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 138995824
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:70e6fb29bbac3ac8deba4dc4919ed5461780689abeaebae628173b367ec12d5a
|
| 3 |
size 138995824
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 71077780
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c728a4c89f732b9a2b2245d6c166ec9f060b948479d811f35c66c4451762905f
|
| 3 |
size 71077780
|
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:a75caa12c9e2db40b89286a803642d8a730328847614ab867df1e648fc1e94f0
|
| 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": 2.
|
| 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": 59.531,
|
| 1090 |
"eval_steps_per_second": 14.885,
|
| 1091 |
"step": 150
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1092 |
}
|
| 1093 |
],
|
| 1094 |
"logging_steps": 1,
|
|
@@ -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": 1.
|
| 1121 |
"train_batch_size": 8,
|
| 1122 |
"trial_name": null,
|
| 1123 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 2.242037057876587,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 4 |
+
"epoch": 0.04071039641748511,
|
| 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": 59.531,
|
| 1090 |
"eval_steps_per_second": 14.885,
|
| 1091 |
"step": 150
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 0.030736349295201262,
|
| 1095 |
+
"grad_norm": 0.4016248285770416,
|
| 1096 |
+
"learning_rate": 1.553232954407171e-05,
|
| 1097 |
+
"loss": 2.4679,
|
| 1098 |
+
"step": 151
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 0.030939901277288687,
|
| 1102 |
+
"grad_norm": 0.4319489300251007,
|
| 1103 |
+
"learning_rate": 1.4938160786375572e-05,
|
| 1104 |
+
"loss": 2.4574,
|
| 1105 |
+
"step": 152
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 0.031143453259376112,
|
| 1109 |
+
"grad_norm": 0.46305328607559204,
|
| 1110 |
+
"learning_rate": 1.435357758543015e-05,
|
| 1111 |
+
"loss": 2.496,
|
| 1112 |
+
"step": 153
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.03134700524146354,
|
| 1116 |
+
"grad_norm": 0.45785951614379883,
|
| 1117 |
+
"learning_rate": 1.3778739760445552e-05,
|
| 1118 |
+
"loss": 2.5628,
|
| 1119 |
+
"step": 154
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 0.03155055722355096,
|
| 1123 |
+
"grad_norm": 0.4514567255973816,
|
| 1124 |
+
"learning_rate": 1.3213804466343421e-05,
|
| 1125 |
+
"loss": 2.616,
|
| 1126 |
+
"step": 155
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 0.03175410920563839,
|
| 1130 |
+
"grad_norm": 0.4534989893436432,
|
| 1131 |
+
"learning_rate": 1.2658926150792322e-05,
|
| 1132 |
+
"loss": 2.6276,
|
| 1133 |
+
"step": 156
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 0.03195766118772581,
|
| 1137 |
+
"grad_norm": 0.4631200134754181,
|
| 1138 |
+
"learning_rate": 1.2114256511983274e-05,
|
| 1139 |
+
"loss": 2.4606,
|
| 1140 |
+
"step": 157
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 0.03216121316981324,
|
| 1144 |
+
"grad_norm": 0.4224139451980591,
|
| 1145 |
+
"learning_rate": 1.157994445715706e-05,
|
| 1146 |
+
"loss": 2.5161,
|
| 1147 |
+
"step": 158
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 0.03236476515190067,
|
| 1151 |
+
"grad_norm": 0.4256625473499298,
|
| 1152 |
+
"learning_rate": 1.1056136061894384e-05,
|
| 1153 |
+
"loss": 2.5777,
