End of training
Browse files- all_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +613 -0
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 40.0,
|
| 3 |
+
"train_loss": 0.4991270453272597,
|
| 4 |
+
"train_runtime": 30553.5264,
|
| 5 |
+
"train_samples": 9801,
|
| 6 |
+
"train_samples_per_second": 12.831,
|
| 7 |
+
"train_steps_per_second": 1.605
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 40.0,
|
| 3 |
+
"train_loss": 0.4991270453272597,
|
| 4 |
+
"train_runtime": 30553.5264,
|
| 5 |
+
"train_samples": 9801,
|
| 6 |
+
"train_samples_per_second": 12.831,
|
| 7 |
+
"train_steps_per_second": 1.605
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,613 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 40.0,
|
| 5 |
+
"global_step": 49040,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.41,
|
| 12 |
+
"learning_rate": 1.9878787878787878e-05,
|
| 13 |
+
"loss": 6.2879,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.82,
|
| 18 |
+
"learning_rate": 1.9674037674037675e-05,
|
| 19 |
+
"loss": 3.3765,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 1.22,
|
| 24 |
+
"learning_rate": 1.9469696969696972e-05,
|
| 25 |
+
"loss": 2.5852,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 1.63,
|
| 30 |
+
"learning_rate": 1.9264946764946766e-05,
|
| 31 |
+
"loss": 1.663,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 2.04,
|
| 36 |
+
"learning_rate": 1.9060196560196563e-05,
|
| 37 |
+
"loss": 1.3474,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 2.45,
|
| 42 |
+
"learning_rate": 1.8855446355446356e-05,
|
| 43 |
+
"loss": 1.1403,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 2.85,
|
| 48 |
+
"learning_rate": 1.8650696150696153e-05,
|
| 49 |
+
"loss": 1.0676,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 3.26,
|
| 54 |
+
"learning_rate": 1.8445945945945947e-05,
|
| 55 |
+
"loss": 0.9461,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 3.67,
|
| 60 |
+
"learning_rate": 1.8241195741195744e-05,
|
| 61 |
+
"loss": 0.8768,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 4.08,
|
| 66 |
+
"learning_rate": 1.8036445536445537e-05,
|
| 67 |
+
"loss": 0.8472,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 4.49,
|
| 72 |
+
"learning_rate": 1.7831695331695334e-05,
|
| 73 |
+
"loss": 0.7818,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 4.89,
|
| 78 |
+
"learning_rate": 1.7627354627354628e-05,
|
| 79 |
+
"loss": 0.759,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 5.3,
|
| 84 |
+
"learning_rate": 1.7422604422604425e-05,
|
| 85 |
+
"loss": 0.6963,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 5.71,
|
| 90 |
+
"learning_rate": 1.721785421785422e-05,
|
| 91 |
+
"loss": 0.6911,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 6.12,
|
| 96 |
+
"learning_rate": 1.7013104013104012e-05,
|
| 97 |
+
"loss": 0.6695,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 6.53,
|
| 102 |
+
"learning_rate": 1.680835380835381e-05,
|
| 103 |
+
"loss": 0.6427,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 6.93,
|
| 108 |
+
"learning_rate": 1.6603603603603603e-05,
|
| 109 |
+
"loss": 0.6228,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 7.34,
|
| 114 |
+
"learning_rate": 1.63992628992629e-05,
|
| 115 |
+
"loss": 0.5752,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 7.75,
|
| 120 |
+
"learning_rate": 1.6194512694512697e-05,
|
| 121 |
+
"loss": 0.5751,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 8.16,
|
| 126 |
+
"learning_rate": 1.598976248976249e-05,
|
| 127 |
+
"loss": 0.5551,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 8.56,
|
| 132 |
+
"learning_rate": 1.5785421785421787e-05,
|
| 133 |
+
"loss": 0.5282,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 8.97,
|
| 138 |
+
"learning_rate": 1.5580671580671584e-05,
|
| 139 |
+
"loss": 0.5259,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 9.38,
|
| 144 |
+
"learning_rate": 1.5375921375921378e-05,
|
| 145 |
+
"loss": 0.4886,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 9.79,
