": 14000
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 1.15,
|
| 193 |
+
"learning_rate": 1.5406304451132586e-05,
|
| 194 |
+
"loss": 0.5544,
|
| 195 |
+
"step": 14500
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 1.19,
|
| 199 |
+
"learning_rate": 1.5247901156344053e-05,
|
| 200 |
+
"loss": 0.566,
|
| 201 |
+
"step": 15000
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 1.23,
|
| 205 |
+
"learning_rate": 1.5089497861555522e-05,
|
| 206 |
+
"loss": 0.5773,
|
| 207 |
+
"step": 15500
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 1.27,
|
| 211 |
+
"learning_rate": 1.4931094566766991e-05,
|
| 212 |
+
"loss": 0.6056,
|
| 213 |
+
"step": 16000
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 1.31,
|
| 217 |
+
"learning_rate": 1.4772691271978458e-05,
|
| 218 |
+
"loss": 0.5532,
|
| 219 |
+
"step": 16500
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 1.35,
|
| 223 |
+
"learning_rate": 1.4614287977189927e-05,
|
| 224 |
+
"loss": 0.5764,
|
| 225 |
+
"step": 17000
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 1.39,
|
| 229 |
+
"learning_rate": 1.4455884682401395e-05,
|
| 230 |
+
"loss": 0.5499,
|
| 231 |
+
"step": 17500
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 1.43,
|
| 235 |
+
"learning_rate": 1.4297481387612864e-05,
|
| 236 |
+
"loss": 0.5695,
|
| 237 |
+
"step": 18000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 1.47,
|
| 241 |
+
"learning_rate": 1.4139078092824333e-05,
|
| 242 |
+
"loss": 0.5823,
|
| 243 |
+
"step": 18500
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 1.5,
|
| 247 |
+
"learning_rate": 1.39806747980358e-05,
|
| 248 |
+
"loss": 0.5329,
|
| 249 |
+
"step": 19000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 1.54,
|
| 253 |
+
"learning_rate": 1.3822271503247269e-05,
|
| 254 |
+
"loss": 0.5661,
|
| 255 |
+
"step": 19500
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 1.58,
|
| 259 |
+
"learning_rate": 1.3663868208458737e-05,
|
| 260 |
+
"loss": 0.5851,
|
| 261 |
+
"step": 20000
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 1.62,
|
| 265 |
+
"learning_rate": 1.3505464913670206e-05,
|
| 266 |
+
"loss": 0.5859,
|
| 267 |
+
"step": 20500
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 1.66,
|
| 271 |
+
"learning_rate": 1.3347061618881673e-05,
|
| 272 |
+
"loss": 0.5569,
|
| 273 |
+
"step": 21000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 1.7,
|
| 277 |
+
"learning_rate": 1.3188658324093142e-05,
|
| 278 |
+
"loss": 0.5473,
|
| 279 |
+
"step": 21500
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 1.74,
|
| 283 |
+
"learning_rate": 1.303025502930461e-05,
|
| 284 |
+
"loss": 0.5635,
|
| 285 |
+
"step": 22000
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 1.78,
|
| 289 |
+
"learning_rate": 1.287185173451608e-05,
|
| 290 |
+
"loss": 0.5848,
|
| 291 |
+
"step": 22500
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 1.82,
|
| 295 |
+
"learning_rate": 1.2713448439727546e-05,
|
| 296 |
+
"loss": 0.5885,
|
| 297 |
+
"step": 23000
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 1.86,
|
| 301 |
+
"learning_rate": 1.2555045144939015e-05,
|
| 302 |
+
"loss": 0.5502,
|
| 303 |
+
"step": 23500
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 1.9,
|
| 307 |
+
"learning_rate": 1.2396641850150484e-05,
|
| 308 |
+
"loss": 0.5999,
|
| 309 |
+
"step": 24000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 1.94,
|
| 313 |
+
"learning_rate": 1.2238238555361953e-05,
|
| 314 |
+
