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 871457448
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b6285a33ebda437688ad29cd8f7e7a0a8915f402bc03986dda38a056a9cb46a6
|
| 3 |
size 871457448
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 443280340
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6f0ea8696cd3edf4e39ecc3cd9695d87f4cea364b87824da66c54dfcdcb11d36
|
| 3 |
size 443280340
|
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 sha256:3e5a201eb618587b99b3c99ba89ac3c2bd3332adfe0182f6b71d70e4afb55c99
|
| 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 sha256:4ac116b8169c53ab649a7f15f2f32735f2c71ec2f803f70de8c655a513ee9cfc
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 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": 7.047,
|
| 1090 |
"eval_steps_per_second": 1.763,
|
| 1091 |
"step": 150
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1092 |
}
|
| 1093 |
],
|
| 1094 |
"logging_steps": 1,
|
|
@@ -1117,7 +1475,7 @@
|
|
| 1117 |
"attributes": {}
|
| 1118 |
}
|
| 1119 |
},
|
| 1120 |
-
"total_flos":
|
| 1121 |
"train_batch_size": 8,
|
| 1122 |
"trial_name": null,
|
| 1123 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.3369649350643158,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 4 |
+
"epoch": 0.1642373229316362,
|
| 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": 7.047,
|
| 1090 |
"eval_steps_per_second": 1.763,
|
| 1091 |
"step": 150
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 0.12399917881338535,
|
| 1095 |
+
"grad_norm": 1.2095229625701904,
|
| 1096 |
+
"learning_rate": 7.107005538862646e-05,
|
| 1097 |
+
"loss": 0.4237,
|
| 1098 |
+
"step": 151
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 0.12482036542804352,
|
| 1102 |
+
"grad_norm": 1.172702670097351,
|
| 1103 |
+
"learning_rate": 7.07041155014006e-05,
|
| 1104 |
+
"loss": 0.351,
|
| 1105 |
+
"step": 152
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 0.1256415520427017,
|
| 1109 |
+
"grad_norm": 1.0826219320297241,
|
| 1110 |
+
"learning_rate": 7.033683215379002e-05,
|
| 1111 |
+
"loss": 0.3055,
|
| 1112 |
+
"step": 153
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.12646273865735988,
|
| 1116 |
+
"grad_norm": 0.3956534266471863,
|
| 1117 |
+
"learning_rate": 6.996822917828477e-05,
|
| 1118 |
+
"loss": 0.2814,
|
| 1119 |
+
"step": 154
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 0.12728392527201807,
|
| 1123 |
+
"grad_norm": 0.26742252707481384,
|
| 1124 |
+
"learning_rate": 6.959833049300377e-05,
|
| 1125 |
+
"loss": 0.2178,
|
| 1126 |
+
"step": 155
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 0.12810511188667625,
|
| 1130 |
+
"grad_norm": 0.2690823972225189,
|
| 1131 |
+
"learning_rate": 6.922716010014255e-05,
|
| 1132 |
+
"loss": 0.2396,
|
| 1133 |
+
"step": 156
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 0.12892629850133444,
|
| 1137 |
+
"grad_norm": 0.3007110357284546,
|
| 1138 |
+
"learning_rate": 6.885474208441603e-05,
|
| 1139 |
+
"loss": 0.2573,
|
| 1140 |
+
"step": 157
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 0.1297474851159926,
|
| 1144 |
+
"grad_norm": 0.27441394329071045,
|
| 1145 |
+
"learning_rate": 6.848110061149556e-05,
|
| 1146 |
+
"loss": 0.2418,
|
| 1147 |
+
"step": 158
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 0.1305686717306508,
|
| 1151 |
+
