Training in progress, step 2000, 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 295488936
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dd91de723bbd7ea7b7dfe87942ef4a89726bd5bdcfdd6abb72301f7a8513b562
|
| 3 |
size 295488936
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 591203178
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8e3bb5fe5d5fa250dea8b06ba61d861c69e2a49b2e59bd1f61d15f4b5e735dbc
|
| 3 |
size 591203178
|
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 sha256:2ec28ea0c416565eeac14a0e9c944f185ac250f4ed4bd15c84ff77ed78ba9301
|
| 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 sha256:2285ab56c032a195010949adc0c02d4df3cdbb2c2798776a12286c09cf74f9be
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 1.
|
| 5 |
"eval_steps": 100,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1177,6 +1177,396 @@
|
|
| 1177 |
"eval_news_finetune_val_samples_per_second": 1.396,
|
| 1178 |
"eval_news_finetune_val_steps_per_second": 1.396,
|
| 1179 |
"step": 1500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1180 |
}
|
| 1181 |
],
|
| 1182 |
"logging_steps": 10,
|
|
@@ -1196,7 +1586,7 @@
|
|
| 1196 |
"attributes": {}
|
| 1197 |
}
|
| 1198 |
},
|
| 1199 |
-
"total_flos":
|
| 1200 |
"train_batch_size": 1,
|
| 1201 |
"trial_name": null,
|
| 1202 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.4287245444801715,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 2000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1177 |
"eval_news_finetune_val_samples_per_second": 1.396,
|
| 1178 |
"eval_news_finetune_val_steps_per_second": 1.396,
|
| 1179 |
"step": 1500
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 1.0785994998213648,
|
| 1183 |
+
"grad_norm": 1.7663507461547852,
|
| 1184 |
+
"learning_rate": 8.082020186215156e-05,
|
| 1185 |
+
"loss": 0.2407,
|
| 1186 |
+
"step": 1510
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 1.0857449088960343,
|
| 1190 |
+
"grad_norm": 1.2081632614135742,
|
| 1191 |
+
"learning_rate": 8.049165944562316e-05,
|
| 1192 |
+
"loss": 0.2483,
|
| 1193 |
+
"step": 1520
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 1.092890317970704,
|
| 1197 |
+
"grad_norm": 0.5045826435089111,
|
| 1198 |
+
"learning_rate": 8.016100750576621e-05,
|
| 1199 |
+
"loss": 0.2013,
|
| 1200 |
+
"step": 1530
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.1000357270453733,
|
| 1204 |
+
"grad_norm": 1.4456278085708618,
|
| 1205 |
+
"learning_rate": 7.98282689182783e-05,
|
| 1206 |
+
"loss": 0.2034,
|
| 1207 |
+
"step": 1540
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.107181136120043,
|
| 1211 |
+
"grad_norm": 1.1558668613433838,
|
| 1212 |
+
"learning_rate": 7.949346670321891e-05,
|
| 1213 |
+
"loss": 0.2386,
|
| 1214 |
+
"step": 1550
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 1.1143265451947124,
|
| 1218 |
+
"grad_norm": 1.4196126461029053,
|
| 1219 |
+
"learning_rate": 7.915662402341664e-05,
|
| 1220 |
+
"loss": 0.2299,
|
| 1221 |
+
"step": 1560
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 1.1214719542693818,
|
| 1225 |
+
"grad_norm": 0.9341222047805786,
|
| 1226 |
+
"learning_rate": 7.88177641828669e-05,
|
| 1227 |
+
"loss": 0.2105,
|
| 1228 |
+
"step": 1570
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 1.1286173633440515,
|
| 1232 |
+
"grad_norm": 1.066001296043396,
|
| 1233 |
+
"learning_rate": 7.847691062511957e-05,
|
| 1234 |
+
"loss": 0.1925,
|
| 1235 |
+
"step": 1580
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 1.135762772418721,
|
| 1239 |
+
"grad_norm": 0.7840182781219482,
|
| 1240 |
+
"learning_rate": 7.813408693165704e-05,
|
| 1241 |
+
"loss": 0.2425,
|
| 1242 |
+
"step": 1590
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 1.1429081814933906,
|
| 1246 |