|
| 1154 |
+
"step": 159
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.03256831713398809,
|
| 1158 |
+
"grad_norm": 0.41553547978401184,
|
| 1159 |
+
"learning_rate": 1.0542974530180327e-05,
|
| 1160 |
+
"loss": 2.4854,
|
| 1161 |
+
"step": 160
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.03277186911607552,
|
| 1165 |
+
"grad_norm": 0.42502570152282715,
|
| 1166 |
+
"learning_rate": 1.0040600155253765e-05,
|
| 1167 |
+
"loss": 2.4952,
|
| 1168 |
+
"step": 161
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 0.03297542109816294,
|
| 1172 |
+
"grad_norm": 0.4329552948474884,
|
| 1173 |
+
"learning_rate": 9.549150281252633e-06,
|
| 1174 |
+
"loss": 2.5339,
|
| 1175 |
+
"step": 162
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 0.03317897308025037,
|
| 1179 |
+
"grad_norm": 0.39022713899612427,
|
| 1180 |
+
"learning_rate": 9.068759265665384e-06,
|
| 1181 |
+
"loss": 2.4271,
|
| 1182 |
+
"step": 163
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 0.0333825250623378,
|
| 1186 |
+
"grad_norm": 0.4360031485557556,
|
| 1187 |
+
"learning_rate": 8.599558442598998e-06,
|
| 1188 |
+
"loss": 2.5676,
|
| 1189 |
+
"step": 164
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 0.03358607704442522,
|
| 1193 |
+
"grad_norm": 0.39230039715766907,
|
| 1194 |
+
"learning_rate": 8.141676086873572e-06,
|
| 1195 |
+
"loss": 2.5392,
|
| 1196 |
+
"step": 165
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.03378962902651265,
|
| 1200 |
+
"grad_norm": 0.4028782844543457,
|
| 1201 |
+
"learning_rate": 7.695237378953223e-06,
|
| 1202 |
+
"loss": 2.5232,
|
| 1203 |
+
"step": 166
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 0.03399318100860007,
|
| 1207 |
+
"grad_norm": 0.3945537209510803,
|
| 1208 |
+
"learning_rate": 7.260364370723044e-06,
|
| 1209 |
+
"loss": 2.4494,
|
| 1210 |
+
"step": 167
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 0.0341967329906875,
|
| 1214 |
+
"grad_norm": 0.4047512412071228,
|
| 1215 |
+
"learning_rate": 6.837175952121306e-06,
|
| 1216 |
+
"loss": 2.5328,
|
| 1217 |
+
"step": 168
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 0.03440028497277492,
|
| 1221 |
+
"grad_norm": 0.3825782835483551,
|
| 1222 |
+
"learning_rate": 6.425787818636131e-06,
|
| 1223 |
+
"loss": 2.4677,
|
| 1224 |
+
"step": 169
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 0.03460383695486235,
|
| 1228 |
+
"grad_norm": 0.40704014897346497,
|
| 1229 |
+
"learning_rate": 6.026312439675552e-06,
|
| 1230 |
+
"loss": 2.423,
|
| 1231 |
+
"step": 170
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 0.034807388936949776,
|
| 1235 |
+
"grad_norm": 0.39246898889541626,
|
| 1236 |
+
"learning_rate": 5.6388590278194096e-06,
|
| 1237 |
+
"loss": 2.6113,
|
| 1238 |
+
"step": 171
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.0350109409190372,
|
| 1242 |
+
"grad_norm": 0.40358856320381165,
|
| 1243 |
+
"learning_rate": 5.263533508961827e-06,
|
| 1244 |
+
"loss": 2.4924,
|
| 1245 |
+
"step": 172
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 0.035214492901124626,
|
| 1249 |
+
"grad_norm": 0.4041445851325989,
|
| 1250 |
+
"learning_rate": 4.900438493352055e-06,
|
| 1251 |
+
"loss": 2.4808,
|
| 1252 |
+
"step": 173
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 0.03541804488321205,
|
