Training in progress, step 191, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 645975704
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e50fb513373efc8d7be1aa405acd51b656eda59fe3bb2a6828a040bc40f028fa
|
| 3 |
size 645975704
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 328468404
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:02eed8f1d6b9be5f98130b2d24376288bd62c4b4ae15100b5bd79cc9a256674f
|
| 3 |
size 328468404
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2c0b3eb76ec3598af64c71a4844be44e294f3ae7cc9e4a726a8760027806eef5
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d8781b403f6e79d31199d91dc5482eb96689fd318472e93e4c83efb756ecf166
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.8613990545272827,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-150",
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 50,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1089,6 +1089,293 @@
|
|
| 1089 |
"eval_samples_per_second": 6.976,
|
| 1090 |
"eval_steps_per_second": 1.76,
|
| 1091 |
"step": 150
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1092 |
}
|
| 1093 |
],
|
| 1094 |
"logging_steps": 1,
|
|
@@ -1112,12 +1399,12 @@
|
|
| 1112 |
"should_evaluate": false,
|
| 1113 |
"should_log": false,
|
| 1114 |
"should_save": true,
|
| 1115 |
-
"should_training_stop":
|
| 1116 |
},
|
| 1117 |
"attributes": {}
|
| 1118 |
}
|
| 1119 |
},
|
| 1120 |
-
"total_flos":
|
| 1121 |
"train_batch_size": 8,
|
| 1122 |
"trial_name": null,
|
| 1123 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.8613990545272827,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-150",
|
| 4 |
+
"epoch": 3.0078740157480315,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 191,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1089 |
"eval_samples_per_second": 6.976,
|
| 1090 |
"eval_steps_per_second": 1.76,
|
| 1091 |
"step": 150
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 2.377952755905512,
|
| 1095 |
+
"grad_norm": 0.23553703725337982,
|
| 1096 |
+
"learning_rate": 1.1574098862709992e-05,
|
| 1097 |
+
"loss": 1.7086,
|
| 1098 |
+
"step": 151
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 2.393700787401575,
|
| 1102 |
+
"grad_norm": 0.2348308116197586,
|
| 1103 |
+
"learning_rate": 1.1024644157551206e-05,
|
| 1104 |
+
"loss": 1.7374,
|
| 1105 |
+
"step": 152
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 2.409448818897638,
|
| 1109 |
+
"grad_norm": 0.24118292331695557,
|
| 1110 |
+
"learning_rate": 1.0486930903705094e-05,
|
| 1111 |
+
"loss": 1.7819,
|
| 1112 |
+
"step": 153
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 2.425196850393701,
|
| 1116 |
+
"grad_norm": 0.22775617241859436,
|
| 1117 |
+
"learning_rate": 9.961121089018932e-06,
|
| 1118 |
+
"loss": 1.7897,
|
| 1119 |
+
"step": 154
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 2.440944881889764,
|
| 1123 |
+
"grad_norm": 0.22295954823493958,
|
| 1124 |
+
"learning_rate": 9.4473731153907e-06,
|
| 1125 |
+
"loss": 1.7652,
|
| 1126 |
+
"step": 155
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 2.456692913385827,
|
| 1130 |
+
"grad_norm": 0.23308737576007843,
|
| 1131 |
+
"learning_rate": 8.945841751049915e-06,
|
| 1132 |
+
"loss": 1.7224,
|
| 1133 |
+
"step": 156
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 2.47244094488189,
|
| 1137 |
+
"grad_norm": 0.25666847825050354,
|
| 1138 |
+
"learning_rate": 8.456678083933289e-06,
|
| 1139 |
+
"loss": 1.7704,
|
| 1140 |
+
"step": 157
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 2.4881889763779528,
|
| 1144 |
+
"grad_norm": 0.24374538660049438,
|
| 1145 |
+
"learning_rate": 7.980029476168944e-06,
|
| 1146 |
+
"loss": 1.958,
|
| 1147 |
+
"step": 158
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 2.5039370078740157,
|
| 1151 |
+
"grad_norm": 0.2336760014295578,
|
| 1152 |
+
"learning_rate": 7.5160395196831046e-06,
|
