Commit ·
db3738d
1
Parent(s): 94414ed
End of training
Browse files- all_results.json +14 -0
- eval_results.json +9 -0
- train_results.json +8 -0
- trainer_state.json +217 -0
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_accuracy": 0.9264140582097748,
|
| 4 |
+
"eval_loss": 0.3421518802642822,
|
| 5 |
+
"eval_runtime": 10.7108,
|
| 6 |
+
"eval_samples": 5463,
|
| 7 |
+
"eval_samples_per_second": 510.048,
|
| 8 |
+
"eval_steps_per_second": 63.768,
|
| 9 |
+
"train_loss": 0.17483176206333578,
|
| 10 |
+
"train_runtime": 6635.2794,
|
| 11 |
+
"train_samples": 104743,
|
| 12 |
+
"train_samples_per_second": 78.929,
|
| 13 |
+
"train_steps_per_second": 2.467
|
| 14 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_accuracy": 0.9264140582097748,
|
| 4 |
+
"eval_loss": 0.3421518802642822,
|
| 5 |
+
"eval_runtime": 10.7108,
|
| 6 |
+
"eval_samples": 5463,
|
| 7 |
+
"eval_samples_per_second": 510.048,
|
| 8 |
+
"eval_steps_per_second": 63.768
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"train_loss": 0.17483176206333578,
|
| 4 |
+
"train_runtime": 6635.2794,
|
| 5 |
+
"train_samples": 104743,
|
| 6 |
+
"train_samples_per_second": 78.929,
|
| 7 |
+
"train_steps_per_second": 2.467
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,217 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"global_step": 16370,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.15,
|
| 12 |
+
"learning_rate": 1.938912645082468e-05,
|
| 13 |
+
"loss": 0.4316,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.31,
|
| 18 |
+
"learning_rate": 1.8778252901649362e-05,
|
| 19 |
+
"loss": 0.3423,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.46,
|
| 24 |
+
"learning_rate": 1.816737935247404e-05,
|
| 25 |
+
"loss": 0.3194,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.61,
|
| 30 |
+
"learning_rate": 1.755650580329872e-05,
|
| 31 |
+
"loss": 0.2965,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.76,
|
| 36 |
+
"learning_rate": 1.69456322541234e-05,
|
| 37 |
+
"loss": 0.29,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.92,
|
| 42 |
+
"learning_rate": 1.6334758704948076e-05,
|
| 43 |
+
"loss": 0.2752,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 1.07,
|
| 48 |
+
"learning_rate": 1.5723885155772757e-05,
|
| 49 |
+
"loss": 0.2424,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 1.22,
|
| 54 |
+
"learning_rate": 1.5113011606597437e-05,
|
| 55 |
+
"loss": 0.2114,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.37,
|
| 60 |
+
"learning_rate": 1.4502138057422115e-05,
|
| 61 |
+
"loss": 0.2084,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 1.53,
|
| 66 |
+
"learning_rate": 1.3891264508246794e-05,
|
| 67 |
+
"loss": 0.2119,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 1.68,
|
| 72 |
+
"learning_rate": 1.3280390959071474e-05,
|
| 73 |
+
"loss": 0.2075,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.83,
|
| 78 |
+
"learning_rate": 1.2669517409896153e-05,
|
| 79 |
+
"loss": 0.2128,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.99,
|
| 84 |
+
"learning_rate": 1.2058643860720831e-05,
|
| 85 |
+
"loss": 0.198,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 2.14,
|
| 90 |
+
"learning_rate": 1.1447770311545512e-05,
|
| 91 |
+
"loss": 0.1527,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 2.29,
|
| 96 |
+
"learning_rate": 1.083689676237019e-05,
|
| 97 |
+
"loss": 0.1485,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 2.44,
|
| 102 |
+
"learning_rate": 1.0226023213194869e-05,
|
| 103 |
+
"loss": 0.1495,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 2.6,
|
| 108 |
+
"learning_rate": 9.615149664019549e-06,
|
| 109 |
+
"loss": 0.1424,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 2.75,
|
| 114 |
+
"learning_rate": 9.004276114844227e-06,
|
| 115 |
+
"loss": 0.1517,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 2.9,
|
| 120 |
+
"learning_rate": 8.393402565668908e-06,
|
| 121 |
+
"loss": 0.1454,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 3.05,
|
| 126 |
+
"learning_rate": 7.782529016493586e-06,
|
| 127 |
+
"loss": 0.1338,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 3.21,
|
| 132 |
+
"learning_rate": 7.171655467318266e-06,
|
| 133 |
+
"loss": 0.1174,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3.36,
|
| 138 |
+
"learning_rate": 6.560781918142944e-06,
|
| 139 |
+
"loss": 0.1114,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 3.51,
|
| 144 |
+
"learning_rate": 5.949908368967624e-06,
|
| 145 |
+
"loss": 0.1183,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 3.67,
|
| 150 |
+
"learning_rate": 5.339034819792304e-06,
|
| 151 |
+
"loss": 0.1147,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 3.82,
|
| 156 |
+
"learning_rate": 4.728161270616982e-06,
|
| 157 |
+
"loss": 0.1107,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 3.97,
|
| 162 |
+
"learning_rate": 4.117287721441662e-06,
|
| 163 |
+
"loss": 0.104,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 4.12,
|
| 168 |
+
"learning_rate": 3.506414172266341e-06,
|
| 169 |
+
"loss": 0.0871,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 4.28,
|
| 174 |
+
"learning_rate": 2.8955406230910206e-06,
|
| 175 |
+
"loss": 0.0858,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 4.43,
|
| 180 |
+
"learning_rate": 2.2846670739156996e-06,
|
| 181 |
+
"loss": 0.0841,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 4.58,
|
| 186 |
+
"learning_rate": 1.6737935247403788e-06,
|
| 187 |
+
"loss": 0.0841,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 4.73,
|
| 192 |
+
"learning_rate": 1.0629199755650582e-06,
|
| 193 |
+
"loss": 0.085,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 4.89,
|
| 198 |
+
"learning_rate": 4.5204642638973736e-07,
|
| 199 |
+
"loss": 0.0858,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 5.0,
|
| 204 |
+
"step": 16370,
|
| 205 |
+
"total_flos": 0.0,
|
| 206 |
+
"train_loss": 0.17483176206333578,
|
| 207 |
+
"train_runtime": 6635.2794,
|
| 208 |
+
"train_samples_per_second": 78.929,
|
| 209 |
+
"train_steps_per_second": 2.467
|
| 210 |
+
}
|
| 211 |
+
],
|
| 212 |
+
"max_steps": 16370,
|
| 213 |
+
"num_train_epochs": 5,
|
| 214 |
+
"total_flos": 0.0,
|
| 215 |
+
"trial_name": null,
|
| 216 |
+
"trial_params": null
|
| 217 |
+
}
|