End of training
Browse files- all_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +397 -0
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 25.0,
|
| 3 |
+
"train_loss": 0.42298770703125,
|
| 4 |
+
"train_runtime": 13299.417,
|
| 5 |
+
"train_samples": 10000,
|
| 6 |
+
"train_samples_per_second": 18.798,
|
| 7 |
+
"train_steps_per_second": 2.35
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 25.0,
|
| 3 |
+
"train_loss": 0.42298770703125,
|
| 4 |
+
"train_runtime": 13299.417,
|
| 5 |
+
"train_samples": 10000,
|
| 6 |
+
"train_samples_per_second": 18.798,
|
| 7 |
+
"train_steps_per_second": 2.35
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,397 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 25.0,
|
| 5 |
+
"global_step": 31250,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.4,
|
| 12 |
+
"learning_rate": 1.9811272141706926e-05,
|
| 13 |
+
"loss": 5.5901,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.8,
|
| 18 |
+
"learning_rate": 1.9489855072463772e-05,
|
| 19 |
+
"loss": 3.3533,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 1.2,
|
| 24 |
+
"learning_rate": 1.916779388083736e-05,
|
| 25 |
+
"loss": 2.693,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 1.6,
|
| 30 |
+
"learning_rate": 1.8847665056360712e-05,
|
| 31 |
+
"loss": 2.1327,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 2.0,
|
| 36 |
+
"learning_rate": 1.8525603864734302e-05,
|
| 37 |
+
"loss": 1.8894,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 2.4,
|
| 42 |
+
"learning_rate": 1.8204186795491144e-05,
|
| 43 |
+
"loss": 1.7199,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 2.8,
|
| 48 |
+
"learning_rate": 1.788276972624799e-05,
|
| 49 |
+
"loss": 1.6236,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 3.2,
|
| 54 |
+
"learning_rate": 1.7561996779388084e-05,
|
| 55 |
+
"loss": 1.5586,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 3.6,
|
| 60 |
+
"learning_rate": 1.724057971014493e-05,
|
| 61 |
+
"loss": 1.4624,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 4.0,
|
| 66 |
+
"learning_rate": 1.6919162640901772e-05,
|
| 67 |
+
"loss": 1.4674,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 4.4,
|
| 72 |
+
"learning_rate": 1.6636392914653787e-05,
|
| 73 |
+
"loss": 1.452,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 4.8,
|
| 78 |
+
"learning_rate": 1.6371014492753626e-05,
|
| 79 |
+
"loss": 1.4946,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 5.2,
|
| 84 |
+
"learning_rate": 1.6048953301127216e-05,
|
| 85 |
+
"loss": 0.0,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 5.6,
|
| 90 |
+
"learning_rate": 1.5726892109500806e-05,
|
| 91 |
+
"loss": 0.0,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 6.0,
|
| 96 |
+
"learning_rate": 1.54048309178744e-05,
|
| 97 |
+
"loss": 0.0,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 6.4,
|
| 102 |
+
"learning_rate": 1.5082769726247988e-05,
|
| 103 |
+
"loss": 0.0,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 6.8,
|
| 108 |
+
"learning_rate": 1.476070853462158e-05,
|
| 109 |
+
"loss": 0.0,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 7.2,
|
| 114 |
+
"learning_rate": 1.4438647342995172e-05,
|
| 115 |
+
"loss": 0.0,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 7.6,
|
| 120 |
+
"learning_rate": 1.411658615136876e-05,
|
| 121 |
+
"loss": 0.0,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 8.0,
|
| 126 |
+
"learning_rate": 1.3794524959742352e-05,
|
| 127 |
+
"loss": 0.0,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 8.4,
|
| 132 |
+
"learning_rate": 1.3472463768115942e-05,
|
| 133 |
+
"loss": 0.0,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 8.8,
|
| 138 |
+
"learning_rate": 1.3150402576489534e-05,
|
| 139 |
+
"loss": 0.0,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 9.2,
|
| 144 |
+
"learning_rate": 1.2828341384863126e-05,
|
| 145 |
+
"loss": 0.0,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 9.6,
|
| 150 |
+
"learning_rate": 1.2506280193236716e-05,
|
| 151 |
+
"loss": 0.0,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 10.0,
|
| 156 |
+
"learning_rate": 1.2184219001610308e-05,
|
| 157 |
+
"loss": 0.0,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 10.4,
|
| 162 |
+
"learning_rate": 1.18621578099839e-05,
|
| 163 |
+
"loss": 0.0,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 10.8,
|
| 168 |
+
"learning_rate": 1.1540096618357488e-05,
|
| 169 |
+
"loss": 0.0,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 11.2,
|
| 174 |
+
"learning_rate": 1.121803542673108e-05,
|
| 175 |
+
"loss": 0.0,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 11.6,
|
| 180 |
+
"learning_rate": 1.0895974235104671e-05,
|
| 181 |
+
"loss": 0.0,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 12.0,
|
| 186 |
+
"learning_rate": 1.0573913043478262e-05,
|
| 187 |
+
"loss": 0.0,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 12.4,
