End of training
Browse files- README.md +6 -4
- all_results.json +11 -11
- eval_results.json +7 -7
- runs/Jan16_13-07-53_vorace/events.out.tfevents.1705415718.vorace.502984.1 +3 -0
- train_results.json +5 -5
- trainer_state.json +1078 -94
README.md
CHANGED
|
@@ -2,6 +2,8 @@
|
|
| 2 |
license: mit
|
| 3 |
base_model: facebook/w2v-bert-2.0
|
| 4 |
tags:
|
|
|
|
|
|
|
| 5 |
- generated_from_trainer
|
| 6 |
metrics:
|
| 7 |
- wer
|
|
@@ -15,11 +17,11 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# wav2vec2-bert-CV16-en-libri
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [facebook/w2v-bert-2.0](https://huggingface.co/facebook/w2v-bert-2.0) on the
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
-
- Loss: 0.
|
| 21 |
-
- Wer: 0.
|
| 22 |
-
- Cer: 0.
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
|
|
|
| 2 |
license: mit
|
| 3 |
base_model: facebook/w2v-bert-2.0
|
| 4 |
tags:
|
| 5 |
+
- automatic-speech-recognition
|
| 6 |
+
- librispeech_asr
|
| 7 |
- generated_from_trainer
|
| 8 |
metrics:
|
| 9 |
- wer
|
|
|
|
| 17 |
|
| 18 |
# wav2vec2-bert-CV16-en-libri
|
| 19 |
|
| 20 |
+
This model is a fine-tuned version of [facebook/w2v-bert-2.0](https://huggingface.co/facebook/w2v-bert-2.0) on the LIBRISPEECH_ASR - CLEAN dataset.
|
| 21 |
It achieves the following results on the evaluation set:
|
| 22 |
+
- Loss: 0.1331
|
| 23 |
+
- Wer: 0.0997
|
| 24 |
+
- Cer: 0.0264
|
| 25 |
|
| 26 |
## Model description
|
| 27 |
|
all_results.json
CHANGED
|
@@ -1,15 +1,15 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_cer": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime": 59.
|
| 6 |
"eval_samples": 2528,
|
| 7 |
-
"eval_samples_per_second": 42.
|
| 8 |
-
"eval_steps_per_second": 1.
|
| 9 |
-
"eval_wer": 0.
|
| 10 |
-
"train_loss":
|
| 11 |
-
"train_runtime":
|
| 12 |
"train_samples": 28538,
|
| 13 |
-
"train_samples_per_second":
|
| 14 |
-
"train_steps_per_second": 0.
|
| 15 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_cer": 0.02643822760039138,
|
| 4 |
+
"eval_loss": 0.13311129808425903,
|
| 5 |
+
"eval_runtime": 59.8158,
|
| 6 |
"eval_samples": 2528,
|
| 7 |
+
"eval_samples_per_second": 42.263,
|
| 8 |
+
"eval_steps_per_second": 1.187,
|
| 9 |
+
"eval_wer": 0.09965613856342333,
|
| 10 |
+
"train_loss": 0.09581804365822763,
|
| 11 |
+
"train_runtime": 4900.8909,
|
| 12 |
"train_samples": 28538,
|
| 13 |
+
"train_samples_per_second": 29.115,
|
| 14 |
+
"train_steps_per_second": 0.404
|
| 15 |
}
|
eval_results.json
CHANGED
|
@@ -1,10 +1,10 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_cer": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime": 59.
|
| 6 |
"eval_samples": 2528,
|
| 7 |
-
"eval_samples_per_second": 42.
|
| 8 |
-
"eval_steps_per_second": 1.
|
| 9 |
-
"eval_wer": 0.