"loss": 0.5639,
|
| 315 |
+
"step": 24500
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 1.98,
|
| 319 |
+
"learning_rate": 1.207983526057342e-05,
|
| 320 |
+
"loss": 0.5664,
|
| 321 |
+
"step": 25000
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 2.0,
|
| 325 |
+
"eval_accuracy": 0.7842274789595817,
|
| 326 |
+
"eval_f1": 0.7824161012799614,
|
| 327 |
+
"eval_kappa": 0.7102691347246439,
|
| 328 |
+
"eval_loss": 0.7084277868270874,
|
| 329 |
+
"eval_precision": 0.7870631351341026,
|
| 330 |
+
"eval_recall": 0.7842274789595817,
|
| 331 |
+
"eval_runtime": 152.0087,
|
| 332 |
+
"eval_samples_per_second": 189.943,
|
| 333 |
+
"eval_steps_per_second": 23.749,
|
| 334 |
+
"step": 25252
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 2.02,
|
| 338 |
+
"learning_rate": 1.1921431965784888e-05,
|
| 339 |
+
"loss": 0.4751,
|
| 340 |
+
"step": 25500
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 2.06,
|
| 344 |
+
"learning_rate": 1.1763028670996357e-05,
|
| 345 |
+
"loss": 0.4208,
|
| 346 |
+
"step": 26000
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 2.1,
|
| 350 |
+
"learning_rate": 1.1604625376207826e-05,
|
| 351 |
+
"loss": 0.4563,
|
| 352 |
+
"step": 26500
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 2.14,
|
| 356 |
+
"learning_rate": 1.1446222081419293e-05,
|
| 357 |
+
"loss": 0.4513,
|
| 358 |
+
"step": 27000
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 2.18,
|
| 362 |
+
"learning_rate": 1.1287818786630762e-05,
|
| 363 |
+
"loss": 0.4385,
|
| 364 |
+
"step": 27500
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 2.22,
|
| 368 |
+
"learning_rate": 1.112941549184223e-05,
|
| 369 |
+
"loss": 0.4493,
|
| 370 |
+
"step": 28000
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 2.26,
|
| 374 |
+
"learning_rate": 1.0971012197053699e-05,
|
| 375 |
+
"loss": 0.4533,
|
| 376 |
+
"step": 28500
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 2.3,
|
| 380 |
+
"learning_rate": 1.0812608902265168e-05,
|
| 381 |
+
"loss": 0.4365,
|
| 382 |
+
"step": 29000
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 2.34,
|
| 386 |
+
"learning_rate": 1.0654205607476635e-05,
|
| 387 |
+
"loss": 0.4205,
|
| 388 |
+
"step": 29500
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 2.38,
|
| 392 |
+
"learning_rate": 1.0495802312688104e-05,
|
| 393 |
+
"loss": 0.4538,
|
| 394 |
+
"step": 30000
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 2.42,
|
| 398 |
+
"learning_rate": 1.0337399017899574e-05,
|
| 399 |
+
"loss": 0.4653,
|
| 400 |
+
"step": 30500
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 2.46,
|
| 404 |
+
"learning_rate": 1.0178995723111043e-05,
|
| 405 |
+
"loss": 0.436,
|
| 406 |
+
"step": 31000
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 2.49,
|
| 410 |
+
"learning_rate": 1.0020592428322511e-05,
|
| 411 |
+
"loss": 0.4249,
|
| 412 |
+
"step": 31500
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 2.53,
|
| 416 |
+
"learning_rate": 9.862189133533979e-06,
|
| 417 |
+
"loss": 0.4441,
|
| 418 |
+
"step": 32000
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 2.57,
|
| 422 |
+
"learning_rate": 9.703785838745447e-06,
|
| 423 |
+
"loss": 0.4544,
|
| 424 |
+
"step": 32500
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 2.61,
|
| 428 |
+
"learning_rate": 9.545382543956914e-06,
|
| 429 |
+
"loss": 0.446,
|
| 430 |
+
"step": 33000
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 2.65,
|
| 434 |
+
"learning_rate": 9.386979249168383e-06,