"grad_norm": 0.3037412166595459,
|
| 1152 |
+
"learning_rate": 6.810625992644085e-05,
|
| 1153 |
+
"loss": 0.2762,
|
| 1154 |
+
"step": 159
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.13138985834530897,
|
| 1158 |
+
"grad_norm": 0.31967610120773315,
|
| 1159 |
+
"learning_rate": 6.773024435212678e-05,
|
| 1160 |
+
"loss": 0.2674,
|
| 1161 |
+
"step": 160
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.13221104495996716,
|
| 1165 |
+
"grad_norm": 0.2789151668548584,
|
| 1166 |
+
"learning_rate": 6.735307828766515e-05,
|
| 1167 |
+
"loss": 0.254,
|
| 1168 |
+
"step": 161
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 0.13303223157462532,
|
| 1172 |
+
"grad_norm": 0.3466937839984894,
|
| 1173 |
+
"learning_rate": 6.697478620682137e-05,
|
| 1174 |
+
"loss": 0.2806,
|
| 1175 |
+
"step": 162
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 0.1338534181892835,
|
| 1179 |
+
"grad_norm": 0.31567007303237915,
|
| 1180 |
+
"learning_rate": 6.659539265642643e-05,
|
| 1181 |
+
"loss": 0.2611,
|
| 1182 |
+
"step": 163
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 0.1346746048039417,
|
| 1186 |
+
"grad_norm": 0.3090943992137909,
|
| 1187 |
+
"learning_rate": 6.621492225478414e-05,
|
| 1188 |
+
"loss": 0.2634,
|
| 1189 |
+
"step": 164
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 0.13549579141859988,
|
| 1193 |
+
"grad_norm": 0.30518144369125366,
|
| 1194 |
+
"learning_rate": 6.583339969007363e-05,
|
| 1195 |
+
"loss": 0.2314,
|
| 1196 |
+
"step": 165
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.13631697803325807,
|
| 1200 |
+
"grad_norm": 0.32459190487861633,
|
| 1201 |
+
"learning_rate": 6.545084971874738e-05,
|
| 1202 |
+
"loss": 0.2531,
|
| 1203 |
+
"step": 166
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 0.13713816464791623,
|
| 1207 |
+
"grad_norm": 0.3627863824367523,
|
| 1208 |
+
"learning_rate": 6.506729716392481e-05,
|
| 1209 |
+
"loss": 0.2555,
|
| 1210 |
+
"step": 167
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 0.13795935126257441,
|
| 1214 |
+
"grad_norm": 0.44058528542518616,
|
| 1215 |
+
"learning_rate": 6.468276691378155e-05,
|
| 1216 |
+
"loss": 0.2623,
|
| 1217 |
+
"step": 168
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 0.1387805378772326,
|
| 1221 |
+
"grad_norm": 0.43444401025772095,
|
| 1222 |
+
"learning_rate": 6.429728391993446e-05,
|
| 1223 |
+
"loss": 0.2796,
|
| 1224 |
+
"step": 169
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 0.1396017244918908,
|
| 1228 |
+
"grad_norm": 0.3341348469257355,
|
| 1229 |
+
"learning_rate": 6.391087319582264e-05,
|
| 1230 |
+
"loss": 0.2392,
|
| 1231 |
+
"step": 170
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 0.14042291110654898,
|
| 1235 |
+
"grad_norm": 0.339946448802948,
|
| 1236 |
+
"learning_rate": 6.35235598150842e-05,
|
| 1237 |
+
"loss": 0.2923,
|
| 1238 |
+
"step": 171
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.14124409772120713,
|
| 1242 |
+
"grad_norm": 0.308361291885376,
|
| 1243 |
+
"learning_rate": 6.313536890992935e-05,
|
| 1244 |
+
"loss": 0.2172,
|
| 1245 |
+
"step": 172
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 0.14206528433586532,
|
| 1249 |
+
"grad_norm": 0.36301669478416443,
|
| 1250 |
+
"learning_rate": 6.274632566950967e-05,
|
| 1251 |
+
"loss": 0.2436,
|
| 1252 |
+
"step": 173