+
"grad_norm": 0.983668327331543,
|
| 1247 |
+
"learning_rate": 7.778931682026293e-05,
|
| 1248 |
+
"loss": 0.2014,
|
| 1249 |
+
"step": 1600
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.1429081814933906,
|
| 1253 |
+
"eval_news_finetune_val_loss": 0.29564452171325684,
|
| 1254 |
+
"eval_news_finetune_val_runtime": 1003.001,
|
| 1255 |
+
"eval_news_finetune_val_samples_per_second": 1.396,
|
| 1256 |
+
"eval_news_finetune_val_steps_per_second": 1.396,
|
| 1257 |
+
"step": 1600
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 1.15005359056806,
|
| 1261 |
+
"grad_norm": 1.63984215259552,
|
| 1262 |
+
"learning_rate": 7.744262414338099e-05,
|
| 1263 |
+
"loss": 0.2863,
|
| 1264 |
+
"step": 1610
|
| 1265 |
+
},
|
| 1266 |
+
{
|
| 1267 |
+
"epoch": 1.1571989996427297,
|
| 1268 |
+
"grad_norm": 0.9211621284484863,
|
| 1269 |
+
"learning_rate": 7.709403288646507e-05,
|
| 1270 |
+
"loss": 0.2175,
|
| 1271 |
+
"step": 1620
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 1.164344408717399,
|
| 1275 |
+
"grad_norm": 1.3369996547698975,
|
| 1276 |
+
"learning_rate": 7.67435671663196e-05,
|
| 1277 |
+
"loss": 0.1893,
|
| 1278 |
+
"step": 1630
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 1.1714898177920685,
|
| 1282 |
+
"grad_norm": 0.7532891631126404,
|
| 1283 |
+
"learning_rate": 7.63912512294312e-05,
|
| 1284 |
+
"loss": 0.2483,
|
| 1285 |
+
"step": 1640
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 1.1786352268667382,
|
| 1289 |
+
"grad_norm": 1.0959442853927612,
|
| 1290 |
+
"learning_rate": 7.603710945029119e-05,
|
| 1291 |
+
"loss": 0.1888,
|
| 1292 |
+
"step": 1650
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 1.1857806359414076,
|
| 1296 |
+
"grad_norm": 0.9019472599029541,
|
| 1297 |
+
"learning_rate": 7.568116632970922e-05,
|
| 1298 |
+
"loss": 0.2144,
|
| 1299 |
+
"step": 1660
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 1.1929260450160772,
|
| 1303 |
+
"grad_norm": 1.1219818592071533,
|
| 1304 |
+
"learning_rate": 7.532344649311829e-05,
|
| 1305 |
+
"loss": 0.191,
|
| 1306 |
+
"step": 1670
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 1.2000714540907467,
|
| 1310 |
+
"grad_norm": 1.0829100608825684,
|
| 1311 |
+
"learning_rate": 7.496397468887106e-05,
|
| 1312 |
+
"loss": 0.2762,
|
| 1313 |
+
"step": 1680
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 1.2072168631654163,
|
| 1317 |
+
"grad_norm": 0.7855832576751709,
|
| 1318 |
+
"learning_rate": 7.460277578652759e-05,
|
| 1319 |
+
"loss": 0.157,
|
| 1320 |
+
"step": 1690
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 1.2143622722400857,
|
| 1324 |
+
"grad_norm": 2.407999038696289,
|
| 1325 |
+
"learning_rate": 7.423987477513488e-05,
|
| 1326 |
+
"loss": 0.2627,
|
| 1327 |
+
"step": 1700
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 1.2143622722400857,
|
| 1331 |
+
"eval_news_finetune_val_loss": 0.28248873353004456,
|
| 1332 |
+
"eval_news_finetune_val_runtime": 1003.1081,
|
| 1333 |
+
"eval_news_finetune_val_samples_per_second": 1.396,
|
| 1334 |
+
"eval_news_finetune_val_steps_per_second": 1.396,
|
| 1335 |
+
"step": 1700
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 1.2215076813147552,
|
| 1339 |
+
"grad_norm": 1.5500895977020264,
|
| 1340 |
+
"learning_rate": 7.387529676149799e-05,
|
| 1341 |
+
"loss": 0.1477,
|
| 1342 |
+
"step": 1710
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"epoch": 1.2286530903894248,
|
| 1346 |
+
"grad_norm": 1.5599130392074585,
|
| 1347 |
+
"learning_rate": 7.350906696844307e-05,
|
| 1348 |
+
"loss": 0.1942,
|
| 1349 |
+
"step": 1720
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 1.2357984994640943,
|
| 1353 |
+
"grad_norm": 1.6327091455459595,
|
| 1354 |
+
"learning_rate": 7.314121073307229e-05,
|
