Upload sld_27.0/panel_counter_v24_rsft/trainer_state.json with huggingface_hub
Browse files
sld_27.0/panel_counter_v24_rsft/trainer_state.json
ADDED
|
@@ -0,0 +1,329 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 80,
|
| 3 |
+
"best_metric": 0.04685089,
|
| 4 |
+
"best_model_checkpoint": "/workspace/training/checkpoints/panel_counter_v24_rsft/v0-20260717-090211/checkpoint-80",
|
| 5 |
+
"epoch": 5.188679245283019,
|
| 6 |
+
"eval_steps": 20,
|
| 7 |
+
"global_step": 140,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.03773584905660377,
|
| 14 |
+
"grad_norm": 1.350296974182129,
|
| 15 |
+
"learning_rate": 5.555555555555556e-06,
|
| 16 |
+
"loss": 0.16312940418720245,
|
| 17 |
+
"step": 1,
|
| 18 |
+
"token_acc": 0.9600725952813067
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"epoch": 0.18867924528301888,
|
| 22 |
+
"grad_norm": 0.6247318387031555,
|
| 23 |
+
"learning_rate": 2.777777777777778e-05,
|
| 24 |
+
"loss": 0.12450525909662247,
|
| 25 |
+
"step": 5,
|
| 26 |
+
"token_acc": 0.9678045274883219
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.37735849056603776,
|
| 30 |
+
"grad_norm": 0.29713693261146545,
|
| 31 |
+
"learning_rate": 4.999472998758978e-05,
|
| 32 |
+
"loss": 0.10083991289138794,
|
| 33 |
+
"step": 10,
|
| 34 |
+
"token_acc": 0.9723655813080604
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 0.5660377358490566,
|
| 38 |
+
"grad_norm": 0.4494168162345886,
|
| 39 |
+
"learning_rate": 4.98105127417984e-05,
|
| 40 |
+
"loss": 0.12630804777145385,
|
| 41 |
+
"step": 15,
|
| 42 |
+
"token_acc": 0.9636058782917447
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 0.7547169811320755,
|
| 46 |
+
"grad_norm": 0.32653141021728516,
|
| 47 |
+
"learning_rate": 4.9365012511037514e-05,
|
| 48 |
+
"loss": 0.11727403402328491,
|
| 49 |
+
"step": 20,
|
| 50 |
+
"token_acc": 0.9663228298243135
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.7547169811320755,
|
| 54 |
+
"eval_loss": 0.05273180827498436,
|
| 55 |
+
"eval_runtime": 36.7642,
|
| 56 |
+
"eval_samples_per_second": 1.387,
|
| 57 |
+
"eval_steps_per_second": 1.387,
|
| 58 |
+
"eval_token_acc": 0.9799651567944251,
|
| 59 |
+
"step": 20
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.9433962264150944,
|
| 63 |
+
"grad_norm": 0.2419557422399521,
|
| 64 |
+
"learning_rate": 4.8662920920639866e-05,
|
| 65 |
+
"loss": 0.06185540556907654,
|
| 66 |
+
"step": 25,
|
| 67 |
+
"token_acc": 0.9806084047668827
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 1.1132075471698113,
|
| 71 |
+
"grad_norm": 0.3515181243419647,
|
| 72 |
+
"learning_rate": 4.7711631795488096e-05,
|
| 73 |
+
"loss": 0.05765472054481506,
|
| 74 |
+
"step": 30,
|
| 75 |
+
"token_acc": 0.9799086757990868
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 1.3018867924528301,
|
| 79 |
+
"grad_norm": 0.22770383954048157,
|
| 80 |
+
"learning_rate": 4.6521163294609196e-05,
|
| 81 |
+
"loss": 0.06323248147964478,
|
| 82 |
+
"step": 35,
|
| 83 |
+
"token_acc": 0.9803469775006777
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 1.490566037735849,
|
| 87 |
+
"grad_norm": 0.39493387937545776,
|
| 88 |
+
"learning_rate": 4.510405240853854e-05,
|
| 89 |
+
"loss": 0.10404094457626342,
|
| 90 |
+
"step": 40,
|
| 91 |
+
"token_acc": 0.9700169339575655
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.490566037735849,
|
| 95 |
+
"eval_loss": 0.048691798001527786,
|
| 96 |
+
"eval_runtime": 36.4442,
|
| 97 |
+
"eval_samples_per_second": 1.399,
|
| 98 |
+
"eval_steps_per_second": 1.399,
|
| 99 |
+
"eval_token_acc": 0.9817508710801394,
|
| 100 |
+
"step": 40
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 1.6792452830188678,
|
| 104 |
+
"grad_norm": 0.3438335955142975,
|
| 105 |
+
"learning_rate": 4.347522293051648e-05,
|
| 106 |
+
"loss": 0.053250497579574584,
|
| 107 |
+
"step": 45,
|
| 108 |
+