|
| 192 |
+
"learning_rate": 1.0251851851851853e-05,
|
| 193 |
+
"loss": 0.0,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 12.8,
|
| 198 |
+
"learning_rate": 9.929790660225444e-06,
|
| 199 |
+
"loss": 0.0,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 13.2,
|
| 204 |
+
"learning_rate": 9.607729468599034e-06,
|
| 205 |
+
"loss": 0.0,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 13.6,
|
| 210 |
+
"learning_rate": 9.285668276972625e-06,
|
| 211 |
+
"loss": 0.0,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 14.0,
|
| 216 |
+
"learning_rate": 8.963607085346217e-06,
|
| 217 |
+
"loss": 0.0,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 14.4,
|
| 222 |
+
"learning_rate": 8.641545893719807e-06,
|
| 223 |
+
"loss": 0.0,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 14.8,
|
| 228 |
+
"learning_rate": 8.3194847020934e-06,
|
| 229 |
+
"loss": 0.0,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 15.2,
|
| 234 |
+
"learning_rate": 7.99742351046699e-06,
|
| 235 |
+
"loss": 0.0,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 15.6,
|
| 240 |
+
"learning_rate": 7.67536231884058e-06,
|
| 241 |
+
"loss": 0.0,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 16.0,
|
| 246 |
+
"learning_rate": 7.3533011272141705e-06,
|
| 247 |
+
"loss": 0.0,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 16.4,
|
| 252 |
+
"learning_rate": 7.031239935587762e-06,
|
| 253 |
+
"loss": 0.0,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 16.8,
|
| 258 |
+
"learning_rate": 6.709178743961353e-06,
|
| 259 |
+
"loss": 0.0,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 17.2,
|
| 264 |
+
"learning_rate": 6.3871175523349435e-06,
|
| 265 |
+
"loss": 0.0,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 17.6,
|
| 270 |
+
"learning_rate": 6.065056360708535e-06,
|
| 271 |
+
"loss": 0.0,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 18.0,
|
| 276 |
+
"learning_rate": 5.742995169082126e-06,
|
| 277 |
+
"loss": 0.0,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 18.4,
|
| 282 |
+
"learning_rate": 5.420933977455716e-06,
|
| 283 |
+
"loss": 0.0,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 18.8,
|
| 288 |
+
"learning_rate": 5.098872785829307e-06,
|
| 289 |
+
"loss": 0.0,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 19.2,
|
| 294 |
+
"learning_rate": 4.776811594202899e-06,
|
| 295 |
+
"loss": 0.0,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 19.6,
|
| 300 |
+
"learning_rate": 4.45475040257649e-06,
|
| 301 |
+
"loss": 0.0,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 20.0,
|
| 306 |
+
"learning_rate": 4.132689210950081e-06,
|
| 307 |
+
"loss": 0.0,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 20.4,
|
| 312 |
+
"learning_rate": 3.8106280193236717e-06,
|
| 313 |
+
"loss": 0.0,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 20.8,
|
| 318 |
+
"learning_rate": 3.4885668276972627e-06,
|
| 319 |
+
"loss": 0.0,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 21.2,
|
| 324 |
+
"learning_rate": 3.1665056360708537e-06,
|
| 325 |
+
"loss": 0.0,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 21.6,
|
| 330 |
+
"learning_rate": 2.8444444444444446e-06,
|
| 331 |
+
"loss": 0.0,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 22.0,
|
| 336 |
+
"learning_rate": 2.522383252818036e-06,
|
| 337 |
+
"loss": 0.0,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 22.4,
|
| 342 |
+
"learning_rate": 2.2003220611916266e-06,
|
| 343 |
+
"loss": 0.0,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 22.8,
|
| 348 |
+
"learning_rate": 1.8782608695652174e-06,
|
| 349 |
+
"loss": 0.0,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 23.2,
|
| 354 |
+
"learning_rate": 1.5561996779388086e-06,
|
| 355 |
+
"loss": 0.0,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 23.6,
|
| 360 |
+
"learning_rate": 1.2341384863123995e-06,
|
| 361 |
+
"loss": 0.0,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 24.0,
|
| 366 |
+
"learning_rate": 9.120772946859904e-07,
|
| 367 |
+
"loss": 0.0,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 24.4,
|
| 372 |
+
"learning_rate": 5.900161030595814e-07,
|
| 373 |
+
"loss": 0.0,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 24.8,
|
| 378 |
+
"learning_rate": 2.679549114331723e-07,
|
| 379 |
+
"loss": 0.0,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 25.0,
|
| 384 |
+
"step": 31250,
|
| 385 |
+
"total_flos": 2.3280874488575558e+19,
|
| 386 |
+
"train_loss": 0.42298770703125,
|
| 387 |
+
"train_runtime": 13299.417,
|
| 388 |
+
"train_samples_per_second": 18.798,
|
| 389 |
+
"train_steps_per_second": 2.35
|
| 390 |
+
}
|
| 391 |
+
],
|
| 392 |
+
"max_steps": 31250,
|
| 393 |
+
"num_train_epochs": 25,
|
| 394 |
+
"total_flos": 2.3280874488575558e+19,
|
| 395 |
+
"trial_name": null,
|
| 396 |
+
"trial_params": null
|
| 397 |
+
}
|