|
| 435 |
+
"loss": 0.4457,
|
| 436 |
+
"step": 33500
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 2.69,
|
| 440 |
+
"learning_rate": 9.228575954379852e-06,
|
| 441 |
+
"loss": 0.4472,
|
| 442 |
+
"step": 34000
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 2.73,
|
| 446 |
+
"learning_rate": 9.07017265959132e-06,
|
| 447 |
+
"loss": 0.4728,
|
| 448 |
+
"step": 34500
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 2.77,
|
| 452 |
+
"learning_rate": 8.911769364802788e-06,
|
| 453 |
+
"loss": 0.4455,
|
| 454 |
+
"step": 35000
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 2.81,
|
| 458 |
+
"learning_rate": 8.753366070014258e-06,
|
| 459 |
+
"loss": 0.4622,
|
| 460 |
+
"step": 35500
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 2.85,
|
| 464 |
+
"learning_rate": 8.594962775225727e-06,
|
| 465 |
+
"loss": 0.4566,
|
| 466 |
+
"step": 36000
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 2.89,
|
| 470 |
+
"learning_rate": 8.436559480437194e-06,
|
| 471 |
+
"loss": 0.4185,
|
| 472 |
+
"step": 36500
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 2.93,
|
| 476 |
+
"learning_rate": 8.278156185648662e-06,
|
| 477 |
+
"loss": 0.4351,
|
| 478 |
+
"step": 37000
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 2.97,
|
| 482 |
+
"learning_rate": 8.119752890860131e-06,
|
| 483 |
+
"loss": 0.4566,
|
| 484 |
+
"step": 37500
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 3.0,
|
| 488 |
+
"eval_accuracy": 0.7792747549613827,
|
| 489 |
+
"eval_f1": 0.7780530572128844,
|
| 490 |
+
"eval_kappa": 0.7039084509242337,
|
| 491 |
+
"eval_loss": 0.7550320029258728,
|
| 492 |
+
"eval_precision": 0.7813618395098462,
|
| 493 |
+
"eval_recall": 0.7792747549613827,
|
| 494 |
+
"eval_runtime": 151.5948,
|
| 495 |
+
"eval_samples_per_second": 190.462,
|
| 496 |
+
"eval_steps_per_second": 23.813,
|
| 497 |
+
"step": 37878
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 3.01,
|
| 501 |
+
"learning_rate": 7.9613495960716e-06,
|
| 502 |
+
"loss": 0.4023,
|
| 503 |
+
"step": 38000
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 3.05,
|
| 507 |
+
"learning_rate": 7.802946301283067e-06,
|
| 508 |
+
"loss": 0.3119,
|
| 509 |
+
"step": 38500
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 3.09,
|
| 513 |
+
"learning_rate": 7.644543006494536e-06,
|
| 514 |
+
"loss": 0.3205,
|
| 515 |
+
"step": 39000
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 3.13,
|
| 519 |
+
"learning_rate": 7.4861397117060044e-06,
|
| 520 |
+
"loss": 0.35,
|
| 521 |
+
"step": 39500
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 3.17,
|
| 525 |
+
"learning_rate": 7.327736416917472e-06,
|
| 526 |
+
"loss": 0.3333,
|
| 527 |
+
"step": 40000
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 3.21,
|
| 531 |
+
"learning_rate": 7.169333122128941e-06,
|
| 532 |
+
"loss": 0.3563,
|
| 533 |
+
"step": 40500
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 3.25,
|
| 537 |
+
"learning_rate": 7.010929827340409e-06,
|
| 538 |
+
"loss": 0.3388,
|
| 539 |
+
"step": 41000
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 3.29,
|
| 543 |
+
"learning_rate": 6.852526532551878e-06,
|
| 544 |
+
"loss": 0.3422,
|
| 545 |
+
"step": 41500
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 3.33,
|
| 549 |
+
"learning_rate": 6.6941232377633456e-06,
|
| 550 |
+
"loss": 0.3336,
|
| 551 |
+
"step": 42000
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 3.37,
|
| 555 |
+