| 1256 |
+
"grad_norm": 0.39526063203811646,
|
| 1257 |
+
"learning_rate": 4.549673247541875e-06,
|
| 1258 |
+
"loss": 2.5216,
|
| 1259 |
+
"step": 174
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 0.035621596865299476,
|
| 1263 |
+
"grad_norm": 0.4017317593097687,
|
| 1264 |
+
"learning_rate": 4.2113336672471245e-06,
|
| 1265 |
+
"loss": 2.3518,
|
| 1266 |
+
"step": 175
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 0.035825148847386905,
|
| 1270 |
+
"grad_norm": 0.39643821120262146,
|
| 1271 |
+
"learning_rate": 3.885512251130763e-06,
|
| 1272 |
+
"loss": 2.5251,
|
| 1273 |
+
"step": 176
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 0.036028700829474326,
|
| 1277 |
+
"grad_norm": 0.43091297149658203,
|
| 1278 |
+
"learning_rate": 3.5722980755146517e-06,
|
| 1279 |
+
"loss": 2.4939,
|
| 1280 |
+
"step": 177
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 0.036232252811561755,
|
| 1284 |
+
"grad_norm": 0.40863290429115295,
|
| 1285 |
+
"learning_rate": 3.271776770026963e-06,
|
| 1286 |
+
"loss": 2.335,
|
| 1287 |
+
"step": 178
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 0.036435804793649176,
|
| 1291 |
+
"grad_norm": 0.43087130784988403,
|
| 1292 |
+
"learning_rate": 2.9840304941919415e-06,
|
| 1293 |
+
"loss": 2.3948,
|
| 1294 |
+
"step": 179
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 0.036639356775736605,
|
| 1298 |
+
"grad_norm": 0.44028177857398987,
|
| 1299 |
+
"learning_rate": 2.7091379149682685e-06,
|
| 1300 |
+
"loss": 2.4948,
|
| 1301 |
+
"step": 180
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 0.036842908757824026,
|
| 1305 |
+
"grad_norm": 0.43982651829719543,
|
| 1306 |
+
"learning_rate": 2.4471741852423237e-06,
|
| 1307 |
+
"loss": 2.414,
|
| 1308 |
+
"step": 181
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 0.037046460739911455,
|
| 1312 |
+
"grad_norm": 0.41899624466896057,
|
| 1313 |
+
"learning_rate": 2.1982109232821178e-06,
|
| 1314 |
+
"loss": 2.1921,
|
| 1315 |
+
"step": 182
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 0.03725001272199888,
|
| 1319 |
+
"grad_norm": 0.49341341853141785,
|
| 1320 |
+
"learning_rate": 1.962316193157593e-06,
|
| 1321 |
+
"loss": 2.3161,
|
| 1322 |
+
"step": 183
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 0.037453564704086305,
|
| 1326 |
+
"grad_norm": 0.433349072933197,
|
| 1327 |
+
"learning_rate": 1.7395544861325718e-06,
|
| 1328 |
+
"loss": 2.2032,
|
| 1329 |
+
"step": 184
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 0.03765711668617373,
|
| 1333 |
+
"grad_norm": 0.46125951409339905,
|
| 1334 |
+
"learning_rate": 1.5299867030334814e-06,
|
| 1335 |
+
"loss": 2.3619,
|
| 1336 |
+
"step": 185
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 0.037860668668261155,
|
| 1340 |
+
"grad_norm": 0.4640481770038605,
|
| 1341 |
+
"learning_rate": 1.333670137599713e-06,
|
| 1342 |
+
"loss": 2.1362,
|
| 1343 |
+
"step": 186
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 0.03806422065034858,
|
| 1347 |
+
"grad_norm": 0.5906166434288025,
|
| 1348 |
+
"learning_rate": 1.1506584608200367e-06,
|
| 1349 |
+
"loss": 2.0228,
|
| 1350 |
+
"step": 187
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 0.03826777263243601,
|
| 1354 |
+
"grad_norm": 0.5172399282455444,
|
| 1355 |
+
"learning_rate": 9.810017062595322e-07,
