Training in progress, step 2000, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 737580392
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8691ff71b090554b749389b543d86609cfacdab476bf9bcda11ba39222817d0d
|
| 3 |
size 737580392
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1475248442
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:25031bd9459b2ecca16cfe1a60ba2badffd9564c789d995f26a7db2a1e4a8e67
|
| 3 |
size 1475248442
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1000
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:db7eaec1ca57d40804cf0ff2a2ceba3277ffbb4ac8dc6548b227514adb077117
|
| 3 |
size 1000
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 2.0,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1057,6 +1057,356 @@
|
|
| 1057 |
"learning_rate": 2.912136832239925e-05,
|
| 1058 |
"loss": 0.3308,
|
| 1059 |
"step": 1500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1060 |
}
|
| 1061 |
],
|
| 1062 |
"logging_steps": 10,
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9372071227741331,
|
| 5 |
"eval_steps": 2.0,
|
| 6 |
+
"global_step": 2000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1057 |
"learning_rate": 2.912136832239925e-05,
|
| 1058 |
"loss": 0.3308,
|
| 1059 |
"step": 1500
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.7075913776944704,
|
| 1063 |
+
"grad_norm": 1.7443350553512573,
|
| 1064 |
+
"learning_rate": 2.9115510777881913e-05,
|
| 1065 |
+
"loss": 0.2742,
|
| 1066 |
+
"step": 1510
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.7122774133083412,
|
| 1070 |
+
"grad_norm": 1.9482985734939575,
|
| 1071 |
+
"learning_rate": 2.9109653233364575e-05,
|
| 1072 |
+
"loss": 0.3158,
|
| 1073 |
+
"step": 1520
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.7169634489222118,
|
| 1077 |
+
"grad_norm": 1.6775048971176147,
|
| 1078 |
+
"learning_rate": 2.9103795688847237e-05,
|
| 1079 |
+
"loss": 0.3222,
|
| 1080 |
+
"step": 1530
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.7216494845360825,
|
| 1084 |
+
"grad_norm": 1.6006051301956177,
|
| 1085 |
+
"learning_rate": 2.9097938144329896e-05,
|
| 1086 |
+
"loss": 0.303,
|
| 1087 |
+
"step": 1540
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.7263355201499532,
|
| 1091 |
+
"grad_norm": 2.021052360534668,
|
| 1092 |
+
"learning_rate": 2.909208059981256e-05,
|
| 1093 |
+
"loss": 0.2953,
|
| 1094 |
+
"step": 1550
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.7310215557638238,
|
| 1098 |
+
"grad_norm": 2.0027759075164795,
|
| 1099 |
+
"learning_rate": 2.908622305529522e-05,
|
| 1100 |
+
"loss": 0.3271,
|
| 1101 |
+
"step": 1560
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.7357075913776945,
|
| 1105 |
+
"grad_norm": 1.6766526699066162,
|
| 1106 |
+
"learning_rate": 2.9080365510777884e-05,
|
| 1107 |
+
"loss": 0.315,
|
| 1108 |
+
"step": 1570
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.7403936269915652,
|
| 1112 |
+
"grad_norm": 1.3581702709197998,
|
| 1113 |
+
"learning_rate": 2.9074507966260543e-05,
|
| 1114 |
+
"loss": 0.2944,
|
| 1115 |
+
"step": 1580
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.7450796626054358,
|
| 1119 |
+
"grad_norm": 1.6048952341079712,
|
| 1120 |
+
"learning_rate": 2.9068650421743205e-05,
|
| 1121 |
+
"loss": 0.3317,
|
| 1122 |
+
"step": 1590
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.7497656982193065,
|
| 1126 |
+
"grad_norm": 2.0446817874908447,
|
| 1127 |
+
"learning_rate": 2.9062792877225867e-05,
|
| 1128 |
+
"loss": 0.3297,
|
| 1129 |
+
"step": 1600
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.7544517338331771,
|
| 1133 |
+
"grad_norm": 2.3005971908569336,
|
| 1134 |
+
"learning_rate": 2.905693533270853e-05,
|
| 1135 |
+
"loss": 0.2981,
|
| 1136 |
+
"step": 1610
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.7591377694470478,
|
| 1140 |
+
"grad_norm": 1.601637840270996,
|
| 1141 |
+
"learning_rate": 2.905107778819119e-05,
|
| 1142 |
+
"loss": 0.3108,
|
| 1143 |
+
"step": 1620
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.7638238050609185,
|
| 1147 |
+
"grad_norm": 1.6310269832611084,
|
| 1148 |
+
"learning_rate": 2.904522024367385e-05,
|
| 1149 |
+
"loss": 0.2934,
|
| 1150 |
+
"step": 1630