|
| 10 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_cer": 0.02643822760039138,
|
| 4 |
+
"eval_loss": 0.13311129808425903,
|
| 5 |
+
"eval_runtime": 59.8158,
|
| 6 |
"eval_samples": 2528,
|
| 7 |
+
"eval_samples_per_second": 42.263,
|
| 8 |
+
"eval_steps_per_second": 1.187,
|
| 9 |
+
"eval_wer": 0.09965613856342333
|
| 10 |
}
|
runs/Jan16_13-07-53_vorace/events.out.tfevents.1705415718.vorace.502984.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9dd9555772ff7031c1763a0f0080c729251887efd088d16d480fd9fec99c0d7e
|
| 3 |
+
size 405
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"train_loss":
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 28538,
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"train_loss": 0.09581804365822763,
|
| 4 |
+
"train_runtime": 4900.8909,
|
| 5 |
"train_samples": 28538,
|
| 6 |
+
"train_samples_per_second": 29.115,
|
| 7 |
+
"train_steps_per_second": 0.404
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 250,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1249,243 +1249,1227 @@
|
|
| 1249 |
"step": 1000
|
| 1250 |
},
|
| 1251 |
{
|
| 1252 |
-
"epoch": 2.
|
| 1253 |
"learning_rate": 3.006e-06,
|
| 1254 |
-
"loss": 0.
|
| 1255 |
"step": 1005
|
| 1256 |
},
|
| 1257 |
{
|
| 1258 |
"epoch": 2.55,
|
| 1259 |
-
"learning_rate": 3.
|
| 1260 |
-
"loss": 0.
|
| 1261 |
"step": 1010
|
| 1262 |
},
|
| 1263 |
{
|
| 1264 |
"epoch": 2.56,
|
| 1265 |
-
"learning_rate": 3.
|
| 1266 |
-
"loss": 0.
|
| 1267 |
"step": 1015
|
| 1268 |
},
|
| 1269 |
{
|
| 1270 |
-
"epoch": 2.
|
| 1271 |
-
"learning_rate": 3.
|
| 1272 |
-
"loss": 0.
|
| 1273 |
"step": 1020
|
| 1274 |
},
|
| 1275 |
{
|
| 1276 |
"epoch": 2.59,
|
| 1277 |
-
"learning_rate": 3.
|
| 1278 |
-
"loss": 0.
|
| 1279 |
"step": 1025
|
| 1280 |
},
|
| 1281 |
{
|
| 1282 |
"epoch": 2.6,
|
| 1283 |
-
"learning_rate": 3.
|
| 1284 |
-
"loss": 0.
|
| 1285 |
"step": 1030
|
| 1286 |
},
|
| 1287 |
{
|
| 1288 |
"epoch": 2.61,
|
| 1289 |
-
"learning_rate": 3.
|
| 1290 |
-
"loss": 0.
|
| 1291 |
"step": 1035
|
| 1292 |
},
|
| 1293 |
{
|
| 1294 |
-
"epoch": 2.
|
| 1295 |
-
"learning_rate": 3.
|
| 1296 |
-
"loss": 0.
|
| 1297 |
"step": 1040
|
| 1298 |
},
|
| 1299 |
{
|
| 1300 |
"epoch": 2.64,
|
| 1301 |
-
"learning_rate": 3.
|
| 1302 |
-
"loss": 0.
|
| 1303 |
"step": 1045
|
| 1304 |
},
|
| 1305 |
{
|
| 1306 |
"epoch": 2.65,
|
| 1307 |
-
"learning_rate": 3.
|
| 1308 |
-
"loss": 0.
|
| 1309 |
"step": 1050
|
| 1310 |
},
|
| 1311 |
{
|
| 1312 |
"epoch": 2.66,
|
| 1313 |
-
"learning_rate": 3.
|
| 1314 |
-
"loss": 0.
|
| 1315 |
"step": 1055
|
| 1316 |
},
|
| 1317 |
{
|
| 1318 |
-
"epoch": 2.
|
| 1319 |
-
"learning_rate": 3.
|
| 1320 |
-
"loss": 0.
|
| 1321 |
"step": 1060
|
| 1322 |
},
|
| 1323 |
{
|
| 1324 |
"epoch": 2.69,
|
| 1325 |
-
"learning_rate": 3.
|
| 1326 |
-
"loss": 0.
|
| 1327 |
"step": 1065
|
| 1328 |
},
|
| 1329 |
{
|
| 1330 |
"epoch": 2.7,
|
| 1331 |
-
"learning_rate": 3.
|
| 1332 |
-
"loss": 0.
|
| 1333 |
"step": 1070
|
| 1334 |
},
|
| 1335 |
{
|
| 1336 |
"epoch": 2.71,
|
| 1337 |
-
"learning_rate": 3.
|
| 1338 |
-
"loss": 0.
