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 34456
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c3434d041e2af161aa9d1fcfd5e8d13b18f7691921ba75b84251ff923f61f8a2
|
| 3 |
size 34456
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 76154
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0e5e26fc1cf7716d55fbe44ac43da7e753301aee222dec9612bc9e5b15157771
|
| 3 |
size 76154
|
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:a0390c1b2ba83e00dadbb33f678bd10aae2b89812bb45dbf0c238503ba1c6afe
|
| 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:d2d754412c61116546142914503e7369d0cc35d3c380a07e5218f595d76b6d96
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 11.5,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-50",
|
| 4 |
-
"epoch": 1.
|
| 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": 160.044,
|
| 1090 |
"eval_steps_per_second": 80.464,
|
| 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": 8,
|
| 1122 |
"trial_name": null,
|
| 1123 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 11.5,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-50",
|
| 4 |
+
"epoch": 1.8691588785046729,
|
| 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": 160.044,
|
| 1090 |
"eval_steps_per_second": 80.464,
|
| 1091 |
"step": 150
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 1.411214953271028,
|
| 1095 |
+
"grad_norm": 0.00040427924250252545,
|
| 1096 |
+
"learning_rate": 1.9136088935510362e-05,
|
| 1097 |
+
"loss": 46.0,
|
| 1098 |
+
"step": 151
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 1.4205607476635513,
|
| 1102 |
+
"grad_norm": 0.000727128644939512,
|
| 1103 |
+
"learning_rate": 1.8414449687337464e-05,
|
| 1104 |
+
"loss": 46.0,
|
| 1105 |
+
"step": 152
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 1.4299065420560748,
|
| 1109 |
+
"grad_norm": 0.0006024191388860345,
|
| 1110 |
+
"learning_rate": 1.7703596875660645e-05,
|
| 1111 |
+
"loss": 46.0,
|
| 1112 |
+
"step": 153
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 1.439252336448598,
|
| 1116 |
+
"grad_norm": 0.00046983369975350797,
|
| 1117 |
+
"learning_rate": 1.700377325606388e-05,
|
| 1118 |
+
"loss": 46.0,
|
| 1119 |
+
"step": 154
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 1.4485981308411215,
|
| 1123 |
+
"grad_norm": 0.0007036470924504101,
|
| 1124 |
+
"learning_rate": 1.631521781767214e-05,
|
| 1125 |
+
"loss": 46.0,
|
| 1126 |
+
"step": 155
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 1.4579439252336448,
|
| 1130 |
+
"grad_norm": 0.0006356867379508913,
|
| 1131 |
+
"learning_rate": 1.5638165701536868e-05,
|
| 1132 |
+
"loss": 46.0,
|
| 1133 |
+
"step": 156
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 1.4672897196261683,
|
| 1137 |
+
"grad_norm": 0.00041911471635103226,
|
| 1138 |
+
"learning_rate": 1.4972848120335453e-05,
|
| 1139 |
+
"loss": 46.0,
|
| 1140 |
+
"step": 157
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 1.4766355140186915,
|
| 1144 |
+
"grad_norm": 0.0004606099391821772,
|
| 1145 |
+
"learning_rate": 1.4319492279412388e-05,
|
| 1146 |
+
"loss": 46.0,
|
| 1147 |
+
"step": 158
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 1.485981308411215,
|
| 1151 |
+
"grad_norm": 0.0005386231350712478,
|
| 1152 |
+
"learning_rate": 1.3678321299188801e-05,
|
| 1153 |
+
"loss": 46.0,
|
| 1154 |
