Training in progress, step 200, 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 37762064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9ed52769f4242edb5e1011762620b55e1d4e461cbeff5d0026745c516054523f
|
| 3 |
size 37762064
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 19283770
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:32eb0a7674586bf0c183c85fa1ca2791078485a0a36abeea5a6f80e8c16815cf
|
| 3 |
size 19283770
|
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 sha256:61dfad1171b39c1a536a6ca4db849c0f4a035448c94a1e995bca2399c4258c7b
|
| 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:f5efc880080e0854e5765c7a5ea108f46d0ccbf30b054c9bbb5a2162a1c5babf
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 3.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 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 +1089,364 @@
|
|
| 1089 |
"eval_samples_per_second": 156.735,
|
| 1090 |
"eval_steps_per_second": 39.342,
|
| 1091 |
"step": 150
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1092 |
}
|
| 1093 |
],
|
| 1094 |
"logging_steps": 1,
|
|
@@ -1103,7 +1461,7 @@
|
|
| 1103 |
"early_stopping_threshold": 0.0
|
| 1104 |
},
|
| 1105 |
"attributes": {
|
| 1106 |
-
"early_stopping_patience_counter":
|
| 1107 |
}
|
| 1108 |
},
|
| 1109 |
"TrainerControl": {
|
|
@@ -1112,12 +1470,12 @@
|
|
| 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": 4,
|
| 1122 |
"trial_name": null,
|
| 1123 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 3.0275626182556152,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 4 |
+
"epoch": 0.11366865586814436,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 200,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1089 |
"eval_samples_per_second": 156.735,
|
| 1090 |
"eval_steps_per_second": 39.342,
|
| 1091 |
"step": 150
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 0.08581983518044899,
|
| 1095 |
+
"grad_norm": 9544.724609375,
|
| 1096 |
+
"learning_rate": 2.5815263157894736e-05,
|
| 1097 |
+
"loss": 5.8488,
|
| 1098 |
+
"step": 151
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 0.08638817845978972,
|
| 1102 |
+
"grad_norm": 14259.00390625,
|
| 1103 |
+
"learning_rate": 2.528842105263158e-05,
|
| 1104 |
+
"loss": 5.3379,
|
| 1105 |
+
"step": 152
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 0.08695652173913043,
|
| 1109 |
+
"grad_norm": 1097.5711669921875,
|
| 1110 |
+
"learning_rate": 2.4761578947368418e-05,
|
| 1111 |
+
"loss": 6.3404,
|
| 1112 |
+
"step": 153
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.08752486501847116,
|
| 1116 |
+
"grad_norm": 1774.6893310546875,
|
| 1117 |
+
"learning_rate": 2.423473684210526e-05,
|
| 1118 |
+
"loss": 6.5448,
|
| 1119 |
+
"step": 154
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 0.08809320829781188,
|
| 1123 |
+
"grad_norm": 4005.585693359375,
|
| 1124 |
+
"learning_rate": 2.3707894736842103e-05,
|
| 1125 |
+
"loss": 5.8295,
|
| 1126 |
+
"step": 155
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 0.0886615515771526,
|
| 1130 |
+
"grad_norm": 3084.43359375,
|
| 1131 |
+
"learning_rate": 2.3181052631578946e-05,
|
| 1132 |
+
"loss": 4.9181,
|
| 1133 |
+
"step": 156
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 0.08922989485649332,
|
| 1137 |
+
"grad_norm": 3153.96923828125,
|
| 1138 |
+
"learning_rate": 2.265421052631579e-05,
|
| 1139 |
+
"loss": 6.0936,
|
| 1140 |
+
"step": 157
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 0.08979823813583404,
|
| 1144 |
+