|
| 1339 |
"step": 1075
|
| 1340 |
},
|
| 1341 |
{
|
| 1342 |
-
"epoch": 2.
|
| 1343 |
-
"learning_rate": 3.
|
| 1344 |
-
"loss": 0.
|
| 1345 |
"step": 1080
|
| 1346 |
},
|
| 1347 |
{
|
| 1348 |
"epoch": 2.74,
|
| 1349 |
-
"learning_rate": 3.
|
| 1350 |
-
"loss": 0.
|
| 1351 |
"step": 1085
|
| 1352 |
},
|
| 1353 |
{
|
| 1354 |
"epoch": 2.75,
|
| 1355 |
-
"learning_rate": 3.
|
| 1356 |
-
"loss": 0.
|
| 1357 |
"step": 1090
|
| 1358 |
},
|
| 1359 |
{
|
| 1360 |
"epoch": 2.76,
|
| 1361 |
-
"learning_rate": 3.
|
| 1362 |
-
"loss": 0.
|
| 1363 |
"step": 1095
|
| 1364 |
},
|
| 1365 |
{
|
| 1366 |
-
"epoch": 2.
|
| 1367 |
-
"learning_rate": 3.
|
| 1368 |
-
"loss": 0.
|
| 1369 |
"step": 1100
|
| 1370 |
},
|
| 1371 |
{
|
| 1372 |
"epoch": 2.79,
|
| 1373 |
-
"learning_rate": 3.
|
| 1374 |
-
"loss": 0.
|
| 1375 |
"step": 1105
|
| 1376 |
},
|
| 1377 |
{
|
| 1378 |
"epoch": 2.8,
|
| 1379 |
-
"learning_rate": 3.
|
| 1380 |
-
"loss": 0.
|
| 1381 |
"step": 1110
|
| 1382 |
},
|
| 1383 |
{
|
| 1384 |
"epoch": 2.81,
|
| 1385 |
-
"learning_rate": 3.
|
| 1386 |
-
"loss": 0.
|
| 1387 |
"step": 1115
|
| 1388 |
},
|
| 1389 |
{
|
| 1390 |
-
"epoch": 2.
|
| 1391 |
-
"learning_rate": 3.
|
| 1392 |
-
"loss": 0.
|
| 1393 |
"step": 1120
|
| 1394 |
},
|
| 1395 |
{
|
| 1396 |
"epoch": 2.84,
|
| 1397 |
-
"learning_rate": 3.
|
| 1398 |
-
"loss": 0.
|
| 1399 |
"step": 1125
|
| 1400 |
},
|
| 1401 |
{
|
| 1402 |
"epoch": 2.85,
|
| 1403 |
-
"learning_rate": 3.
|
| 1404 |
-
"loss": 0.
|
| 1405 |
"step": 1130
|
| 1406 |
},
|
| 1407 |
{
|
| 1408 |
-
"epoch": 2.
|
| 1409 |
-
"learning_rate": 3.
|
| 1410 |
-
"loss": 0.
|
| 1411 |
"step": 1135
|
| 1412 |
},
|
| 1413 |
{
|
| 1414 |
"epoch": 2.88,
|
| 1415 |
-
"learning_rate": 3.
|
| 1416 |
-
"loss": 0.
|
| 1417 |
"step": 1140
|
| 1418 |
},
|
| 1419 |
{
|
| 1420 |
"epoch": 2.89,
|
| 1421 |
-
"learning_rate": 3.
|
| 1422 |
-
"loss": 0.
|
| 1423 |
"step": 1145
|
| 1424 |
},
|
| 1425 |
{
|
| 1426 |
"epoch": 2.9,
|
| 1427 |
-
"learning_rate": 3.
|
| 1428 |
-
"loss": 0.
|
| 1429 |
"step": 1150
|
| 1430 |
},
|
| 1431 |
{
|
| 1432 |
-
"epoch": 2.
|
| 1433 |
-
"learning_rate": 3.
|
| 1434 |
-
"loss": 0.
|
| 1435 |
"step": 1155
|
| 1436 |
},
|
| 1437 |
{
|
| 1438 |
"epoch": 2.93,
|
| 1439 |
-
"learning_rate": 3.
|
| 1440 |
-
"loss": 0.