|
| 1356 |
+
"loss": 2.2062,
|
| 1357 |
+
"step": 188
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 0.03847132461452343,
|
| 1361 |
+
"grad_norm": 0.5385973453521729,
|
| 1362 |
+
"learning_rate": 8.247462563808817e-07,
|
| 1363 |
+
"loss": 1.9455,
|
| 1364 |
+
"step": 189
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 0.03867487659661086,
|
| 1368 |
+
"grad_norm": 0.5821766257286072,
|
| 1369 |
+
"learning_rate": 6.819348298638839e-07,
|
| 1370 |
+
"loss": 2.0129,
|
| 1371 |
+
"step": 190
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 0.03887842857869828,
|
| 1375 |
+
"grad_norm": 0.5611380338668823,
|
| 1376 |
+
"learning_rate": 5.526064699265753e-07,
|
| 1377 |
+
"loss": 1.851,
|
| 1378 |
+
"step": 191
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 0.03908198056078571,
|
| 1382 |
+
"grad_norm": 0.5985032320022583,
|
| 1383 |
+
"learning_rate": 4.367965336512403e-07,
|
| 1384 |
+
"loss": 2.0951,
|
| 1385 |
+
"step": 192
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 0.03928553254287313,
|
| 1389 |
+
"grad_norm": 0.5864668488502502,
|
| 1390 |
+
"learning_rate": 3.3453668231809286e-07,
|
| 1391 |
+
"loss": 1.7182,
|
| 1392 |
+
"step": 193
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 0.03948908452496056,
|
| 1396 |
+
"grad_norm": 0.6117426156997681,
|
| 1397 |
+
"learning_rate": 2.458548727494292e-07,
|
| 1398 |
+
"loss": 1.804,
|
| 1399 |
+
"step": 194
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 0.03969263650704799,
|
| 1403 |
+
"grad_norm": 0.5681461691856384,
|
| 1404 |
+
"learning_rate": 1.7077534966650766e-07,
|
| 1405 |
+
"loss": 1.6146,
|
| 1406 |
+
"step": 195
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 0.03989618848913541,
|
| 1410 |
+
"grad_norm": 0.5706256628036499,
|
| 1411 |
+
"learning_rate": 1.0931863906127327e-07,
|
| 1412 |
+
"loss": 1.587,
|
| 1413 |
+
"step": 196
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 0.04009974047122284,
|
| 1417 |
+
"grad_norm": 0.5642743110656738,
|
| 1418 |
+
"learning_rate": 6.150154258476315e-08,
|
| 1419 |
+
"loss": 1.5379,
|
| 1420 |
+
"step": 197
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 0.04030329245331026,
|
| 1424 |
+
"grad_norm": 0.618344783782959,
|
| 1425 |
+
"learning_rate": 2.7337132953697554e-08,
|
| 1426 |
+
"loss": 1.626,
|
| 1427 |
+
"step": 198
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 0.04050684443539769,
|
| 1431 |
+
"grad_norm": 0.7449570298194885,
|
| 1432 |
+
"learning_rate": 6.834750376549792e-09,
|
| 1433 |
+
"loss": 1.7495,
|
| 1434 |
+
"step": 199
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 0.04071039641748511,
|
| 1438 |
+
"grad_norm": 0.7824847102165222,
|
| 1439 |
+
"learning_rate": 0.0,
|
| 1440 |
+
"loss": 1.0957,
|
| 1441 |
+
"step": 200
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 0.04071039641748511,
|
| 1445 |
+
"eval_loss": 2.242037057876587,
|
| 1446 |
+
"eval_runtime": 139.1251,
|
| 1447 |
+
"eval_samples_per_second": 59.479,
|
| 1448 |
+
"eval_steps_per_second": 14.872,
|
| 1449 |
+
"step": 200
|
| 1450 |
}
|
| 1451 |
],
|
| 1452 |
"logging_steps": 1,
|
|
|
|
| 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": 1.3737759080448e+16,
|
| 1479 |
"train_batch_size": 8,
|
| 1480 |
"trial_name": null,
|
| 1481 |
"trial_params": null
|