"learning_rate": 6.535719942974814e-06,
|
| 556 |
+
"loss": 0.3433,
|
| 557 |
+
"step": 42500
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 3.41,
|
| 561 |
+
"learning_rate": 6.377316648186282e-06,
|
| 562 |
+
"loss": 0.3574,
|
| 563 |
+
"step": 43000
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 3.45,
|
| 567 |
+
"learning_rate": 6.218913353397751e-06,
|
| 568 |
+
"loss": 0.3105,
|
| 569 |
+
"step": 43500
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 3.48,
|
| 573 |
+
"learning_rate": 6.060510058609219e-06,
|
| 574 |
+
"loss": 0.3545,
|
| 575 |
+
"step": 44000
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 3.52,
|
| 579 |
+
"learning_rate": 5.9021067638206875e-06,
|
| 580 |
+
"loss": 0.3469,
|
| 581 |
+
"step": 44500
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 3.56,
|
| 585 |
+
"learning_rate": 5.7437034690321554e-06,
|
| 586 |
+
"loss": 0.35,
|
| 587 |
+
"step": 45000
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 3.6,
|
| 591 |
+
"learning_rate": 5.585300174243625e-06,
|
| 592 |
+
"loss": 0.3425,
|
| 593 |
+
"step": 45500
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 3.64,
|
| 597 |
+
"learning_rate": 5.426896879455094e-06,
|
| 598 |
+
"loss": 0.3203,
|
| 599 |
+
"step": 46000
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 3.68,
|
| 603 |
+
"learning_rate": 5.268493584666562e-06,
|
| 604 |
+
"loss": 0.3549,
|
| 605 |
+
"step": 46500
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 3.72,
|
| 609 |
+
"learning_rate": 5.11009028987803e-06,
|
| 610 |
+
"loss": 0.3326,
|
| 611 |
+
"step": 47000
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 3.76,
|
| 615 |
+
"learning_rate": 4.951686995089498e-06,
|
| 616 |
+
"loss": 0.3328,
|
| 617 |
+
"step": 47500
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 3.8,
|
| 621 |
+
"learning_rate": 4.793283700300967e-06,
|
| 622 |
+
"loss": 0.3376,
|
| 623 |
+
"step": 48000
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 3.84,
|
| 627 |
+
"learning_rate": 4.634880405512435e-06,
|
| 628 |
+
"loss": 0.3285,
|
| 629 |
+
"step": 48500
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 3.88,
|
| 633 |
+
"learning_rate": 4.476477110723904e-06,
|
| 634 |
+
"loss": 0.361,
|
| 635 |
+
"step": 49000
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 3.92,
|
| 639 |
+
"learning_rate": 4.3180738159353715e-06,
|
| 640 |
+
"loss": 0.3487,
|
| 641 |
+
"step": 49500
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 3.96,
|
| 645 |
+
"learning_rate": 4.15967052114684e-06,
|
| 646 |
+
"loss": 0.3511,
|
| 647 |
+
"step": 50000
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 4.0,
|
| 651 |
+
"learning_rate": 4.001267226358309e-06,
|
| 652 |
+
"loss": 0.3528,
|
| 653 |
+
"step": 50500
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 4.0,
|
| 657 |
+
"eval_accuracy": 0.7778201087521214,
|
| 658 |
+
"eval_f1": 0.7760540283760836,
|
| 659 |
+
"eval_kappa": 0.7007853270915267,
|
| 660 |
+
"eval_loss": 0.9983726739883423,
|
| 661 |
+
"eval_precision": 0.7770574865770817,
|
| 662 |
+
"eval_recall": 0.7778201087521214,
|
| 663 |
+
"eval_runtime": 151.818,
|
| 664 |
+
"eval_samples_per_second": 190.182,
|
| 665 |
+
"eval_steps_per_second": 23.778,
|
| 666 |
+
"step": 50504
|
| 667 |
+
}
|
| 668 |
+
],
|
| 669 |
+
"max_steps": 63130,
|
| 670 |
+
"num_train_epochs": 5,
|
| 671 |
+
"total_flos": 5.330919857118713e+16,
|
| 672 |
+
"trial_name": null,
|
| 673 |
+
"trial_params": null
|
| 674 |
+
}
|