|
| 1441 |
"step": 1160
|
| 1442 |
},
|
| 1443 |
{
|
| 1444 |
"epoch": 2.94,
|
| 1445 |
-
"learning_rate": 3.
|
| 1446 |
-
"loss": 0.
|
| 1447 |
"step": 1165
|
| 1448 |
},
|
| 1449 |
{
|
| 1450 |
"epoch": 2.95,
|
| 1451 |
-
"learning_rate": 3.
|
| 1452 |
-
"loss": 0.
|
| 1453 |
"step": 1170
|
| 1454 |
},
|
| 1455 |
{
|
| 1456 |
-
"epoch": 2.
|
| 1457 |
-
"learning_rate": 3.
|
| 1458 |
-
"loss": 0.
|
| 1459 |
"step": 1175
|
| 1460 |
},
|
| 1461 |
{
|
| 1462 |
"epoch": 2.98,
|
| 1463 |
-
"learning_rate": 3.
|
| 1464 |
-
"loss": 0.
|
| 1465 |
"step": 1180
|
| 1466 |
},
|
| 1467 |
{
|
| 1468 |
"epoch": 2.99,
|
| 1469 |
-
"learning_rate": 3.
|
| 1470 |
-
"loss": 0.
|
| 1471 |
"step": 1185
|
| 1472 |
},
|
| 1473 |
{
|
| 1474 |
"epoch": 3.0,
|
| 1475 |
-
"
|
| 1476 |
-
"
|
| 1477 |
-
"
|
| 1478 |
-
|
| 1479 |
-
|
| 1480 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1481 |
}
|
| 1482 |
],
|
| 1483 |
"logging_steps": 5,
|
| 1484 |
-
"max_steps":
|
| 1485 |
"num_input_tokens_seen": 0,
|
| 1486 |
-
"num_train_epochs":
|
| 1487 |
"save_steps": 500,
|
| 1488 |
-
"total_flos":
|
| 1489 |
"train_batch_size": 12,
|
| 1490 |
"trial_name": null,
|
| 1491 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 4.996216897856242,
|
| 5 |
"eval_steps": 250,
|
| 6 |
+
"global_step": 1980,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1249 |
"step": 1000
|
| 1250 |
},
|
| 1251 |
{
|
| 1252 |
+
"epoch": 2.54,
|
| 1253 |
"learning_rate": 3.006e-06,
|
| 1254 |
+
"loss": 0.292,
|
| 1255 |
"step": 1005
|
| 1256 |
},
|
| 1257 |
{
|
| 1258 |
"epoch": 2.55,
|
| 1259 |
+
"learning_rate": 3.0179999999999997e-06,
|
| 1260 |
+
"loss": 0.3656,
|
| 1261 |
"step": 1010
|
| 1262 |
},
|
| 1263 |
{
|
| 1264 |
"epoch": 2.56,
|
| 1265 |
+
"learning_rate": 3.033e-06,
|
| 1266 |
+
"loss": 0.3029,
|
| 1267 |
"step": 1015
|
| 1268 |
},
|
| 1269 |
{
|
| 1270 |
+
"epoch": 2.58,
|
| 1271 |
+
"learning_rate": 3.048e-06,
|
| 1272 |
+
"loss": 0.2976,
|
| 1273 |
"step": 1020
|
| 1274 |
},
|
| 1275 |
{
|
| 1276 |
"epoch": 2.59,
|
| 1277 |
+
"learning_rate": 3.063e-06,
|
| 1278 |
+
"loss": 0.3674,
|
| 1279 |
"step": 1025
|
| 1280 |
},
|
| 1281 |
{
|
| 1282 |
"epoch": 2.6,
|
| 1283 |
+
"learning_rate": 3.078e-06,
|
| 1284 |
+
"loss": 0.3063,
|
| 1285 |
"step": 1030
|
| 1286 |
},
|
| 1287 |
{
|
| 1288 |
"epoch": 2.61,
|
| 1289 |
+
"learning_rate": 3.093e-06,
|
| 1290 |
+
"loss": 0.2821,
|
| 1291 |
"step": 1035
|
| 1292 |
},
|
| 1293 |
{