+
"step": 159
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 1.4953271028037383,
|
| 1158 |
+
"grad_norm": 0.0004815705760847777,
|
| 1159 |
+
"learning_rate": 1.3049554138967051e-05,
|
| 1160 |
+
"loss": 46.0,
|
| 1161 |
+
"step": 160
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 1.5046728971962615,
|
| 1165 |
+
"grad_norm": 0.000545655726455152,
|
| 1166 |
+
"learning_rate": 1.2433405522156332e-05,
|
| 1167 |
+
"loss": 46.0,
|
| 1168 |
+
"step": 161
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 1.514018691588785,
|
| 1172 |
+
"grad_norm": 0.00040522534982301295,
|
| 1173 |
+
"learning_rate": 1.183008586294485e-05,
|
| 1174 |
+
"loss": 46.0,
|
| 1175 |
+
"step": 162
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 1.5233644859813085,
|
| 1179 |
+
"grad_norm": 0.0005212983232922852,
|
| 1180 |
+
"learning_rate": 1.1239801194443506e-05,
|
| 1181 |
+
"loss": 46.0,
|
| 1182 |
+
"step": 163
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 1.5327102803738317,
|
| 1186 |
+
"grad_norm": 0.0006503637996502221,
|
| 1187 |
+
"learning_rate": 1.066275309832584e-05,
|
| 1188 |
+
"loss": 46.0,
|
| 1189 |
+
"step": 164
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 1.542056074766355,
|
| 1193 |
+
"grad_norm": 0.0006243106327019632,
|
| 1194 |
+
"learning_rate": 1.0099138635988026e-05,
|
| 1195 |
+
"loss": 46.0,
|
| 1196 |
+
"step": 165
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 1.5514018691588785,
|
| 1200 |
+
"grad_norm": 0.0005656758439727128,
|
| 1201 |
+
"learning_rate": 9.549150281252633e-06,
|
| 1202 |
+
"loss": 46.0,
|
| 1203 |
+
"step": 166
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 1.560747663551402,
|
| 1207 |
+
"grad_norm": 0.000596692378167063,
|
| 1208 |
+
"learning_rate": 9.012975854638949e-06,
|
| 1209 |
+
"loss": 46.0,
|
| 1210 |
+
"step": 167
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 1.5700934579439252,
|
| 1214 |
+
"grad_norm": 0.0006169018452055752,
|
| 1215 |
+
"learning_rate": 8.490798459222476e-06,
|
| 1216 |
+
"loss": 46.0,
|
| 1217 |
+
"step": 168
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 1.5794392523364484,
|
| 1221 |
+
"grad_norm": 0.00040590198477730155,
|
| 1222 |
+
"learning_rate": 7.982796418105371e-06,
|
| 1223 |
+
"loss": 46.0,
|
| 1224 |
+
"step": 169
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 1.588785046728972,
|
| 1228 |
+
"grad_norm": 0.000642999482806772,
|
| 1229 |
+
"learning_rate": 7.489143213519301e-06,
|
| 1230 |
+
"loss": 46.0,
|
| 1231 |
+
"step": 170
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 1.5981308411214954,
|
| 1235 |
+
"grad_norm": 0.00047177704982459545,
|
| 1236 |
+
"learning_rate": 7.010007427581378e-06,
|
| 1237 |
+
"loss": 46.0,
|
| 1238 |
+
"step": 171
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 1.6074766355140186,
|
| 1242 |
+
"grad_norm": 0.0006852488149888813,
|
| 1243 |
+
"learning_rate": 6.5455526847235825e-06,
|
| 1244 |
+
"loss": 46.0,
|
| 1245 |
+
"step": 172
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 1.616822429906542,
|
| 1249 |
+
"grad_norm": 0.0005702375201508403,
|
| 1250 |
+
"learning_rate": 6.0959375958151045e-06,
|
| 1251 |
+
"loss": 46.0,
|
| 1252 |
+
"step": 173
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 1.6261682242990654,
|
| 1256 |
+
"grad_norm": 0.0003300758544355631,
