Upload trainer_state.json
Browse files- trainer_state.json +556 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,556 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.767357481089724,
|
| 5 |
+
"global_step": 45000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 4.948752639239079e-05,
|
| 13 |
+
"loss": 7.0829,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 4.897505278478159e-05,
|
| 19 |
+
"loss": 6.3316,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.09,
|
| 24 |
+
"learning_rate": 4.846257917717238e-05,
|
| 25 |
+
"loss": 5.9867,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.12,
|
| 30 |
+
"learning_rate": 4.795010556956317e-05,
|
| 31 |
+
"loss": 5.7109,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.15,
|
| 36 |
+
"learning_rate": 4.743763196195396e-05,
|
| 37 |
+
"loss": 5.374,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.18,
|
| 42 |
+
"learning_rate": 4.6925158354344756e-05,
|
| 43 |
+
"loss": 4.9609,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.22,
|
| 48 |
+
"learning_rate": 4.6412684746735545e-05,
|
| 49 |
+
"loss": 4.576,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.25,
|
| 54 |
+
"learning_rate": 4.5900211139126335e-05,
|
| 55 |
+
"loss": 4.2963,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.28,
|
| 60 |
+
"learning_rate": 4.5387737531517125e-05,
|
| 61 |
+
"loss": 4.0911,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.31,
|
| 66 |
+
"learning_rate": 4.487526392390792e-05,
|
| 67 |
+
"loss": 3.9354,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.34,
|
| 72 |
+
"learning_rate": 4.436279031629871e-05,
|
| 73 |
+
"loss": 3.7992,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.37,
|
| 78 |
+
"learning_rate": 4.38503167086895e-05,
|
| 79 |
+
"loss": 3.6783,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.4,
|
| 84 |
+
"learning_rate": 4.333784310108029e-05,
|
| 85 |
+
"loss": 3.5666,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.43,
|
| 90 |
+
"learning_rate": 4.282536949347109e-05,
|
| 91 |
+
"loss": 3.472,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.46,
|
| 96 |
+
"learning_rate": 4.231289588586188e-05,
|
| 97 |
+
"loss": 3.3692,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.49,
|
| 102 |
+
"learning_rate": 4.180042227825267e-05,
|
| 103 |
+
"loss": 3.2752,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.52,
|
| 108 |
+
"learning_rate": 4.128794867064346e-05,
|
| 109 |
+
"loss": 3.1837,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.55,
|
| 114 |
+
"learning_rate": 4.0775475063034255e-05,
|
| 115 |
+
"loss": 3.1124,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.58,
|
| 120 |
+
"learning_rate": 4.0263001455425045e-05,
|
| 121 |
+
"loss": 3.0376,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.61,
|
| 126 |
+
"learning_rate": 3.9750527847815835e-05,
|
| 127 |
+
"loss": 2.9563,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.65,
|
| 132 |
+
"learning_rate": 3.923805424020663e-05,
|
| 133 |
+
"loss": 2.892,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.68,
|
| 138 |
+
"learning_rate": 3.872558063259742e-05,
|
| 139 |
+
"loss": 2.8688,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.71,
|
| 144 |
+
"learning_rate": 3.821310702498822e-05,
|
| 145 |
+
"loss": 2.8093,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.74,
|
| 150 |
+
"learning_rate": 3.770063341737901e-05,
|
| 151 |
+
"loss": 2.7535,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.77,
|
| 156 |
+
"learning_rate": 3.71881598097698e-05,
|
| 157 |
+
"loss": 2.6915,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.8,
|
| 162 |
+
"learning_rate": 3.6675686202160595e-05,
|
| 163 |
+
"loss": 2.6404,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.83,
|
| 168 |
+
"learning_rate": 3.6163212594551385e-05,
|
| 169 |
+
"loss": 2.5972,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.86,
|
| 174 |
+
"learning_rate": 3.5650738986942175e-05,
|
| 175 |
+
"loss": 2.5741,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.89,
|
| 180 |
+
"learning_rate": 3.513826537933297e-05,
|
| 181 |
+
"loss": 2.5309,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.92,
|
| 186 |
+
"learning_rate": 3.462579177172376e-05,
|
| 187 |
+
"loss": 2.5045,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.95,
|
| 192 |
+
"learning_rate": 3.411331816411455e-05,
|
| 193 |
+
"loss": 2.4531,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.98,
|
| 198 |
+
"learning_rate": 3.360084455650534e-05,