| 1355 |
+
"loss": 0.2,
|
| 1356 |
+
"step": 1730
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 1.242943908538764,
|
| 1360 |
+
"grad_norm": 0.6044666767120361,
|
| 1361 |
+
"learning_rate": 7.277175350501111e-05,
|
| 1362 |
+
"loss": 0.185,
|
| 1363 |
+
"step": 1740
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 1.2500893176134333,
|
| 1367 |
+
"grad_norm": 1.317089319229126,
|
| 1368 |
+
"learning_rate": 7.240072084464729e-05,
|
| 1369 |
+
"loss": 0.196,
|
| 1370 |
+
"step": 1750
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 1.257234726688103,
|
| 1374 |
+
"grad_norm": 1.089105486869812,
|
| 1375 |
+
"learning_rate": 7.202813842136283e-05,
|
| 1376 |
+
"loss": 0.1322,
|
| 1377 |
+
"step": 1760
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 1.2643801357627724,
|
| 1381 |
+
"grad_norm": 1.4972888231277466,
|
| 1382 |
+
"learning_rate": 7.165403201175787e-05,
|
| 1383 |
+
"loss": 0.2176,
|
| 1384 |
+
"step": 1770
|
| 1385 |
+
},
|
| 1386 |
+
{
|
| 1387 |
+
"epoch": 1.2715255448374418,
|
| 1388 |
+
"grad_norm": 1.4998830556869507,
|
| 1389 |
+
"learning_rate": 7.127842749786747e-05,
|
| 1390 |
+
"loss": 0.218,
|
| 1391 |
+
"step": 1780
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 1.2786709539121115,
|
| 1395 |
+
"grad_norm": 0.9759517908096313,
|
| 1396 |
+
"learning_rate": 7.090135086537095e-05,
|
| 1397 |
+
"loss": 0.1653,
|
| 1398 |
+
"step": 1790
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 1.285816362986781,
|
| 1402 |
+
"grad_norm": 0.9713583588600159,
|
| 1403 |
+
"learning_rate": 7.052282820179412e-05,
|
| 1404 |
+
"loss": 0.175,
|
| 1405 |
+
"step": 1800
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 1.285816362986781,
|
| 1409 |
+
"eval_news_finetune_val_loss": 0.2936909794807434,
|
| 1410 |
+
"eval_news_finetune_val_runtime": 1003.12,
|
| 1411 |
+
"eval_news_finetune_val_samples_per_second": 1.396,
|
| 1412 |
+
"eval_news_finetune_val_steps_per_second": 1.396,
|
| 1413 |
+
"step": 1800
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 1.2929617720614506,
|
| 1417 |
+
"grad_norm": 0.6328814625740051,
|
| 1418 |
+
"learning_rate": 7.014288569470446e-05,
|
| 1419 |
+
"loss": 0.1727,
|
| 1420 |
+
"step": 1810
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 1.30010718113612,
|
| 1424 |
+
"grad_norm": 1.622104525566101,
|
| 1425 |
+
"learning_rate": 6.976154962989934e-05,
|
| 1426 |
+
"loss": 0.2363,
|
| 1427 |
+
"step": 1820
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 1.3072525902107897,
|
| 1431 |
+
"grad_norm": 1.8254674673080444,
|
| 1432 |
+
"learning_rate": 6.937884638958757e-05,
|
| 1433 |
+
"loss": 0.1897,
|
| 1434 |
+
"step": 1830
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 1.314397999285459,
|
| 1438 |
+
"grad_norm": 0.8813793063163757,
|
| 1439 |
+
"learning_rate": 6.899480245056396e-05,
|
| 1440 |
+
"loss": 0.2029,
|
| 1441 |
+
"step": 1840
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 1.3215434083601285,
|
| 1445 |
+
"grad_norm": 0.7675999999046326,
|
| 1446 |
+
"learning_rate": 6.860944438237788e-05,
|
| 1447 |
+
"loss": 0.2025,
|
| 1448 |
+
"step": 1850
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 1.3286888174347982,
|
| 1452 |
+
"grad_norm": 1.1973013877868652,
|
| 1453 |
+
"learning_rate": 6.82227988454948e-05,
|
| 1454 |
+
"loss": 0.2317,
|
| 1455 |
+
"step": 1860
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 1.3358342265094676,
|
| 1459 |
+
"grad_norm": 0.7864009737968445,
|
| 1460 |
+
"learning_rate": 6.783489258945195e-05,
|
| 1461 |
+
"loss": 0.2318,
|
| 1462 |
+
"step": 1870
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 1.3429796355841372,
|
| 1466 |
+
"grad_norm": 1.0866330862045288,
|
| 1467 |
+
"learning_rate": 6.74457524510077e-05,
|