|
| 1257 |
+
"learning_rate": 5.6613157039969055e-06,
|
| 1258 |
+
"loss": 46.0,
|
| 1259 |
+
"step": 174
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 1.6355140186915889,
|
| 1263 |
+
"grad_norm": 0.0005260781035758555,
|
| 1264 |
+
"learning_rate": 5.241835432246889e-06,
|
| 1265 |
+
"loss": 46.0,
|
| 1266 |
+
"step": 175
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 1.644859813084112,
|
| 1270 |
+
"grad_norm": 0.0005789488204754889,
|
| 1271 |
+
"learning_rate": 4.837640032693558e-06,
|
| 1272 |
+
"loss": 46.0,
|
| 1273 |
+
"step": 176
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 1.6542056074766354,
|
| 1277 |
+
"grad_norm": 0.0008496741647832096,
|
| 1278 |
+
"learning_rate": 4.448867537695578e-06,
|
| 1279 |
+
"loss": 46.0,
|
| 1280 |
+
"step": 177
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 1.6635514018691588,
|
| 1284 |
+
"grad_norm": 0.0009718951187096536,
|
| 1285 |
+
"learning_rate": 4.075650712703849e-06,
|
| 1286 |
+
"loss": 46.0,
|
| 1287 |
+
"step": 178
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 1.6728971962616823,
|
| 1291 |
+
"grad_norm": 0.00039575292612425983,
|
| 1292 |
+
"learning_rate": 3.71811701092219e-06,
|
| 1293 |
+
"loss": 46.0,
|
| 1294 |
+
"step": 179
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 1.6822429906542056,
|
| 1298 |
+
"grad_norm": 0.00040399827412329614,
|
| 1299 |
+
"learning_rate": 3.376388529782215e-06,
|
| 1300 |
+
"loss": 46.0,
|
| 1301 |
+
"step": 180
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 1.6915887850467288,
|
| 1305 |
+
"grad_norm": 0.0006307908333837986,
|
| 1306 |
+
"learning_rate": 3.0505819692471792e-06,
|
| 1307 |
+
"loss": 46.0,
|
| 1308 |
+
"step": 181
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 1.7009345794392523,
|
| 1312 |
+
"grad_norm": 0.0008413252071477473,
|
| 1313 |
+
"learning_rate": 2.7408085919590264e-06,
|
| 1314 |
+
"loss": 46.0,
|
| 1315 |
+
"step": 182
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 1.7102803738317758,
|
| 1319 |
+
"grad_norm": 0.0006041501183062792,
|
| 1320 |
+
"learning_rate": 2.4471741852423237e-06,
|
| 1321 |
+
"loss": 46.0,
|
| 1322 |
+
"step": 183
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 1.719626168224299,
|
| 1326 |
+
"grad_norm": 0.0006907913484610617,
|
| 1327 |
+
"learning_rate": 2.1697790249779636e-06,
|
| 1328 |
+
"loss": 46.0,
|
| 1329 |
+
"step": 184
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 1.7289719626168223,
|
| 1333 |
+
"grad_norm": 0.0008346527465619147,
|
| 1334 |
+
"learning_rate": 1.908717841359048e-06,
|
| 1335 |
+
"loss": 46.0,
|
| 1336 |
+
"step": 185
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 1.7383177570093458,
|
| 1340 |
+
"grad_norm": 0.00033760842052288353,
|
| 1341 |
+
"learning_rate": 1.6640797865406288e-06,
|
| 1342 |
+
"loss": 46.0,
|
| 1343 |
+
"step": 186
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 1.7476635514018692,
|
| 1347 |
+
"grad_norm": 0.000893876887857914,
|
| 1348 |
+
"learning_rate": 1.4359484041943038e-06,
|
| 1349 |
+
"loss": 46.0,
|
| 1350 |
+
"step": 187
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 1.7570093457943925,
|
| 1354 |
+
"grad_norm": 0.000611237483099103,
|
| 1355 |
+
"learning_rate": 1.2244016009781701e-06,
|
| 1356 |
+
"loss": 46.0,
|
| 1357 |
+
"step": 188
