Training in progress, step 195, 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 140815952
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:04f72de55f528d8ea8742a046bb1497e86c2b72142c11c42c065b5d323d499b0
|
| 3 |
size 140815952
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 71878612
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cf923b202fe9c5a0306932baadf74064198cd05e038a96060a9d584a6fc82910
|
| 3 |
size 71878612
|
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:a44578a30646300ac5f324ca30388c75c7181bc597ed808a77a04b54fd005aab
|
| 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 sha256:7535d6d6d3346211338a559c66a34e5433ea456734f0f5c94e8703828d95ba57
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.4009778499603271,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-50",
|
| 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 +1089,321 @@
|
|
| 1089 |
"eval_samples_per_second": 57.273,
|
| 1090 |
"eval_steps_per_second": 14.579,
|
| 1091 |
"step": 150
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1092 |
}
|
| 1093 |
],
|
| 1094 |
"logging_steps": 1,
|
|
@@ -1112,12 +1427,12 @@
|
|
| 1112 |
"should_evaluate": false,
|
| 1113 |
"should_log": false,
|
| 1114 |
"should_save": true,
|
| 1115 |
-
"should_training_stop":
|
| 1116 |
},
|
| 1117 |
"attributes": {}
|
| 1118 |
}
|
| 1119 |
},
|
| 1120 |
-
"total_flos": 1.
|
| 1121 |
"train_batch_size": 8,
|
| 1122 |
"trial_name": null,
|
| 1123 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.4009778499603271,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-50",
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 195,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1089 |
"eval_samples_per_second": 57.273,
|
| 1090 |
"eval_steps_per_second": 14.579,
|
| 1091 |
"step": 150
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 2.3230769230769233,
|
| 1095 |
+
"grad_norm": 4.365710258483887,
|
| 1096 |
+
"learning_rate": 1.3319917421293182e-05,
|
| 1097 |
+
"loss": 0.7168,
|
| 1098 |
+
"step": 151
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 2.3384615384615386,
|
| 1102 |
+
"grad_norm": 4.27862548828125,
|
| 1103 |
+
"learning_rate": 1.2748216872812745e-05,
|
| 1104 |
+
"loss": 0.7446,
|
| 1105 |
+
"step": 152
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 2.353846153846154,
|
| 1109 |
+
"grad_norm": 3.9131150245666504,
|
| 1110 |
+
"learning_rate": 1.2187258516515642e-05,
|
| 1111 |
+
"loss": 0.6219,
|
| 1112 |
+
"step": 153
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 2.3692307692307693,
|
| 1116 |
+
"grad_norm": 3.5391688346862793,
|
| 1117 |
+
"learning_rate": 1.1637204114389177e-05,
|
| 1118 |
+
"loss": 0.5397,
|
| 1119 |
+
"step": 154
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 2.3846153846153846,
|
| 1123 |
+
"grad_norm": 3.1678285598754883,
|
| 1124 |
+
"learning_rate": 1.1098212284078036e-05,
|
| 1125 |
+
"loss": 0.5958,
|
| 1126 |
+
"step": 155
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 2.4,
|
| 1130 |
+
"grad_norm": 2.9963395595550537,
|
| 1131 |
+
"learning_rate": 1.0570438453144043e-05,
|
| 1132 |
+
"loss": 0.5378,
|
| 1133 |
+
"step": 156
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 2.4153846153846152,
|
| 1137 |
+
"grad_norm": 3.137876510620117,
|
| 1138 |
+
"learning_rate": 1.0054034814246094e-05,
|
| 1139 |
+
"loss": 0.4767,
|
| 1140 |
+
"step": 157
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 2.430769230769231,
|
| 1144 |
+
"grad_norm": 3.598328113555908,
|
| 1145 |
+
"learning_rate": 9.549150281252633e-06,
|
| 1146 |
+
"loss": 0.5279,
|
| 1147 |
+
"step": 158
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 2.4461538461538463,
|
| 1151 |
+
"grad_norm": 2.911151885986328,
|
| 1152 |
+
"learning_rate": 9.055930446299915e-06,
|
| 1153 |
+
"loss": 0.4513,
|
| 1154 |
+
"step": 159
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 2.4615384615384617,
|
| 1158 |
+
"grad_norm": 3.33564829826355,
|
| 1159 |
+
"learning_rate": 8.574517537807897e-06,
|
| 1160 |
+
"loss": 0.5493,
|
| 1161 |
+
"step": 160
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 2.476923076923077,