|
| 199 |
+
"loss": 2.4361,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.01,
|
| 204 |
+
"learning_rate": 3.308837094889614e-05,
|
| 205 |
+
"loss": 2.3951,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.05,
|
| 210 |
+
"learning_rate": 3.257589734128693e-05,
|
| 211 |
+
"loss": 2.3564,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.08,
|
| 216 |
+
"learning_rate": 3.206342373367772e-05,
|
| 217 |
+
"loss": 2.3385,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.11,
|
| 222 |
+
"learning_rate": 3.155095012606851e-05,
|
| 223 |
+
"loss": 2.3167,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.14,
|
| 228 |
+
"learning_rate": 3.1038476518459304e-05,
|
| 229 |
+
"loss": 2.285,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.17,
|
| 234 |
+
"learning_rate": 3.0526002910850094e-05,
|
| 235 |
+
"loss": 2.2793,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.2,
|
| 240 |
+
"learning_rate": 3.0013529303240884e-05,
|
| 241 |
+
"loss": 2.2349,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.23,
|
| 246 |
+
"learning_rate": 2.9501055695631674e-05,
|
| 247 |
+
"loss": 2.2214,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.26,
|
| 252 |
+
"learning_rate": 2.898858208802247e-05,
|
| 253 |
+
"loss": 2.1906,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.29,
|
| 258 |
+
"learning_rate": 2.847610848041326e-05,
|
| 259 |
+
"loss": 2.1662,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.32,
|
| 264 |
+
"learning_rate": 2.796363487280405e-05,
|
| 265 |
+
"loss": 2.1746,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.35,
|
| 270 |
+
"learning_rate": 2.745116126519484e-05,
|
| 271 |
+
"loss": 2.1394,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.38,
|
| 276 |
+
"learning_rate": 2.6938687657585637e-05,
|
| 277 |
+
"loss": 2.1186,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.41,
|
| 282 |
+
"learning_rate": 2.6426214049976427e-05,
|
| 283 |
+
"loss": 2.0967,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.45,
|
| 288 |
+
"learning_rate": 2.5913740442367217e-05,
|
| 289 |
+
"loss": 2.0976,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.48,
|
| 294 |
+
"learning_rate": 2.5401266834758007e-05,
|
| 295 |
+
"loss": 2.0787,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.51,
|
| 300 |
+
"learning_rate": 2.48887932271488e-05,
|
| 301 |
+
"loss": 2.0739,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.54,
|
| 306 |
+
"learning_rate": 2.4376319619539594e-05,
|
| 307 |
+
"loss": 2.0548,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.57,
|
| 312 |
+
"learning_rate": 2.3863846011930387e-05,
|
| 313 |
+
"loss": 2.0068,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.6,
|
| 318 |
+
"learning_rate": 2.3351372404321177e-05,
|
| 319 |
+
"loss": 2.0166,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.63,
|
| 324 |
+
"learning_rate": 2.283889879671197e-05,
|
| 325 |
+
"loss": 1.9973,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.66,
|
| 330 |
+
"learning_rate": 2.232642518910276e-05,
|
| 331 |
+
"loss": 1.9874,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.69,
|
| 336 |
+
"learning_rate": 2.1813951581493554e-05,
|
| 337 |
+
"loss": 1.9941,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.72,
|
| 342 |
+
"learning_rate": 2.1301477973884347e-05,
|
| 343 |
+
"loss": 1.9653,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.75,
|
| 348 |
+
"learning_rate": 2.078900436627514e-05,
|
| 349 |
+
"loss": 1.9548,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.78,
|
| 354 |
+
"learning_rate": 2.027653075866593e-05,
|
| 355 |
+
"loss": 1.9431,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.81,
|
| 360 |
+
"learning_rate": 1.9764057151056724e-05,
|
| 361 |
+
"loss": 1.9327,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.84,
|
| 366 |
+
"learning_rate": 1.9251583543447514e-05,
|
| 367 |
+
"loss": 1.9279,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.88,
|
| 372 |
+
"learning_rate": 1.8739109935838307e-05,
|
| 373 |
+
"loss": 1.9013,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.91,
|
| 378 |
+
"learning_rate": 1.8226636328229097e-05,
|
| 379 |
+
"loss": 1.8981,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 1.94,
|
| 384 |
+
"learning_rate": 1.771416272061989e-05,
|
| 385 |
+
"loss": 1.9058,
|
| 386 |
+
"step": 31500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 1.97,
|
| 390 |
+
"learning_rate": 1.720168911301068e-05,
|
| 391 |
+
"loss": 1.8961,
|
| 392 |
+