|
| 150 |
+
"learning_rate": 1.5171171171171173e-05,
|
| 151 |
+
"loss": 0.489,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 10.2,
|
| 156 |
+
"learning_rate": 1.4966830466830468e-05,
|
| 157 |
+
"loss": 0.476,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 10.6,
|
| 162 |
+
"learning_rate": 1.4762080262080263e-05,
|
| 163 |
+
"loss": 0.4426,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 11.01,
|
| 168 |
+
"learning_rate": 1.4557330057330059e-05,
|
| 169 |
+
"loss": 0.4613,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 11.42,
|
| 174 |
+
"learning_rate": 1.4352989352989356e-05,
|
| 175 |
+
"loss": 0.4205,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 11.83,
|
| 180 |
+
"learning_rate": 1.414823914823915e-05,
|
| 181 |
+
"loss": 0.4196,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 12.23,
|
| 186 |
+
"learning_rate": 1.3943488943488944e-05,
|
| 187 |
+
"loss": 0.4189,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 12.64,
|
| 192 |
+
"learning_rate": 1.373873873873874e-05,
|
| 193 |
+
"loss": 0.3959,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 13.05,
|
| 198 |
+
"learning_rate": 1.3533988533988535e-05,
|
| 199 |
+
"loss": 0.3857,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 13.46,
|
| 204 |
+
"learning_rate": 1.3329647829647832e-05,
|
| 205 |
+
"loss": 0.3672,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 13.87,
|
| 210 |
+
"learning_rate": 1.3124897624897626e-05,
|
| 211 |
+
"loss": 0.3936,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 14.27,
|
| 216 |
+
"learning_rate": 1.2920147420147423e-05,
|
| 217 |
+
"loss": 0.3475,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 14.68,
|
| 222 |
+
"learning_rate": 1.2715397215397216e-05,
|
| 223 |
+
"loss": 0.3541,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 15.09,
|
| 228 |
+
"learning_rate": 1.2510647010647013e-05,
|
| 229 |
+
"loss": 0.3613,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 15.5,
|
| 234 |
+
"learning_rate": 1.2305896805896807e-05,
|
| 235 |
+
"loss": 0.3399,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 15.91,
|
| 240 |
+
"learning_rate": 1.2101146601146602e-05,
|
| 241 |
+
"loss": 0.3298,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 16.31,
|
| 246 |
+
"learning_rate": 1.1896396396396397e-05,
|
| 247 |
+
"loss": 0.3221,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 16.72,
|
| 252 |
+
"learning_rate": 1.1691646191646192e-05,
|
| 253 |
+
"loss": 0.3166,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 17.13,
|
| 258 |
+
"learning_rate": 1.1486895986895988e-05,
|
| 259 |
+
"loss": 0.3352,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 17.54,
|
| 264 |
+
"learning_rate": 1.1282145782145783e-05,
|
| 265 |
+
"loss": 0.3161,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 17.94,
|
| 270 |
+
"learning_rate": 1.1077395577395578e-05,
|
| 271 |
+
"loss": 0.3143,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 18.35,
|
| 276 |
+
"learning_rate": 1.0872645372645374e-05,
|
| 277 |
+
"loss": 0.3005,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 18.76,
|
| 282 |
+
"learning_rate": 1.0667895167895169e-05,
|
| 283 |
+
"loss": 0.2967,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 19.17,
|
| 288 |
+
"learning_rate": 1.0463144963144964e-05,
|
| 289 |
+
"loss": 0.2963,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 19.58,
|
| 294 |
+
"learning_rate": 1.0258394758394758e-05,
|
| 295 |
+
"loss": 0.2882,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 19.98,
|
| 300 |
+
"learning_rate": 1.0054054054054055e-05,
|
| 301 |
+
"loss": 0.282,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 20.39,
|
| 306 |
+
"learning_rate": 9.84930384930385e-06,
|
| 307 |
+
"loss": 0.2895,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 20.8,
|
| 312 |
+
"learning_rate": 9.644963144963145e-06,
|
| 313 |
+
"loss": 0.2699,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 21.21,
|
| 318 |
+