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.7685098406747891,
|
| 1154 |
+
"grad_norm": 1.7142064571380615,
|
| 1155 |
+
"learning_rate": 2.9039362699156517e-05,
|
| 1156 |
+
"loss": 0.3136,
|
| 1157 |
+
"step": 1640
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.7731958762886598,
|
| 1161 |
+
"grad_norm": 1.5031518936157227,
|
| 1162 |
+
"learning_rate": 2.9033505154639176e-05,
|
| 1163 |
+
"loss": 0.3295,
|
| 1164 |
+
"step": 1650
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.7778819119025304,
|
| 1168 |
+
"grad_norm": 1.8852213621139526,
|
| 1169 |
+
"learning_rate": 2.902764761012184e-05,
|
| 1170 |
+
"loss": 0.3582,
|
| 1171 |
+
"step": 1660
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.7825679475164011,
|
| 1175 |
+
"grad_norm": 2.1107571125030518,
|
| 1176 |
+
"learning_rate": 2.90217900656045e-05,
|
| 1177 |
+
"loss": 0.2892,
|
| 1178 |
+
"step": 1670
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.7872539831302718,
|
| 1182 |
+
"grad_norm": 2.2586848735809326,
|
| 1183 |
+
"learning_rate": 2.9015932521087163e-05,
|
| 1184 |
+
"loss": 0.3282,
|
| 1185 |
+
"step": 1680
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.7919400187441424,
|
| 1189 |
+
"grad_norm": 1.761014461517334,
|
| 1190 |
+
"learning_rate": 2.9010074976569822e-05,
|
| 1191 |
+
"loss": 0.3297,
|
| 1192 |
+
"step": 1690
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.7966260543580131,
|
| 1196 |
+
"grad_norm": 1.51633882522583,
|
| 1197 |
+
"learning_rate": 2.9004217432052484e-05,
|
| 1198 |
+
"loss": 0.3342,
|
| 1199 |
+
"step": 1700
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.8013120899718837,
|
| 1203 |
+
"grad_norm": 1.6662074327468872,
|
| 1204 |
+
"learning_rate": 2.8998359887535147e-05,
|
| 1205 |
+
"loss": 0.3184,
|
| 1206 |
+
"step": 1710
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.8059981255857545,
|
| 1210 |
+
"grad_norm": 1.925766944885254,
|
| 1211 |
+
"learning_rate": 2.899250234301781e-05,
|
| 1212 |
+
"loss": 0.3004,
|
| 1213 |
+
"step": 1720
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.8106841611996252,
|
| 1217 |
+
"grad_norm": 2.039914846420288,
|
| 1218 |
+
"learning_rate": 2.8986644798500468e-05,
|
| 1219 |
+
"loss": 0.3238,
|
| 1220 |
+
"step": 1730
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.8153701968134958,
|
| 1224 |
+
"grad_norm": 1.7044464349746704,
|
| 1225 |
+
"learning_rate": 2.898078725398313e-05,
|
| 1226 |
+
"loss": 0.3099,
|
| 1227 |
+
"step": 1740
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.8200562324273665,
|
| 1231 |
+
"grad_norm": 2.0358059406280518,
|
| 1232 |
+
"learning_rate": 2.8974929709465793e-05,
|
| 1233 |
+
"loss": 0.309,
|
| 1234 |
+
"step": 1750
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.8247422680412371,
|
| 1238 |
+
"grad_norm": 1.9068161249160767,
|
| 1239 |
+
"learning_rate": 2.8969072164948455e-05,
|
| 1240 |
+
"loss": 0.3073,
|
| 1241 |
+
"step": 1760
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.8294283036551078,
|
| 1245 |
+
"grad_norm": 1.69276762008667,
|
| 1246 |
+
"learning_rate": 2.8963214620431114e-05,
|
| 1247 |
+
"loss": 0.2794,
|
| 1248 |
+
"step": 1770
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.8341143392689785,
|
| 1252 |
+
"grad_norm": 1.8938722610473633,
|
| 1253 |
+
"learning_rate": 2.8957357075913777e-05,
|
| 1254 |
+
"loss": 0.2988,
|
| 1255 |
+
"step": 1780
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.8388003748828491,
|
| 1259 |
+
"grad_norm": 1.361587405204773,
|
| 1260 |
+
"learning_rate": 2.895149953139644e-05,
|
| 1261 |
+
"loss": 0.3045,
|
| 1262 |
+
"step": 1790
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.8434864104967198,
|
| 1266 |
+
"grad_norm": 1.8083635568618774,
|
| 1267 |
+
"learning_rate": 2.89456419868791e-05,
|
| 1268 |
+
"loss": 0.2373,
|
| 1269 |
+
"step": 1800
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.8481724461105904,
|
| 1273 |
+
"grad_norm": 1.8941372632980347,
|
| 1274 |
+
"learning_rate": 2.893978444236176e-05,
|
| 1275 |
+
"loss": 0.2969,
|
| 1276 |
+
"step": 1810
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.8528584817244611,
|
| 1280 |
+
"grad_norm": 1.6231937408447266,
|
| 1281 |
+