"token_acc": 0.9824094084535357
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 1.8679245283018868,
|
| 112 |
+
"grad_norm": 0.24651823937892914,
|
| 113 |
+
"learning_rate": 4.1651828291931264e-05,
|
| 114 |
+
"loss": 0.036376011371612546,
|
| 115 |
+
"step": 50,
|
| 116 |
+
"token_acc": 0.9876480455446551
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 2.0377358490566038,
|
| 120 |
+
"grad_norm": 0.20605504512786865,
|
| 121 |
+
"learning_rate": 3.965307091713037e-05,
|
| 122 |
+
"loss": 0.02665190100669861,
|
| 123 |
+
"step": 55,
|
| 124 |
+
"token_acc": 0.9911268855368234
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 2.2264150943396226,
|
| 128 |
+
"grad_norm": 0.2606625556945801,
|
| 129 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 130 |
+
"loss": 0.02710295617580414,
|
| 131 |
+
"step": 60,
|
| 132 |
+
"token_acc": 0.9903079206084214
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 2.2264150943396226,
|
| 136 |
+
"eval_loss": 0.047727711498737335,
|
| 137 |
+
"eval_runtime": 36.6729,
|
| 138 |
+
"eval_samples_per_second": 1.391,
|
| 139 |
+
"eval_steps_per_second": 1.391,
|
| 140 |
+
"eval_token_acc": 0.9814895470383276,
|
| 141 |
+
"step": 60
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 2.4150943396226414,
|
| 145 |
+
"grad_norm": 0.25210270285606384,
|
| 146 |
+
"learning_rate": 3.5215289831955786e-05,
|
| 147 |
+
"loss": 0.029737859964370728,
|
| 148 |
+
"step": 65,
|
| 149 |
+
"token_acc": 0.9909777484920675
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 2.6037735849056602,
|
| 153 |
+
"grad_norm": 0.4128791391849518,
|
| 154 |
+
"learning_rate": 3.282300101580386e-05,
|
| 155 |
+
"loss": 0.03639853894710541,
|
| 156 |
+
"step": 70,
|
| 157 |
+
"token_acc": 0.9874800478850758
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 2.7924528301886795,
|
| 161 |
+
"grad_norm": 0.18674048781394958,
|
| 162 |
+
"learning_rate": 3.0348327080162435e-05,
|
| 163 |
+
"loss": 0.031123530864715577,
|
| 164 |
+
"step": 75,
|
| 165 |
+
"token_acc": 0.9894353177328192
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.981132075471698,
|
| 169 |
+
"grad_norm": 0.3658820688724518,
|
| 170 |
+
"learning_rate": 2.781732916288303e-05,
|
| 171 |
+
"loss": 0.04066756665706635,
|
| 172 |
+
"step": 80,
|
| 173 |
+
"token_acc": 0.9867752803338865
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 2.981132075471698,
|
| 177 |
+
"eval_loss": 0.046850889921188354,
|
| 178 |
+
"eval_runtime": 36.7683,
|
| 179 |
+
"eval_samples_per_second": 1.387,
|
| 180 |
+
"eval_steps_per_second": 1.387,
|
| 181 |
+
"eval_token_acc": 0.982273519163763,
|
| 182 |
+
"step": 80
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.150943396226415,
|
| 186 |
+
"grad_norm": 0.13613182306289673,
|
| 187 |
+
"learning_rate": 2.5256661557557247e-05,
|
| 188 |
+
"loss": 0.05700427293777466,
|
| 189 |
+
"step": 85,
|
| 190 |
+
"token_acc": 0.9842603370615922
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.339622641509434,
|
| 194 |
+
"grad_norm": 0.31659549474716187,
|
| 195 |
+
"learning_rate": 2.2693291013417453e-05,
|
| 196 |
+
"loss": 0.03171400725841522,
|
| 197 |
+
"step": 90,
|
| 198 |
+
"token_acc": 0.9895714220372318
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 3.5283018867924527,
|
| 202 |
+
"grad_norm": 0.4015447497367859,
|
| 203 |
+
"learning_rate": 2.015421274472325e-05,
|
| 204 |
+
"loss": 0.012700113654136657,
|
| 205 |
+
"step": 95,
|
| 206 |
+
"token_acc": 0.9953211867672153
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.7169811320754715,
|
| 210 |
+
"grad_norm": 0.23558788001537323,
|
| 211 |
+
"learning_rate": 1.7666166140378852e-05,
|
| 212 |
+
"loss": 0.025953733921051027,
|
| 213 |
+
"step": 100,
|
| 214 |
+
"token_acc": 0.9913788807494892
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 3.7169811320754715,