"learning_rate": 9.44021294021294e-06,
|
| 319 |
+
"loss": 0.2781,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 21.62,
|
| 324 |
+
"learning_rate": 9.235462735462736e-06,
|
| 325 |
+
"loss": 0.2748,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 22.02,
|
| 330 |
+
"learning_rate": 9.031122031122031e-06,
|
| 331 |
+
"loss": 0.2679,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 22.43,
|
| 336 |
+
"learning_rate": 8.826371826371826e-06,
|
| 337 |
+
"loss": 0.2581,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 22.84,
|
| 342 |
+
"learning_rate": 8.621621621621622e-06,
|
| 343 |
+
"loss": 0.2643,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 23.25,
|
| 348 |
+
"learning_rate": 8.416871416871417e-06,
|
| 349 |
+
"loss": 0.2547,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 23.65,
|
| 354 |
+
"learning_rate": 8.212121212121212e-06,
|
| 355 |
+
"loss": 0.253,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 24.06,
|
| 360 |
+
"learning_rate": 8.007371007371007e-06,
|
| 361 |
+
"loss": 0.2576,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 24.47,
|
| 366 |
+
"learning_rate": 7.802620802620803e-06,
|
| 367 |
+
"loss": 0.2373,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 24.88,
|
| 372 |
+
"learning_rate": 7.597870597870598e-06,
|
| 373 |
+
"loss": 0.2571,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 25.29,
|
| 378 |
+
"learning_rate": 7.393120393120393e-06,
|
| 379 |
+
"loss": 0.2461,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 25.69,
|
| 384 |
+
"learning_rate": 7.1883701883701885e-06,
|
| 385 |
+
"loss": 0.243,
|
| 386 |
+
"step": 31500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 26.1,
|
| 390 |
+
"learning_rate": 6.9840294840294846e-06,
|
| 391 |
+
"loss": 0.2399,
|
| 392 |
+
"step": 32000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 26.51,
|
| 396 |
+
"learning_rate": 6.77927927927928e-06,
|
| 397 |
+
"loss": 0.2395,
|
| 398 |
+
"step": 32500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 26.92,
|
| 402 |
+
"learning_rate": 6.574529074529074e-06,
|
| 403 |
+
"loss": 0.2387,
|
| 404 |
+
"step": 33000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 27.32,
|
| 408 |
+
"learning_rate": 6.3697788697788695e-06,
|
| 409 |
+
"loss": 0.2298,
|
| 410 |
+
"step": 33500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 27.73,
|
| 414 |
+
"learning_rate": 6.165028665028665e-06,
|
| 415 |
+
"loss": 0.2285,
|
| 416 |
+
"step": 34000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 28.14,
|
| 420 |
+
"learning_rate": 5.960687960687961e-06,
|
| 421 |
+
"loss": 0.2311,
|
| 422 |
+
"step": 34500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 28.55,
|
| 426 |
+
"learning_rate": 5.755937755937756e-06,
|
| 427 |
+
"loss": 0.2307,
|
| 428 |
+
"step": 35000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 28.96,
|
| 432 |
+
"learning_rate": 5.5511875511875515e-06,
|
| 433 |
+
"loss": 0.231,
|
| 434 |
+
"step": 35500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 29.36,
|
| 438 |
+
"learning_rate": 5.346437346437347e-06,
|
| 439 |
+
"loss": 0.2229,
|
| 440 |
+
"step": 36000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 29.77,
|
| 444 |
+
"learning_rate": 5.141687141687142e-06,
|
| 445 |
+
"loss": 0.2224,
|
| 446 |
+
"step": 36500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 30.18,
|
| 450 |
+
"learning_rate": 4.936936936936937e-06,
|
| 451 |
+
"loss": 0.2209,
|
| 452 |
+
"step": 37000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 30.59,
|
| 456 |
+
"learning_rate": 4.7321867321867326e-06,
|
| 457 |
+
"loss": 0.2277,
|
| 458 |
+
"step": 37500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 31.0,
|
| 462 |
+
"learning_rate": 4.527436527436528e-06,
|
| 463 |
+
"loss": 0.222,
|
| 464 |
+
"step": 38000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 31.4,
|
| 468 |
+
"learning_rate": 4.323095823095824e-06,
|
| 469 |
+
"loss": 0.2129,
|
| 470 |
+
"step": 38500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 31.81,