| 1468 |
+
"loss": 0.1871,
|
| 1469 |
+
"step": 1880
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 1.3501250446588067,
|
| 1473 |
+
"grad_norm": 0.8745126724243164,
|
| 1474 |
+
"learning_rate": 6.705540535228485e-05,
|
| 1475 |
+
"loss": 0.211,
|
| 1476 |
+
"step": 1890
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 1.3572704537334763,
|
| 1480 |
+
"grad_norm": 1.3401581048965454,
|
| 1481 |
+
"learning_rate": 6.66638782989081e-05,
|
| 1482 |
+
"loss": 0.2307,
|
| 1483 |
+
"step": 1900
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 1.3572704537334763,
|
| 1487 |
+
"eval_news_finetune_val_loss": 0.2787444591522217,
|
| 1488 |
+
"eval_news_finetune_val_runtime": 1002.9344,
|
| 1489 |
+
"eval_news_finetune_val_samples_per_second": 1.396,
|
| 1490 |
+
"eval_news_finetune_val_steps_per_second": 1.396,
|
| 1491 |
+
"step": 1900
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 1.3644158628081458,
|
| 1495 |
+
"grad_norm": 0.6149284839630127,
|
| 1496 |
+
"learning_rate": 6.627119837813564e-05,
|
| 1497 |
+
"loss": 0.2128,
|
| 1498 |
+
"step": 1910
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 1.3715612718828152,
|
| 1502 |
+
"grad_norm": 1.7847625017166138,
|
| 1503 |
+
"learning_rate": 6.587739275698525e-05,
|
| 1504 |
+
"loss": 0.1551,
|
| 1505 |
+
"step": 1920
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 1.3787066809574848,
|
| 1509 |
+
"grad_norm": 1.1973716020584106,
|
| 1510 |
+
"learning_rate": 6.54824886803547e-05,
|
| 1511 |
+
"loss": 0.2335,
|
| 1512 |
+
"step": 1930
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 1.3858520900321543,
|
| 1516 |
+
"grad_norm": 1.5757859945297241,
|
| 1517 |
+
"learning_rate": 6.508651346913687e-05,
|
| 1518 |
+
"loss": 0.1504,
|
| 1519 |
+
"step": 1940
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 1.392997499106824,
|
| 1523 |
+
"grad_norm": 1.7269341945648193,
|
| 1524 |
+
"learning_rate": 6.468949451832968e-05,
|
| 1525 |
+
"loss": 0.2679,
|
| 1526 |
+
"step": 1950
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 1.4001429081814933,
|
| 1530 |
+
"grad_norm": 1.6860129833221436,
|
| 1531 |
+
"learning_rate": 6.429145929514063e-05,
|
| 1532 |
+
"loss": 0.1942,
|
| 1533 |
+
"step": 1960
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 1.407288317256163,
|
| 1537 |
+
"grad_norm": 1.1732631921768188,
|
| 1538 |
+
"learning_rate": 6.389243533708671e-05,
|
| 1539 |
+
"loss": 0.2025,
|
| 1540 |
+
"step": 1970
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 1.4144337263308324,
|
| 1544 |
+
"grad_norm": 0.9073033332824707,
|
| 1545 |
+
"learning_rate": 6.349245025008912e-05,
|
| 1546 |
+
"loss": 0.1836,
|
| 1547 |
+
"step": 1980
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 1.4215791354055018,
|
| 1551 |
+
"grad_norm": 1.133843183517456,
|
| 1552 |
+
"learning_rate": 6.309153170656342e-05,
|
| 1553 |
+
"loss": 0.1526,
|
| 1554 |
+
"step": 1990
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 1.4287245444801715,
|
| 1558 |
+
"grad_norm": 2.656296968460083,
|
| 1559 |
+
"learning_rate": 6.268970744350515e-05,
|
| 1560 |
+
"loss": 0.1939,
|
| 1561 |
+
"step": 2000
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 1.4287245444801715,
|
| 1565 |
+
"eval_news_finetune_val_loss": 0.27414408326148987,
|
| 1566 |
+
"eval_news_finetune_val_runtime": 1003.0949,
|
| 1567 |
+
"eval_news_finetune_val_samples_per_second": 1.396,
|
| 1568 |
+
"eval_news_finetune_val_steps_per_second": 1.396,
|
| 1569 |
+
"step": 2000
|
| 1570 |
}
|
| 1571 |
],
|
| 1572 |
"logging_steps": 10,
|
|
|
|
| 1586 |
"attributes": {}
|
| 1587 |
}
|
| 1588 |
},
|
| 1589 |
+
"total_flos": 5.538125336973312e+16,
|
| 1590 |
"train_batch_size": 1,
|
| 1591 |
"trial_name": null,
|
| 1592 |
"trial_params": null
|