"grad_norm": 3639.78125,
|
| 1145 |
+
"learning_rate": 2.212736842105263e-05,
|
| 1146 |
+
"loss": 5.4882,
|
| 1147 |
+
"step": 158
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 0.09036658141517477,
|
| 1151 |
+
"grad_norm": 1438.063232421875,
|
| 1152 |
+
"learning_rate": 2.1600526315789474e-05,
|
| 1153 |
+
"loss": 6.7295,
|
| 1154 |
+
"step": 159
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.09093492469451549,
|
| 1158 |
+
"grad_norm": 4263.53662109375,
|
| 1159 |
+
"learning_rate": 2.1073684210526313e-05,
|
| 1160 |
+
"loss": 5.478,
|
| 1161 |
+
"step": 160
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.0915032679738562,
|
| 1165 |
+
"grad_norm": 1958.2095947265625,
|
| 1166 |
+
"learning_rate": 2.0546842105263155e-05,
|
| 1167 |
+
"loss": 5.7273,
|
| 1168 |
+
"step": 161
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 0.09207161125319693,
|
| 1172 |
+
"grad_norm": 2313.61083984375,
|
| 1173 |
+
"learning_rate": 2.002e-05,
|
| 1174 |
+
"loss": 5.5683,
|
| 1175 |
+
"step": 162
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 0.09263995453253765,
|
| 1179 |
+
"grad_norm": 1184.456298828125,
|
| 1180 |
+
"learning_rate": 1.949315789473684e-05,
|
| 1181 |
+
"loss": 6.1653,
|
| 1182 |
+
"step": 163
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 0.09320829781187838,
|
| 1186 |
+
"grad_norm": 2392.1455078125,
|
| 1187 |
+
"learning_rate": 1.8966315789473683e-05,
|
| 1188 |
+
"loss": 6.6287,
|
| 1189 |
+
"step": 164
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 0.0937766410912191,
|
| 1193 |
+
"grad_norm": 3100.29736328125,
|
| 1194 |
+
"learning_rate": 1.8439473684210522e-05,
|
| 1195 |
+
"loss": 6.3648,
|
| 1196 |
+
"step": 165
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.09434498437055981,
|
| 1200 |
+
"grad_norm": 1175.8050537109375,
|
| 1201 |
+
"learning_rate": 1.791263157894737e-05,
|
| 1202 |
+
"loss": 6.3125,
|
| 1203 |
+
"step": 166
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 0.09491332764990054,
|
| 1207 |
+
"grad_norm": 2080.410888671875,
|
| 1208 |
+
"learning_rate": 1.738578947368421e-05,
|
| 1209 |
+
"loss": 5.5606,
|
| 1210 |
+
"step": 167
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 0.09548167092924126,
|
| 1214 |
+
"grad_norm": 3428.589599609375,
|
| 1215 |
+
"learning_rate": 1.685894736842105e-05,
|
| 1216 |
+
"loss": 6.6513,
|
| 1217 |
+
"step": 168
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 0.09605001420858199,
|
| 1221 |
+
"grad_norm": 3019.490478515625,
|
| 1222 |
+
"learning_rate": 1.6332105263157893e-05,
|
| 1223 |
+
"loss": 4.9957,
|
| 1224 |
+
"step": 169
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 0.0966183574879227,
|
| 1228 |
+
"grad_norm": 1557.4947509765625,
|
| 1229 |
+
"learning_rate": 1.5805263157894735e-05,
|
| 1230 |
+
"loss": 6.3384,
|
| 1231 |
+
"step": 170
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 0.09718670076726342,
|
| 1235 |
+
"grad_norm": 902.9617919921875,
|
| 1236 |
+
"learning_rate": 1.5278421052631578e-05,
|
| 1237 |
+
"loss": 6.0823,
|
| 1238 |
+
"step": 171
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.09775504404660415,
|
| 1242 |
+
"grad_norm": 1549.714599609375,
|
| 1243 |
+
"learning_rate": 1.4751578947368419e-05,
|
| 1244 |
+
"loss": 5.5164,
|
| 1245 |
+
"step": 172
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 0.09832338732594487,
|
| 1249 |
+
"grad_norm": 1734.38330078125,
|
| 1250 |
+
"learning_rate": 1.4224736842105262e-05,
|
| 1251 |
+
"loss": 5.9773,
|
| 1252 |