"learning_rate": 2.8933926897844423e-05,
|
| 1282 |
+
"loss": 0.3047,
|
| 1283 |
+
"step": 1820
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.8575445173383318,
|
| 1287 |
+
"grad_norm": 1.9185701608657837,
|
| 1288 |
+
"learning_rate": 2.8928069353327085e-05,
|
| 1289 |
+
"loss": 0.2988,
|
| 1290 |
+
"step": 1830
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.8622305529522024,
|
| 1294 |
+
"grad_norm": 2.167684316635132,
|
| 1295 |
+
"learning_rate": 2.8922211808809748e-05,
|
| 1296 |
+
"loss": 0.3174,
|
| 1297 |
+
"step": 1840
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.8669165885660731,
|
| 1301 |
+
"grad_norm": 1.5613234043121338,
|
| 1302 |
+
"learning_rate": 2.8916354264292407e-05,
|
| 1303 |
+
"loss": 0.3034,
|
| 1304 |
+
"step": 1850
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.8716026241799437,
|
| 1308 |
+
"grad_norm": 1.938776969909668,
|
| 1309 |
+
"learning_rate": 2.8910496719775072e-05,
|
| 1310 |
+
"loss": 0.3212,
|
| 1311 |
+
"step": 1860
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.8762886597938144,
|
| 1315 |
+
"grad_norm": 1.797566294670105,
|
| 1316 |
+
"learning_rate": 2.8904639175257735e-05,
|
| 1317 |
+
"loss": 0.2966,
|
| 1318 |
+
"step": 1870
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.8809746954076851,
|
| 1322 |
+
"grad_norm": 1.7625051736831665,
|
| 1323 |
+
"learning_rate": 2.8898781630740394e-05,
|
| 1324 |
+
"loss": 0.2957,
|
| 1325 |
+
"step": 1880
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.8856607310215557,
|
| 1329 |
+
"grad_norm": 2.0054209232330322,
|
| 1330 |
+
"learning_rate": 2.8892924086223056e-05,
|
| 1331 |
+
"loss": 0.3161,
|
| 1332 |
+
"step": 1890
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.8903467666354264,
|
| 1336 |
+
"grad_norm": 1.7838996648788452,
|
| 1337 |
+
"learning_rate": 2.888706654170572e-05,
|
| 1338 |
+
"loss": 0.3208,
|
| 1339 |
+
"step": 1900
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.895032802249297,
|
| 1343 |
+
"grad_norm": 2.5746254920959473,
|
| 1344 |
+
"learning_rate": 2.888120899718838e-05,
|
| 1345 |
+
"loss": 0.3338,
|
| 1346 |
+
"step": 1910
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.8997188378631678,
|
| 1350 |
+
"grad_norm": 2.0083072185516357,
|
| 1351 |
+
"learning_rate": 2.887535145267104e-05,
|
| 1352 |
+
"loss": 0.3,
|
| 1353 |
+
"step": 1920
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.9044048734770385,
|
| 1357 |
+
"grad_norm": 1.6170719861984253,
|
| 1358 |
+
"learning_rate": 2.8869493908153702e-05,
|
| 1359 |
+
"loss": 0.2895,
|
| 1360 |
+
"step": 1930
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.9090909090909091,
|
| 1364 |
+
"grad_norm": 1.9582558870315552,
|
| 1365 |
+
"learning_rate": 2.8863636363636365e-05,
|
| 1366 |
+
"loss": 0.2849,
|
| 1367 |
+
"step": 1940
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.9137769447047798,
|
| 1371 |
+
"grad_norm": 1.850935697555542,
|
| 1372 |
+
"learning_rate": 2.8857778819119027e-05,
|
| 1373 |
+
"loss": 0.2822,
|
| 1374 |
+
"step": 1950
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.9184629803186504,
|
| 1378 |
+
"grad_norm": 2.373908042907715,
|
| 1379 |
+
"learning_rate": 2.8851921274601686e-05,
|
| 1380 |
+
"loss": 0.3264,
|
| 1381 |
+
"step": 1960
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.9231490159325211,
|
| 1385 |
+
"grad_norm": 2.1133532524108887,
|
| 1386 |
+
"learning_rate": 2.884606373008435e-05,
|
| 1387 |
+
"loss": 0.3108,
|
| 1388 |
+
"step": 1970
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.9278350515463918,
|
| 1392 |
+
"grad_norm": 1.951541543006897,
|
| 1393 |
+
"learning_rate": 2.884020618556701e-05,
|
| 1394 |
+
"loss": 0.264,
|
| 1395 |
+
"step": 1980
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.9325210871602624,
|
| 1399 |
+
"grad_norm": 1.9710968732833862,
|
| 1400 |
+
"learning_rate": 2.8834348641049673e-05,
|
| 1401 |
+
"loss": 0.3616,
|
| 1402 |
+
"step": 1990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.9372071227741331,
|
| 1406 |
+
"grad_norm": 1.7492647171020508,
|
| 1407 |
+
"learning_rate": 2.8828491096532332e-05,
|
| 1408 |
+
"loss": 0.3261,
|
| 1409 |
+
"step": 2000
|
| 1410 |
}
|
| 1411 |
],
|
| 1412 |
"logging_steps": 10,
|