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 0.1428864709505235,
|
| 1256 |
+
"grad_norm": 0.37350887060165405,
|
| 1257 |
+
"learning_rate": 6.235645533828349e-05,
|
| 1258 |
+
"loss": 0.2499,
|
| 1259 |
+
"step": 174
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 0.1437076575651817,
|
| 1263 |
+
"grad_norm": 0.33035343885421753,
|
| 1264 |
+
"learning_rate": 6.19657832143779e-05,
|
| 1265 |
+
"loss": 0.2233,
|
| 1266 |
+
"step": 175
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 0.14452884417983985,
|
| 1270 |
+
"grad_norm": 0.3460734486579895,
|
| 1271 |
+
"learning_rate": 6.157433464794716e-05,
|
| 1272 |
+
"loss": 0.238,
|
| 1273 |
+
"step": 176
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 0.14535003079449804,
|
| 1277 |
+
"grad_norm": 0.3977898359298706,
|
| 1278 |
+
"learning_rate": 6.118213503952779e-05,
|
| 1279 |
+
"loss": 0.2669,
|
| 1280 |
+
"step": 177
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 0.14617121740915623,
|
| 1284 |
+
"grad_norm": 0.442679226398468,
|
| 1285 |
+
"learning_rate": 6.078920983839031e-05,
|
| 1286 |
+
"loss": 0.2615,
|
| 1287 |
+
"step": 178
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 0.14699240402381442,
|
| 1291 |
+
"grad_norm": 0.3600601553916931,
|
| 1292 |
+
"learning_rate": 6.0395584540887963e-05,
|
| 1293 |
+
"loss": 0.2502,
|
| 1294 |
+
"step": 179
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 0.1478135906384726,
|
| 1298 |
+
"grad_norm": 0.3798171877861023,
|
| 1299 |
+
"learning_rate": 6.0001284688802226e-05,
|
| 1300 |
+
"loss": 0.2574,
|
| 1301 |
+
"step": 180
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 0.14863477725313076,
|
| 1305 |
+
"grad_norm": 0.3883853256702423,
|
| 1306 |
+
"learning_rate": 5.960633586768543e-05,
|
| 1307 |
+
"loss": 0.2288,
|
| 1308 |
+
"step": 181
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 0.14945596386778895,
|
| 1312 |
+
"grad_norm": 0.48229384422302246,
|
| 1313 |
+
"learning_rate": 5.921076370520058e-05,
|
| 1314 |
+
"loss": 0.313,
|
| 1315 |
+
"step": 182
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 0.15027715048244714,
|
| 1319 |
+
"grad_norm": 0.43548011779785156,
|
| 1320 |
+
"learning_rate": 5.8814593869458455e-05,
|
| 1321 |
+
"loss": 0.2908,
|
| 1322 |
+
"step": 183
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 0.15109833709710532,
|
| 1326 |
+
"grad_norm": 0.4758455455303192,
|
| 1327 |
+
"learning_rate": 5.841785206735192e-05,
|
| 1328 |
+
"loss": 0.2527,
|
| 1329 |
+
"step": 184
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 0.1519195237117635,
|
| 1333 |
+
"grad_norm": 0.38382488489151,
|
| 1334 |
+
"learning_rate": 5.8020564042888015e-05,
|
| 1335 |
+
"loss": 0.2506,
|
| 1336 |
+
"step": 185
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 0.15274071032642167,
|
| 1340 |
+
"grad_norm": 0.4354500472545624,
|
| 1341 |
+
"learning_rate": 5.762275557551727e-05,
|
| 1342 |
+
"loss": 0.2753,
|
| 1343 |
+
"step": 186
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 0.15356189694107986,
|
| 1347 |
+
"grad_norm": 0.4720785319805145,
|
| 1348 |
+
"learning_rate": 5.7224452478461064e-05,
|
| 1349 |
+
"loss": 0.2799,
|
| 1350 |
+
"step": 187
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 0.15438308355573804,
|
| 1354 |
+
"grad_norm": 0.44493159651756287,
|
| 1355 |
+
"learning_rate": 5.682568059703659e-05,