"step": 32000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.0,
|
| 396 |
+
"learning_rate": 1.6689215505401473e-05,
|
| 397 |
+
"loss": 1.8763,
|
| 398 |
+
"step": 32500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.03,
|
| 402 |
+
"learning_rate": 1.6176741897792263e-05,
|
| 403 |
+
"loss": 1.8504,
|
| 404 |
+
"step": 33000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.06,
|
| 408 |
+
"learning_rate": 1.5664268290183057e-05,
|
| 409 |
+
"loss": 1.8633,
|
| 410 |
+
"step": 33500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.09,
|
| 414 |
+
"learning_rate": 1.5151794682573847e-05,
|
| 415 |
+
"loss": 1.8527,
|
| 416 |
+
"step": 34000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.12,
|
| 420 |
+
"learning_rate": 1.463932107496464e-05,
|
| 421 |
+
"loss": 1.837,
|
| 422 |
+
"step": 34500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.15,
|
| 426 |
+
"learning_rate": 1.4126847467355431e-05,
|
| 427 |
+
"loss": 1.8145,
|
| 428 |
+
"step": 35000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.18,
|
| 432 |
+
"learning_rate": 1.3614373859746223e-05,
|
| 433 |
+
"loss": 1.8368,
|
| 434 |
+
"step": 35500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.21,
|
| 438 |
+
"learning_rate": 1.3101900252137015e-05,
|
| 439 |
+
"loss": 1.8084,
|
| 440 |
+
"step": 36000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.24,
|
| 444 |
+
"learning_rate": 1.2589426644527808e-05,
|
| 445 |
+
"loss": 1.8235,
|
| 446 |
+
"step": 36500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.28,
|
| 450 |
+
"learning_rate": 1.20769530369186e-05,
|
| 451 |
+
"loss": 1.8042,
|
| 452 |
+
"step": 37000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.31,
|
| 456 |
+
"learning_rate": 1.1564479429309391e-05,
|
| 457 |
+
"loss": 1.7855,
|
| 458 |
+
"step": 37500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.34,
|
| 462 |
+
"learning_rate": 1.1052005821700183e-05,
|
| 463 |
+
"loss": 1.7904,
|
| 464 |
+
"step": 38000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.37,
|
| 468 |
+
"learning_rate": 1.0539532214090975e-05,
|
| 469 |
+
"loss": 1.7749,
|
| 470 |
+
"step": 38500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.4,
|
| 474 |
+
"learning_rate": 1.0027058606481766e-05,
|
| 475 |
+
"loss": 1.785,
|
| 476 |
+
"step": 39000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.43,
|
| 480 |
+
"learning_rate": 9.514584998872558e-06,
|
| 481 |
+
"loss": 1.7599,
|
| 482 |
+
"step": 39500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.46,
|
| 486 |
+
"learning_rate": 9.00211139126335e-06,
|
| 487 |
+
"loss": 1.774,
|
| 488 |
+
"step": 40000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.49,
|
| 492 |
+
"learning_rate": 8.489637783654141e-06,
|
| 493 |
+
"loss": 1.7442,
|
| 494 |
+
"step": 40500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.52,
|
| 498 |
+
"learning_rate": 7.977164176044934e-06,
|
| 499 |
+
"loss": 1.7571,
|
| 500 |
+
"step": 41000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.55,
|
| 504 |
+
"learning_rate": 7.464690568435725e-06,
|
| 505 |
+
"loss": 1.751,
|
| 506 |
+
"step": 41500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.58,
|
| 510 |
+
"learning_rate": 6.952216960826518e-06,
|
| 511 |
+
"loss": 1.7367,
|
| 512 |
+
"step": 42000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.61,
|
| 516 |
+
"learning_rate": 6.439743353217309e-06,
|
| 517 |
+
"loss": 1.7437,
|
| 518 |
+
"step": 42500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.64,
|
| 522 |
+
"learning_rate": 5.927269745608102e-06,
|
| 523 |
+
"loss": 1.7467,
|
| 524 |
+
"step": 43000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.68,
|
| 528 |
+
"learning_rate": 5.414796137998893e-06,
|
| 529 |
+
"loss": 1.7417,
|
| 530 |
+
"step": 43500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.71,
|
| 534 |
+
"learning_rate": 4.902322530389685e-06,
|
| 535 |
+
"loss": 1.7392,
|
| 536 |
+
"step": 44000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.74,
|
| 540 |
+
"learning_rate": 4.389848922780477e-06,
|
| 541 |
+
"loss": 1.7313,
|
| 542 |
+
"step": 44500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.77,
|
| 546 |
+
"learning_rate": 3.877375315171269e-06,
|
| 547 |
+
"loss": 1.745,
|
| 548 |
+
"step": 45000
|
| 549 |
+
}
|
| 550 |
+
],
|
| 551 |
+
"max_steps": 48783,
|
| 552 |
+
"num_train_epochs": 3,
|
| 553 |
+
"total_flos": 9.548707870117478e+16,
|
| 554 |
+
"trial_name": null,
|
| 555 |
+
"trial_params": null
|
| 556 |
+
}
|