+
"step": 173
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 0.0988917306052856,
|
| 1256 |
+
"grad_norm": 4998.42333984375,
|
| 1257 |
+
"learning_rate": 1.3697894736842106e-05,
|
| 1258 |
+
"loss": 5.3307,
|
| 1259 |
+
"step": 174
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 0.09946007388462631,
|
| 1263 |
+
"grad_norm": 2184.949462890625,
|
| 1264 |
+
"learning_rate": 1.3171052631578945e-05,
|
| 1265 |
+
"loss": 6.3121,
|
| 1266 |
+
"step": 175
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 0.10002841716396704,
|
| 1270 |
+
"grad_norm": 949.4255981445312,
|
| 1271 |
+
"learning_rate": 1.264421052631579e-05,
|
| 1272 |
+
"loss": 5.8982,
|
| 1273 |
+
"step": 176
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 0.10059676044330776,
|
| 1277 |
+
"grad_norm": 2043.7244873046875,
|
| 1278 |
+
"learning_rate": 1.211736842105263e-05,
|
| 1279 |
+
"loss": 6.0588,
|
| 1280 |
+
"step": 177
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 0.10116510372264848,
|
| 1284 |
+
"grad_norm": 3321.778076171875,
|
| 1285 |
+
"learning_rate": 1.1590526315789473e-05,
|
| 1286 |
+
"loss": 6.2381,
|
| 1287 |
+
"step": 178
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 0.1017334470019892,
|
| 1291 |
+
"grad_norm": 516.5933227539062,
|
| 1292 |
+
"learning_rate": 1.1063684210526316e-05,
|
| 1293 |
+
"loss": 5.9794,
|
| 1294 |
+
"step": 179
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 0.10230179028132992,
|
| 1298 |
+
"grad_norm": 1424.5836181640625,
|
| 1299 |
+
"learning_rate": 1.0536842105263156e-05,
|
| 1300 |
+
"loss": 6.3459,
|
| 1301 |
+
"step": 180
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 0.10287013356067065,
|
| 1305 |
+
"grad_norm": 3069.05712890625,
|
| 1306 |
+
"learning_rate": 1.001e-05,
|
| 1307 |
+
"loss": 6.2964,
|
| 1308 |
+
"step": 181
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 0.10343847684001137,
|
| 1312 |
+
"grad_norm": 615.3517456054688,
|
| 1313 |
+
"learning_rate": 9.483157894736842e-06,
|
| 1314 |
+
"loss": 6.2804,
|
| 1315 |
+
"step": 182
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 0.10400682011935208,
|
| 1319 |
+
"grad_norm": 2094.96240234375,
|
| 1320 |
+
"learning_rate": 8.956315789473684e-06,
|
| 1321 |
+
"loss": 5.9543,
|
| 1322 |
+
"step": 183
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 0.10457516339869281,
|
| 1326 |
+
"grad_norm": 2743.946533203125,
|
| 1327 |
+
"learning_rate": 8.429473684210525e-06,
|
| 1328 |
+
"loss": 5.9415,
|
| 1329 |
+
"step": 184
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 0.10514350667803353,
|
| 1333 |
+
"grad_norm": 3393.304931640625,
|
| 1334 |
+
"learning_rate": 7.902631578947368e-06,
|
| 1335 |
+
"loss": 6.3838,
|
| 1336 |
+
"step": 185
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 0.10571184995737426,
|
| 1340 |
+
"grad_norm": 1315.7086181640625,
|
| 1341 |
+
"learning_rate": 7.3757894736842095e-06,
|
| 1342 |
+
"loss": 5.8147,
|
| 1343 |
+
"step": 186
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 0.10628019323671498,
|
| 1347 |
+
"grad_norm": 1178.4144287109375,
|
| 1348 |
+
"learning_rate": 6.848947368421053e-06,
|
| 1349 |
+
"loss": 5.7014,
|
| 1350 |
+
"step": 187
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 0.10684853651605569,
|
| 1354 |
+
"grad_norm": 3066.94091796875,
|
| 1355 |
+
"learning_rate": 6.322105263157895e-06,
|
| 1356 |
+
"loss": 6.9476,
|
| 1357 |
+
"step": 188
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 0.10741687979539642,
|
| 1361 |
+