| 1153 |
+
"loss": 1.6695,
|
| 1154 |
+
"step": 159
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 2.5196850393700787,
|
| 1158 |
+
"grad_norm": 0.23040024936199188,
|
| 1159 |
+
"learning_rate": 7.064847992942614e-06,
|
| 1160 |
+
"loss": 1.748,
|
| 1161 |
+
"step": 160
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 2.5354330708661417,
|
| 1165 |
+
"grad_norm": 0.2296517789363861,
|
| 1166 |
+
"learning_rate": 6.626590818846162e-06,
|
| 1167 |
+
"loss": 1.7122,
|
| 1168 |
+
"step": 161
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 2.5511811023622046,
|
| 1172 |
+
"grad_norm": 0.2303856462240219,
|
| 1173 |
+
"learning_rate": 6.201400023777104e-06,
|
| 1174 |
+
"loss": 1.7648,
|
| 1175 |
+
"step": 162
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 2.5669291338582676,
|
| 1179 |
+
"grad_norm": 0.2246444672346115,
|
| 1180 |
+
"learning_rate": 5.7894036978301035e-06,
|
| 1181 |
+
"loss": 1.6923,
|
| 1182 |
+
"step": 163
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 2.5826771653543306,
|
| 1186 |
+
"grad_norm": 0.23633837699890137,
|
| 1187 |
+
"learning_rate": 5.39072595622353e-06,
|
| 1188 |
+
"loss": 1.7337,
|
| 1189 |
+
"step": 164
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 2.5984251968503935,
|
| 1193 |
+
"grad_norm": 0.24737262725830078,
|
| 1194 |
+
"learning_rate": 5.005486901909428e-06,
|
| 1195 |
+
"loss": 1.7537,
|
| 1196 |
+
"step": 165
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 2.6141732283464565,
|
| 1200 |
+
"grad_norm": 0.24044926464557648,
|
| 1201 |
+
"learning_rate": 4.6338025893920166e-06,
|
| 1202 |
+
"loss": 1.8285,
|
| 1203 |
+
"step": 166
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 2.6299212598425195,
|
| 1207 |
+
"grad_norm": 0.2352256029844284,
|
| 1208 |
+
"learning_rate": 4.275784989765985e-06,
|
| 1209 |
+
"loss": 1.7291,
|
| 1210 |
+
"step": 167
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 2.6456692913385824,
|
| 1214 |
+
"grad_norm": 0.23979999125003815,
|
| 1215 |
+
"learning_rate": 3.93154195698478e-06,
|
| 1216 |
+
"loss": 1.7735,
|
| 1217 |
+
"step": 168
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 2.661417322834646,
|
| 1221 |
+
"grad_norm": 0.23789139091968536,
|
| 1222 |
+
"learning_rate": 3.601177195369304e-06,
|
| 1223 |
+
"loss": 1.8119,
|
| 1224 |
+
"step": 169
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 2.677165354330709,
|
| 1228 |
+
"grad_norm": 0.2280394583940506,
|
| 1229 |
+
"learning_rate": 3.2847902283666022e-06,
|
| 1230 |
+
"loss": 1.7269,
|
| 1231 |
+
"step": 170
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 2.6929133858267718,
|
| 1235 |
+
"grad_norm": 0.22925442457199097,
|
| 1236 |
+
"learning_rate": 2.9824763685681766e-06,
|
| 1237 |
+
"loss": 1.7427,
|
| 1238 |
+
"step": 171
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 2.7086614173228347,
|
| 1242 |
+
"grad_norm": 0.25694581866264343,
|
| 1243 |
+
"learning_rate": 2.694326688996662e-06,
|
| 1244 |
+
"loss": 1.6969,
|
| 1245 |
+
"step": 172
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 2.7244094488188977,
|
| 1249 |
+
"grad_norm": 0.24270634353160858,
|
| 1250 |
+
"learning_rate": 2.4204279956698995e-06,
|
| 1251 |
+
"loss": 1.8398,
|
| 1252 |
+
"step": 173
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 2.7401574803149606,
|
| 1256 |
+
"grad_norm": 0.23506343364715576,
|
| 1257 |
+
"learning_rate": 2.1608628014502365e-06,
|
| 1258 |
+
"loss": 1.6692,
|
| 1259 |
+
"step": 174
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 2.7559055118110236,
|
| 1263 |
+
"grad_norm": 0.2271362543106079,
|
| 1264 |
+
"learning_rate": 1.915709301187335e-06,
|
| 1265 |
+
"loss": 1.6801,
|
| 1266 |
+
"step": 175
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 2.7716535433070866,
|
| 1270 |
+
"grad_norm": 0.22768786549568176,
|
| 1271 |
+
"learning_rate": 1.6850413481616868e-06,
|