|
| 218 |
+
"eval_loss": 0.05166962742805481,
|
| 219 |
+
"eval_runtime": 36.6478,
|
| 220 |
+
"eval_samples_per_second": 1.392,
|
| 221 |
+
"eval_steps_per_second": 1.392,
|
| 222 |
+
"eval_token_acc": 0.982273519163763,
|
| 223 |
+
"step": 100
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 3.9056603773584904,
|
| 227 |
+
"grad_norm": 0.284417062997818,
|
| 228 |
+
"learning_rate": 1.5255353167683017e-05,
|
| 229 |
+
"loss": 0.01755093187093735,
|
| 230 |
+
"step": 105,
|
| 231 |
+
"token_acc": 0.9936604723190089
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 4.0754716981132075,
|
| 235 |
+
"grad_norm": 0.21325848996639252,
|
| 236 |
+
"learning_rate": 1.2947162435741278e-05,
|
| 237 |
+
"loss": 0.014751245081424714,
|
| 238 |
+
"step": 110,
|
| 239 |
+
"token_acc": 0.9949184721911994
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.264150943396227,
|
| 243 |
+
"grad_norm": 0.1644717901945114,
|
| 244 |
+
"learning_rate": 1.0765901824467167e-05,
|
| 245 |
+
"loss": 0.01028374806046486,
|
| 246 |
+
"step": 115,
|
| 247 |
+
"token_acc": 0.9962836791576339
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 4.452830188679245,
|
| 251 |
+
"grad_norm": 0.17199504375457764,
|
| 252 |
+
"learning_rate": 8.734542494893955e-06,
|
| 253 |
+
"loss": 0.010910643637180329,
|
| 254 |
+
"step": 120,
|
| 255 |
+
"token_acc": 0.9961574902621328
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 4.452830188679245,
|
| 259 |
+
"eval_loss": 0.050852350890636444,
|
| 260 |
+
"eval_runtime": 36.5898,
|
| 261 |
+
"eval_samples_per_second": 1.394,
|
| 262 |
+
"eval_steps_per_second": 1.394,
|
| 263 |
+
"eval_token_acc": 0.9827526132404181,
|
| 264 |
+
"step": 120
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.6415094339622645,
|
| 268 |
+
"grad_norm": 0.4569353759288788,
|
| 269 |
+
"learning_rate": 6.8744769766601854e-06,
|
| 270 |
+
"loss": 0.021015277504920958,
|
| 271 |
+
"step": 125,
|
| 272 |
+
"token_acc": 0.992462311557789
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 4.830188679245283,
|
| 276 |
+
"grad_norm": 0.34078502655029297,
|
| 277 |
+
"learning_rate": 5.205293880283552e-06,
|
| 278 |
+
"loss": 0.013351918756961822,
|
| 279 |
+
"step": 130,
|
| 280 |
+
"token_acc": 0.996004160052548
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 5.0,
|
| 284 |
+
"grad_norm": 0.23289847373962402,
|
| 285 |
+
"learning_rate": 3.7445716067596503e-06,
|
| 286 |
+
"loss": 0.009969624876976012,
|
| 287 |
+
"step": 135,
|
| 288 |
+
"token_acc": 0.9963277610304193
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 5.188679245283019,
|
| 292 |
+
"grad_norm": 0.17062047123908997,
|
| 293 |
+
"learning_rate": 2.507693226958871e-06,
|
| 294 |
+
"loss": 0.005502679198980331,
|
| 295 |
+
"step": 140,
|
| 296 |
+
"token_acc": 0.9982894103811407
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 5.188679245283019,
|
| 300 |
+
"eval_loss": 0.051256969571113586,
|
| 301 |
+
"eval_runtime": 37.2526,
|
| 302 |
+
"eval_samples_per_second": 1.369,
|
| 303 |
+
"eval_steps_per_second": 1.369,
|
| 304 |
+
"eval_token_acc": 0.9825348432055749,
|
| 305 |
+
"step": 140
|
| 306 |
+
}
|
| 307 |
+
],
|
| 308 |
+
"logging_steps": 5,
|
| 309 |
+
"max_steps": 162,
|
| 310 |
+
"num_input_tokens_seen": 0,
|
| 311 |
+
"num_train_epochs": 6,
|
| 312 |
+
"save_steps": 20,
|
| 313 |
+
"stateful_callbacks": {
|
| 314 |
+
"TrainerControl": {
|
| 315 |
+
"args": {
|
| 316 |
+
"should_epoch_stop": false,
|
| 317 |
+
"should_evaluate": false,
|
| 318 |
+
"should_log": false,
|
| 319 |
+
"should_save": true,
|
| 320 |
+
"should_training_stop": false
|
| 321 |
+
},
|
| 322 |
+
"attributes": {}
|
| 323 |
+
}
|
| 324 |
+
},
|
| 325 |
+
"total_flos": 1.1153563097586624e+17,
|
| 326 |
+
"train_batch_size": 1,
|
| 327 |
+
"trial_name": null,
|
| 328 |
+
"trial_params": null
|
| 329 |
+
}
|