|
| 474 |
+
"learning_rate": 4.118345618345619e-06,
|
| 475 |
+
"loss": 0.2287,
|
| 476 |
+
"step": 39000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 32.22,
|
| 480 |
+
"learning_rate": 3.9140049140049144e-06,
|
| 481 |
+
"loss": 0.214,
|
| 482 |
+
"step": 39500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 32.63,
|
| 486 |
+
"learning_rate": 3.7092547092547093e-06,
|
| 487 |
+
"loss": 0.211,
|
| 488 |
+
"step": 40000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 33.03,
|
| 492 |
+
"learning_rate": 3.5045045045045046e-06,
|
| 493 |
+
"loss": 0.2204,
|
| 494 |
+
"step": 40500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 33.44,
|
| 498 |
+
"learning_rate": 3.2997542997543003e-06,
|
| 499 |
+
"loss": 0.2155,
|
| 500 |
+
"step": 41000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 33.85,
|
| 504 |
+
"learning_rate": 3.0954135954135955e-06,
|
| 505 |
+
"loss": 0.2163,
|
| 506 |
+
"step": 41500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 34.26,
|
| 510 |
+
"learning_rate": 2.890663390663391e-06,
|
| 511 |
+
"loss": 0.2082,
|
| 512 |
+
"step": 42000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 34.67,
|
| 516 |
+
"learning_rate": 2.685913185913186e-06,
|
| 517 |
+
"loss": 0.203,
|
| 518 |
+
"step": 42500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 35.07,
|
| 522 |
+
"learning_rate": 2.4811629811629813e-06,
|
| 523 |
+
"loss": 0.2044,
|
| 524 |
+
"step": 43000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 35.48,
|
| 528 |
+
"learning_rate": 2.2764127764127766e-06,
|
| 529 |
+
"loss": 0.2084,
|
| 530 |
+
"step": 43500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 35.89,
|
| 534 |
+
"learning_rate": 2.0720720720720723e-06,
|
| 535 |
+
"loss": 0.2088,
|
| 536 |
+
"step": 44000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 36.3,
|
| 540 |
+
"learning_rate": 1.8673218673218673e-06,
|
| 541 |
+
"loss": 0.2004,
|
| 542 |
+
"step": 44500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 36.7,
|
| 546 |
+
"learning_rate": 1.6625716625716626e-06,
|
| 547 |
+
"loss": 0.2051,
|
| 548 |
+
"step": 45000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 37.11,
|
| 552 |
+
"learning_rate": 1.4578214578214581e-06,
|
| 553 |
+
"loss": 0.2077,
|
| 554 |
+
"step": 45500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 37.52,
|
| 558 |
+
"learning_rate": 1.2530712530712532e-06,
|
| 559 |
+
"loss": 0.2019,
|
| 560 |
+
"step": 46000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 37.93,
|
| 564 |
+
"learning_rate": 1.0483210483210485e-06,
|
| 565 |
+
"loss": 0.2011,
|
| 566 |
+
"step": 46500
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 38.34,
|
| 570 |
+
"learning_rate": 8.435708435708436e-07,
|
| 571 |
+
"loss": 0.2092,
|
| 572 |
+
"step": 47000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 38.74,
|
| 576 |
+
"learning_rate": 6.392301392301394e-07,
|
| 577 |
+
"loss": 0.1992,
|
| 578 |
+
"step": 47500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 39.15,
|
| 582 |
+
"learning_rate": 4.3447993447993446e-07,
|
| 583 |
+
"loss": 0.2025,
|
| 584 |
+
"step": 48000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 39.56,
|
| 588 |
+
"learning_rate": 2.2972972972972977e-07,
|
| 589 |
+
"loss": 0.2005,
|
| 590 |
+
"step": 48500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 39.97,
|
| 594 |
+
"learning_rate": 2.4979524979524984e-08,
|
| 595 |
+
"loss": 0.2026,
|
| 596 |
+
"step": 49000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 40.0,
|
| 600 |
+
"step": 49040,
|
| 601 |
+
"total_flos": 7.6145025653737e+19,
|
| 602 |
+
"train_loss": 0.4991270453272597,
|
| 603 |
+
"train_runtime": 30553.5264,
|
| 604 |
+
"train_samples_per_second": 12.831,
|
| 605 |
+
"train_steps_per_second": 1.605
|
| 606 |
+
}
|
| 607 |
+
],
|
| 608 |
+
"max_steps": 49040,
|
| 609 |
+
"num_train_epochs": 40,
|
| 610 |
+
"total_flos": 7.6145025653737e+19,
|
| 611 |
+
"trial_name": null,
|
| 612 |
+
"trial_params": null
|
| 613 |
+
}
|