| 1272 |
+
"loss": 1.5849,
|
| 1273 |
+
"step": 176
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 2.7874015748031495,
|
| 1277 |
+
"grad_norm": 0.24064046144485474,
|
| 1278 |
+
"learning_rate": 1.4689284318360918e-06,
|
| 1279 |
+
"loss": 1.7468,
|
| 1280 |
+
"step": 177
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 2.8031496062992125,
|
| 1284 |
+
"grad_norm": 0.2282429337501526,
|
| 1285 |
+
"learning_rate": 1.2674356569217282e-06,
|
| 1286 |
+
"loss": 1.7783,
|
| 1287 |
+
"step": 178
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 2.8188976377952755,
|
| 1291 |
+
"grad_norm": 0.2347092181444168,
|
| 1292 |
+
"learning_rate": 1.080623723765134e-06,
|
| 1293 |
+
"loss": 1.7443,
|
| 1294 |
+
"step": 179
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 2.8346456692913384,
|
| 1298 |
+
"grad_norm": 0.23791402578353882,
|
| 1299 |
+
"learning_rate": 9.085489100620737e-07,
|
| 1300 |
+
"loss": 1.8306,
|
| 1301 |
+
"step": 180
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 2.850393700787402,
|
| 1305 |
+
"grad_norm": 0.23669809103012085,
|
| 1306 |
+
"learning_rate": 7.512630539036502e-07,
|
| 1307 |
+
"loss": 1.7913,
|
| 1308 |
+
"step": 181
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 2.866141732283465,
|
| 1312 |
+
"grad_norm": 0.24042251706123352,
|
| 1313 |
+
"learning_rate": 6.088135381599414e-07,
|
| 1314 |
+
"loss": 1.7201,
|
| 1315 |
+
"step": 182
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 2.8818897637795278,
|
| 1319 |
+
"grad_norm": 0.23241114616394043,
|
| 1320 |
+
"learning_rate": 4.812432762057673e-07,
|
| 1321 |
+
"loss": 1.7816,
|
| 1322 |
+
"step": 183
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 2.8976377952755907,
|
| 1326 |
+
"grad_norm": 0.24429729580879211,
|
| 1327 |
+
"learning_rate": 3.685906989928656e-07,
|
| 1328 |
+
"loss": 1.8598,
|
| 1329 |
+
"step": 184
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 2.9133858267716537,
|
| 1333 |
+
"grad_norm": 0.23608095943927765,
|
| 1334 |
+
"learning_rate": 2.7088974347246887e-07,
|
| 1335 |
+
"loss": 1.7586,
|
| 1336 |
+
"step": 185
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 2.9291338582677167,
|
| 1340 |
+
"grad_norm": 0.2519053816795349,
|
| 1341 |
+
"learning_rate": 1.8816984237169376e-07,
|
| 1342 |
+
"loss": 1.7577,
|
| 1343 |
+
"step": 186
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 2.9448818897637796,
|
| 1347 |
+
"grad_norm": 0.27554014325141907,
|
| 1348 |
+
"learning_rate": 1.2045591532681145e-07,
|
| 1349 |
+
"loss": 1.7541,
|
| 1350 |
+
"step": 187
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 2.9606299212598426,
|
| 1354 |
+
"grad_norm": 0.24321898818016052,
|
| 1355 |
+
"learning_rate": 6.776836137615262e-08,
|
| 1356 |
+
"loss": 1.7011,
|
| 1357 |
+
"step": 188
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 2.9763779527559056,
|
| 1361 |
+
"grad_norm": 0.24500428140163422,
|
| 1362 |
+
"learning_rate": 3.0123052814812206e-08,
|
| 1363 |
+
"loss": 1.7679,
|
| 1364 |
+
"step": 189
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 2.9921259842519685,
|
| 1368 |
+
"grad_norm": 0.2374401092529297,
|
| 1369 |
+
"learning_rate": 7.53133041307974e-09,
|
| 1370 |
+
"loss": 1.7637,
|
| 1371 |
+
"step": 190
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 3.0078740157480315,
|
| 1375 |
+
"grad_norm": 0.47424909472465515,
|
| 1376 |
+
"learning_rate": 0.0,
|
| 1377 |
+
"loss": 2.8832,
|
| 1378 |
+
"step": 191
|
| 1379 |
}
|
| 1380 |
],
|
| 1381 |
"logging_steps": 1,
|
|
|
|
| 1399 |
"should_evaluate": false,
|
| 1400 |
"should_log": false,
|
| 1401 |
"should_save": true,
|
| 1402 |
+
"should_training_stop": true
|
| 1403 |
},
|
| 1404 |
"attributes": {}
|
| 1405 |
}
|
| 1406 |
},
|
| 1407 |
+
"total_flos": 5.319641044038451e+17,
|
| 1408 |
"train_batch_size": 8,
|
| 1409 |
"trial_name": null,
|
| 1410 |
"trial_params": null
|