"grad_norm": 818.8816528320312,
|
| 1362 |
+
"learning_rate": 5.7952631578947365e-06,
|
| 1363 |
+
"loss": 5.4968,
|
| 1364 |
+
"step": 189
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 0.10798522307473714,
|
| 1368 |
+
"grad_norm": 6053.64111328125,
|
| 1369 |
+
"learning_rate": 5.268421052631578e-06,
|
| 1370 |
+
"loss": 6.8307,
|
| 1371 |
+
"step": 190
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 0.10855356635407787,
|
| 1375 |
+
"grad_norm": 1458.8125,
|
| 1376 |
+
"learning_rate": 4.741578947368421e-06,
|
| 1377 |
+
"loss": 5.5373,
|
| 1378 |
+
"step": 191
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 0.10912190963341858,
|
| 1382 |
+
"grad_norm": 3176.13330078125,
|
| 1383 |
+
"learning_rate": 4.2147368421052626e-06,
|
| 1384 |
+
"loss": 6.041,
|
| 1385 |
+
"step": 192
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 0.1096902529127593,
|
| 1389 |
+
"grad_norm": 1368.4093017578125,
|
| 1390 |
+
"learning_rate": 3.6878947368421047e-06,
|
| 1391 |
+
"loss": 6.1969,
|
| 1392 |
+
"step": 193
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 0.11025859619210003,
|
| 1396 |
+
"grad_norm": 2060.69921875,
|
| 1397 |
+
"learning_rate": 3.1610526315789474e-06,
|
| 1398 |
+
"loss": 6.4491,
|
| 1399 |
+
"step": 194
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 0.11082693947144075,
|
| 1403 |
+
"grad_norm": 7935.82177734375,
|
| 1404 |
+
"learning_rate": 2.634210526315789e-06,
|
| 1405 |
+
"loss": 7.1167,
|
| 1406 |
+
"step": 195
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 0.11139528275078148,
|
| 1410 |
+
"grad_norm": 2467.206298828125,
|
| 1411 |
+
"learning_rate": 2.1073684210526313e-06,
|
| 1412 |
+
"loss": 6.9577,
|
| 1413 |
+
"step": 196
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 0.11196362603012219,
|
| 1417 |
+
"grad_norm": 5464.1357421875,
|
| 1418 |
+
"learning_rate": 1.5805263157894737e-06,
|
| 1419 |
+
"loss": 6.1997,
|
| 1420 |
+
"step": 197
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 0.11253196930946291,
|
| 1424 |
+
"grad_norm": 2190.900146484375,
|
| 1425 |
+
"learning_rate": 1.0536842105263156e-06,
|
| 1426 |
+
"loss": 7.1513,
|
| 1427 |
+
"step": 198
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 0.11310031258880364,
|
| 1431 |
+
"grad_norm": 2982.100830078125,
|
| 1432 |
+
"learning_rate": 5.268421052631578e-07,
|
| 1433 |
+
"loss": 6.7664,
|
| 1434 |
+
"step": 199
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 0.11366865586814436,
|
| 1438 |
+
"grad_norm": 2327.494140625,
|
| 1439 |
+
"learning_rate": 0.0,
|
| 1440 |
+
"loss": 7.4672,
|
| 1441 |
+
"step": 200
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 0.11366865586814436,
|
| 1445 |
+
"eval_loss": 3.0275626182556152,
|
| 1446 |
+
"eval_runtime": 4.7155,
|
| 1447 |
+
"eval_samples_per_second": 157.141,
|
| 1448 |
+
"eval_steps_per_second": 39.444,
|
| 1449 |
+
"step": 200
|
| 1450 |
}
|
| 1451 |
],
|
| 1452 |
"logging_steps": 1,
|
|
|
|
| 1461 |
"early_stopping_threshold": 0.0
|
| 1462 |
},
|
| 1463 |
"attributes": {
|
| 1464 |
+
"early_stopping_patience_counter": 0
|
| 1465 |
}
|
| 1466 |
},
|
| 1467 |
"TrainerControl": {
|
|
|
|
| 1470 |
"should_evaluate": false,
|
| 1471 |
"should_log": false,
|
| 1472 |
"should_save": true,
|
| 1473 |
+
"should_training_stop": true
|
| 1474 |
},
|
| 1475 |
"attributes": {}
|
| 1476 |
}
|
| 1477 |
},
|
| 1478 |
+
"total_flos": 1387209575890944.0,
|
| 1479 |
"train_batch_size": 4,
|
| 1480 |
"trial_name": null,
|
| 1481 |
"trial_params": null
|