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 1.7663551401869158,
|
| 1361 |
+
"grad_norm": 0.0004157586663495749,
|
| 1362 |
+
"learning_rate": 1.0295116199317057e-06,
|
| 1363 |
+
"loss": 46.0,
|
| 1364 |
+
"step": 189
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 1.7757009345794392,
|
| 1368 |
+
"grad_norm": 0.00046051564277149737,
|
| 1369 |
+
"learning_rate": 8.513450158049108e-07,
|
| 1370 |
+
"loss": 46.0,
|
| 1371 |
+
"step": 190
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 1.7850467289719627,
|
| 1375 |
+
"grad_norm": 0.0004915129975415766,
|
| 1376 |
+
"learning_rate": 6.899626323298713e-07,
|
| 1377 |
+
"loss": 46.0,
|
| 1378 |
+
"step": 191
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 1.794392523364486,
|
| 1382 |
+
"grad_norm": 0.0005064051947556436,
|
| 1383 |
+
"learning_rate": 5.454195814427021e-07,
|
| 1384 |
+
"loss": 46.0,
|
| 1385 |
+
"step": 192
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 1.8037383177570092,
|
| 1389 |
+
"grad_norm": 0.0005573919625021517,
|
| 1390 |
+
"learning_rate": 4.177652244628627e-07,
|
| 1391 |
+
"loss": 46.0,
|
| 1392 |
+
"step": 193
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 1.8130841121495327,
|
| 1396 |
+
"grad_norm": 0.0006023433525115252,
|
| 1397 |
+
"learning_rate": 3.0704315523631953e-07,
|
| 1398 |
+
"loss": 46.0,
|
| 1399 |
+
"step": 194
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 1.8224299065420562,
|
| 1403 |
+
"grad_norm": 0.0005069277249276638,
|
| 1404 |
+
"learning_rate": 2.1329118524827662e-07,
|
| 1405 |
+
"loss": 46.0,
|
| 1406 |
+
"step": 195
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 1.8317757009345794,
|
| 1410 |
+
"grad_norm": 0.0007788498187437654,
|
| 1411 |
+
"learning_rate": 1.3654133071059893e-07,
|
| 1412 |
+
"loss": 46.0,
|
| 1413 |
+
"step": 196
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 1.8411214953271027,
|
| 1417 |
+
"grad_norm": 0.0005983134033158422,
|
| 1418 |
+
"learning_rate": 7.681980162830282e-08,
|
| 1419 |
+
"loss": 46.0,
|
| 1420 |
+
"step": 197
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 1.8504672897196262,
|
| 1424 |
+
"grad_norm": 0.0005851113237440586,
|
| 1425 |
+
"learning_rate": 3.4146992848854695e-08,
|
| 1426 |
+
"loss": 46.0,
|
| 1427 |
+
"step": 198
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 1.8598130841121496,
|
| 1431 |
+
"grad_norm": 0.0003598025650717318,
|
| 1432 |
+
"learning_rate": 8.537477097364522e-09,
|
| 1433 |
+
"loss": 46.0,
|
| 1434 |
+
"step": 199
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 1.8691588785046729,
|
| 1438 |
+
"grad_norm": 0.0006664404063485563,
|
| 1439 |
+
"learning_rate": 0.0,
|
| 1440 |
+
"loss": 46.0,
|
| 1441 |
+
"step": 200
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 1.8691588785046729,
|
| 1445 |
+
"eval_loss": 11.5,
|
| 1446 |
+
"eval_runtime": 1.1238,
|
| 1447 |
+
"eval_samples_per_second": 161.063,
|
| 1448 |
+
"eval_steps_per_second": 80.977,
|
| 1449 |
+
"step": 200
|
| 1450 |
}
|
| 1451 |
],
|
| 1452 |
"logging_steps": 1,
|
|
|
|
| 1461 |
"early_stopping_threshold": 0.0
|
| 1462 |
},
|
| 1463 |
"attributes": {
|
| 1464 |
+
"early_stopping_patience_counter": 3
|
| 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": 32158462771200.0,
|
| 1479 |
"train_batch_size": 8,
|
| 1480 |
"trial_name": null,
|
| 1481 |
"trial_params": null
|