|
| 1165 |
+
"grad_norm": 3.348562240600586,
|
| 1166 |
+
"learning_rate": 8.105050379466333e-06,
|
| 1167 |
+
"loss": 0.6329,
|
| 1168 |
+
"step": 161
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 2.4923076923076923,
|
| 1172 |
+
"grad_norm": 3.2942681312561035,
|
| 1173 |
+
"learning_rate": 7.64766435020246e-06,
|
| 1174 |
+
"loss": 0.4799,
|
| 1175 |
+
"step": 162
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 2.5076923076923077,
|
| 1179 |
+
"grad_norm": 2.794443368911743,
|
| 1180 |
+
"learning_rate": 7.2024913451422875e-06,
|
| 1181 |
+
"loss": 0.636,
|
| 1182 |
+
"step": 163
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 2.523076923076923,
|
| 1186 |
+
"grad_norm": 3.167689085006714,
|
| 1187 |
+
"learning_rate": 6.769659737576229e-06,
|
| 1188 |
+
"loss": 0.7126,
|
| 1189 |
+
"step": 164
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 2.5384615384615383,
|
| 1193 |
+
"grad_norm": 2.878326654434204,
|
| 1194 |
+
"learning_rate": 6.349294341940593e-06,
|
| 1195 |
+
"loss": 0.6393,
|
| 1196 |
+
"step": 165
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 2.5538461538461537,
|
| 1200 |
+
"grad_norm": 3.4416472911834717,
|
| 1201 |
+
"learning_rate": 5.941516377825102e-06,
|
| 1202 |
+
"loss": 0.7256,
|
| 1203 |
+
"step": 166
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 2.569230769230769,
|
| 1207 |
+
"grad_norm": 3.32590651512146,
|
| 1208 |
+
"learning_rate": 5.546443435017146e-06,
|
| 1209 |
+
"loss": 0.6606,
|
| 1210 |
+
"step": 167
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 2.5846153846153848,
|
| 1214 |
+
"grad_norm": 3.124803066253662,
|
| 1215 |
+
"learning_rate": 5.1641894395928205e-06,
|
| 1216 |
+
"loss": 0.5744,
|
| 1217 |
+
"step": 168
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 2.6,
|
| 1221 |
+
"grad_norm": 3.3172974586486816,
|
| 1222 |
+
"learning_rate": 4.794864621064266e-06,
|
| 1223 |
+
"loss": 0.6565,
|
| 1224 |
+
"step": 169
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 2.6153846153846154,
|
| 1228 |
+
"grad_norm": 2.9242913722991943,
|
| 1229 |
+
"learning_rate": 4.43857548059321e-06,
|
| 1230 |
+
"loss": 0.5911,
|
| 1231 |
+
"step": 170
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 2.6307692307692307,
|
| 1235 |
+
"grad_norm": 3.0626587867736816,
|
| 1236 |
+
"learning_rate": 4.095424760279454e-06,
|
| 1237 |
+
"loss": 0.5574,
|
| 1238 |
+
"step": 171
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 2.646153846153846,
|
| 1242 |
+
"grad_norm": 3.1158361434936523,
|
| 1243 |
+
"learning_rate": 3.765511413533429e-06,
|
| 1244 |
+
"loss": 0.5037,
|
| 1245 |
+
"step": 172
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 2.6615384615384614,
|
| 1249 |
+
"grad_norm": 2.95619797706604,
|
| 1250 |
+
"learning_rate": 3.448930576541309e-06,
|
| 1251 |
+
"loss": 0.5908,
|
| 1252 |
+
"step": 173
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 2.676923076923077,
|
| 1256 |
+
"grad_norm": 3.1260483264923096,
|
| 1257 |
+
"learning_rate": 3.1457735408308154e-06,
|
| 1258 |
+
"loss": 0.564,
|
| 1259 |
+
"step": 174
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 2.6923076923076925,
|
| 1263 |
+
"grad_norm": 3.150573253631592,
|
| 1264 |
+
"learning_rate": 2.85612772694579e-06,
|
| 1265 |
+
"loss": 0.5157,
|
| 1266 |
+
"step": 175
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 2.707692307692308,
|
| 1270 |
+
"grad_norm": 3.3056576251983643,
|
| 1271 |
+
"learning_rate": 2.5800766592369073e-06,
|
| 1272 |
+
"loss": 0.4903,
|
| 1273 |
+
"step": 176
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 2.723076923076923,
|
| 1277 |
+
"grad_norm": 3.4288980960845947,
|
| 1278 |
+
"learning_rate": 2.317699941776064e-06,
|
| 1279 |
+
"loss": 0.6538,
|
| 1280 |
+
"step": 177
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 2.7384615384615385,
|
| 1284 |
+
"grad_norm": 2.5557122230529785,
|
| 1285 |
+
"learning_rate": 2.069073235401109e-06,
|
| 1286 |
+
"loss": 0.331,
|
| 1287 |
+
"step": 178
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 2.753846153846154,
|
| 1291 |
+
"grad_norm": 2.8993163108825684,