|
| 1356 |
+
"loss": 0.2565,
|
| 1357 |
+
"step": 188
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 0.15520427017039623,
|
| 1361 |
+
"grad_norm": 0.430894136428833,
|
| 1362 |
+
"learning_rate": 5.642646580697973e-05,
|
| 1363 |
+
"loss": 0.2542,
|
| 1364 |
+
"step": 189
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 0.15602545678505442,
|
| 1368 |
+
"grad_norm": 0.4223925769329071,
|
| 1369 |
+
"learning_rate": 5.602683401276615e-05,
|
| 1370 |
+
"loss": 0.2331,
|
| 1371 |
+
"step": 190
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 0.15684664339971258,
|
| 1375 |
+
"grad_norm": 0.44736889004707336,
|
| 1376 |
+
"learning_rate": 5.562681114593028e-05,
|
| 1377 |
+
"loss": 0.2476,
|
| 1378 |
+
"step": 191
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 0.15766783001437076,
|
| 1382 |
+
"grad_norm": 0.5344662666320801,
|
| 1383 |
+
"learning_rate": 5.522642316338268e-05,
|
| 1384 |
+
"loss": 0.3117,
|
| 1385 |
+
"step": 192
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 0.15848901662902895,
|
| 1389 |
+
"grad_norm": 0.4708305895328522,
|
| 1390 |
+
"learning_rate": 5.482569604572576e-05,
|
| 1391 |
+
"loss": 0.213,
|
| 1392 |
+
"step": 193
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 0.15931020324368714,
|
| 1396 |
+
"grad_norm": 0.5267696976661682,
|
| 1397 |
+
"learning_rate": 5.442465579556793e-05,
|
| 1398 |
+
"loss": 0.2536,
|
| 1399 |
+
"step": 194
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 0.1601313898583453,
|
| 1403 |
+
"grad_norm": 0.5681753754615784,
|
| 1404 |
+
"learning_rate": 5.402332843583631e-05,
|
| 1405 |
+
"loss": 0.3035,
|
| 1406 |
+
"step": 195
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 0.16095257647300348,
|
| 1410 |
+
"grad_norm": 0.6239408850669861,
|
| 1411 |
+
"learning_rate": 5.3621740008088126e-05,
|
| 1412 |
+
"loss": 0.3042,
|
| 1413 |
+
"step": 196
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 0.16177376308766167,
|
| 1417 |
+
"grad_norm": 0.5667141079902649,
|
| 1418 |
+
"learning_rate": 5.321991657082097e-05,
|
| 1419 |
+
"loss": 0.2375,
|
| 1420 |
+
"step": 197
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 0.16259494970231986,
|
| 1424 |
+
"grad_norm": 0.6225094795227051,
|
| 1425 |
+
"learning_rate": 5.281788419778187e-05,
|
| 1426 |
+
"loss": 0.2812,
|
| 1427 |
+
"step": 198
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 0.16341613631697804,
|
| 1431 |
+
"grad_norm": 0.8053914904594421,
|
| 1432 |
+
"learning_rate": 5.2415668976275355e-05,
|
| 1433 |
+
"loss": 0.2652,
|
| 1434 |
+
"step": 199
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 0.1642373229316362,
|
| 1438 |
+
"grad_norm": 0.7740500569343567,
|
| 1439 |
+
"learning_rate": 5.201329700547076e-05,
|
| 1440 |
+
"loss": 0.3169,
|
| 1441 |
+
"step": 200
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 0.1642373229316362,
|
| 1445 |
+
"eval_loss": 0.3369649350643158,
|
| 1446 |
+
"eval_runtime": 290.8356,
|
| 1447 |
+
"eval_samples_per_second": 7.052,
|
| 1448 |
+
"eval_steps_per_second": 1.764,
|
| 1449 |
+
"step": 200
|
| 1450 |
}
|
| 1451 |
],
|
| 1452 |
"logging_steps": 1,
|
|
|
|
| 1475 |
"attributes": {}
|
| 1476 |
}
|
| 1477 |
},
|
| 1478 |
+
"total_flos": 3.4760340805779456e+17,
|
| 1479 |
"train_batch_size": 8,
|
| 1480 |
"trial_name": null,
|
| 1481 |
"trial_params": null
|