|
| 1292 |
+
"learning_rate": 1.8342682358978069e-06,
|
| 1293 |
+
"loss": 0.6332,
|
| 1294 |
+
"step": 179
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 2.769230769230769,
|
| 1298 |
+
"grad_norm": 3.0784480571746826,
|
| 1299 |
+
"learning_rate": 1.6133526533250565e-06,
|
| 1300 |
+
"loss": 0.6591,
|
| 1301 |
+
"step": 180
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 2.7846153846153845,
|
| 1305 |
+
"grad_norm": 3.570091485977173,
|
| 1306 |
+
"learning_rate": 1.4063901924895984e-06,
|
| 1307 |
+
"loss": 0.7268,
|
| 1308 |
+
"step": 181
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 2.8,
|
| 1312 |
+
"grad_norm": 3.490947723388672,
|
| 1313 |
+
"learning_rate": 1.2134405345755773e-06,
|
| 1314 |
+
"loss": 0.8519,
|
| 1315 |
+
"step": 182
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 2.815384615384615,
|
| 1319 |
+
"grad_norm": 3.4577839374542236,
|
| 1320 |
+
"learning_rate": 1.034559319934497e-06,
|
| 1321 |
+
"loss": 0.6088,
|
| 1322 |
+
"step": 183
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 2.830769230769231,
|
| 1326 |
+
"grad_norm": 3.753882884979248,
|
| 1327 |
+
"learning_rate": 8.697981320403337e-07,
|
| 1328 |
+
"loss": 0.7528,
|
| 1329 |
+
"step": 184
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 2.8461538461538463,
|
| 1333 |
+
"grad_norm": 3.232329845428467,
|
| 1334 |
+
"learning_rate": 7.192044826145771e-07,
|
| 1335 |
+
"loss": 0.5498,
|
| 1336 |
+
"step": 185
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 2.8615384615384616,
|
| 1340 |
+
"grad_norm": 2.8896830081939697,
|
| 1341 |
+
"learning_rate": 5.828217979253869e-07,
|
| 1342 |
+
"loss": 0.6241,
|
| 1343 |
+
"step": 186
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 2.876923076923077,
|
| 1347 |
+
"grad_norm": 3.106513261795044,
|
| 1348 |
+
"learning_rate": 4.60689406264897e-07,
|
| 1349 |
+
"loss": 0.5324,
|
| 1350 |
+
"step": 187
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 2.8923076923076922,
|
| 1354 |
+
"grad_norm": 3.1418616771698,
|
| 1355 |
+
"learning_rate": 3.528425266082325e-07,
|
| 1356 |
+
"loss": 0.5798,
|
| 1357 |
+
"step": 188
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 2.9076923076923076,
|
| 1361 |
+
"grad_norm": 2.6466550827026367,
|
| 1362 |
+
"learning_rate": 2.593122584574892e-07,
|
| 1363 |
+
"loss": 0.5133,
|
| 1364 |
+
"step": 189
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 2.9230769230769234,
|
| 1368 |
+
"grad_norm": 3.7142605781555176,
|
| 1369 |
+
"learning_rate": 1.8012557287367392e-07,
|
| 1370 |
+
"loss": 0.6735,
|
| 1371 |
+
"step": 190
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 2.9384615384615387,
|
| 1375 |
+
"grad_norm": 3.4207851886749268,
|
| 1376 |
+
"learning_rate": 1.1530530469914259e-07,
|
| 1377 |
+
"loss": 0.6708,
|
| 1378 |
+
"step": 191
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 2.953846153846154,
|
| 1382 |
+
"grad_norm": 3.6083593368530273,
|
| 1383 |
+
"learning_rate": 6.487014597275631e-08,
|
| 1384 |
+
"loss": 0.6976,
|
| 1385 |
+
"step": 192
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 2.9692307692307693,
|
| 1389 |
+
"grad_norm": 3.2904462814331055,
|
| 1390 |
+
"learning_rate": 2.8834640539737723e-08,
|
| 1391 |
+
"loss": 0.5393,
|
| 1392 |
+
"step": 193
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 2.9846153846153847,
|
| 1396 |
+
"grad_norm": 3.3836851119995117,
|
| 1397 |
+
"learning_rate": 7.209179857675663e-09,
|
| 1398 |
+
"loss": 0.4512,
|
| 1399 |
+
"step": 194
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 3.0,
|
| 1403 |
+
"grad_norm": 3.2520158290863037,
|
| 1404 |
+
"learning_rate": 0.0,
|
| 1405 |
+
"loss": 0.5668,
|
| 1406 |
+
"step": 195
|
| 1407 |
}
|
| 1408 |
],
|
| 1409 |
"logging_steps": 1,
|
|
|
|
| 1427 |
"should_evaluate": false,
|
| 1428 |
"should_log": false,
|
| 1429 |
"should_save": true,
|
| 1430 |
+
"should_training_stop": true
|
| 1431 |
},
|
| 1432 |
"attributes": {}
|
| 1433 |
}
|
| 1434 |
},
|
| 1435 |
+
"total_flos": 1.507054074200064e+16,
|
| 1436 |
"train_batch_size": 8,
|
| 1437 |
"trial_name": null,
|
| 1438 |
"trial_params": null
|