|
| 1294 |
+
"epoch": 2.63,
|
| 1295 |
+
"learning_rate": 3.108e-06,
|
| 1296 |
+
"loss": 0.2804,
|
| 1297 |
"step": 1040
|
| 1298 |
},
|
| 1299 |
{
|
| 1300 |
"epoch": 2.64,
|
| 1301 |
+
"learning_rate": 3.123e-06,
|
| 1302 |
+
"loss": 0.3475,
|
| 1303 |
"step": 1045
|
| 1304 |
},
|
| 1305 |
{
|
| 1306 |
"epoch": 2.65,
|
| 1307 |
+
"learning_rate": 3.138e-06,
|
| 1308 |
+
"loss": 0.2895,
|
| 1309 |
"step": 1050
|
| 1310 |
},
|
| 1311 |
{
|
| 1312 |
"epoch": 2.66,
|
| 1313 |
+
"learning_rate": 3.153e-06,
|
| 1314 |
+
"loss": 0.278,
|
| 1315 |
"step": 1055
|
| 1316 |
},
|
| 1317 |
{
|
| 1318 |
+
"epoch": 2.68,
|
| 1319 |
+
"learning_rate": 3.168e-06,
|
| 1320 |
+
"loss": 0.3217,
|
| 1321 |
"step": 1060
|
| 1322 |
},
|
| 1323 |
{
|
| 1324 |
"epoch": 2.69,
|
| 1325 |
+
"learning_rate": 3.1830000000000003e-06,
|
| 1326 |
+
"loss": 0.2822,
|
| 1327 |
"step": 1065
|
| 1328 |
},
|
| 1329 |
{
|
| 1330 |
"epoch": 2.7,
|
| 1331 |
+
"learning_rate": 3.198e-06,
|
| 1332 |
+
"loss": 0.2736,
|
| 1333 |
"step": 1070
|
| 1334 |
},
|
| 1335 |
{
|
| 1336 |
"epoch": 2.71,
|
| 1337 |
+
"learning_rate": 3.213e-06,
|
| 1338 |
+
"loss": 0.3326,
|
| 1339 |
"step": 1075
|
| 1340 |
},
|
| 1341 |
{
|
| 1342 |
+
"epoch": 2.73,
|
| 1343 |
+
"learning_rate": 3.2280000000000003e-06,
|
| 1344 |
+
"loss": 0.2738,
|
| 1345 |
"step": 1080
|
| 1346 |
},
|
| 1347 |
{
|
| 1348 |
"epoch": 2.74,
|
| 1349 |
+
"learning_rate": 3.243e-06,
|
| 1350 |
+
"loss": 0.2712,
|
| 1351 |
"step": 1085
|
| 1352 |
},
|
| 1353 |
{
|
| 1354 |
"epoch": 2.75,
|
| 1355 |
+
"learning_rate": 3.258e-06,
|
| 1356 |
+
"loss": 0.2858,
|
| 1357 |
"step": 1090
|
| 1358 |
},
|
| 1359 |
{
|
| 1360 |
"epoch": 2.76,
|
| 1361 |
+
"learning_rate": 3.2730000000000003e-06,
|
| 1362 |
+
"loss": 0.2962,
|
| 1363 |
"step": 1095
|
| 1364 |
},
|
| 1365 |
{
|
| 1366 |
+
"epoch": 2.78,
|
| 1367 |
+
"learning_rate": 3.288e-06,
|
| 1368 |
+
"loss": 0.2485,
|
| 1369 |
"step": 1100
|
| 1370 |
},
|
| 1371 |
{
|
| 1372 |
"epoch": 2.79,
|
| 1373 |
+
"learning_rate": 3.3030000000000004e-06,
|
| 1374 |
+
"loss": 0.2566,
|
| 1375 |
"step": 1105
|
| 1376 |
},
|
| 1377 |
{
|
| 1378 |
"epoch": 2.8,
|
| 1379 |
+
"learning_rate": 3.315e-06,
|
| 1380 |
+
"loss": 0.2978,
|
| 1381 |
"step": 1110
|
| 1382 |
},
|
| 1383 |
{
|
| 1384 |
"epoch": 2.81,
|
| 1385 |
+
"learning_rate": 3.3300000000000003e-06,
|
| 1386 |
+
"loss": 0.2629,
|
| 1387 |
"step": 1115
|
| 1388 |
},
|
| 1389 |
{
|
| 1390 |
+
"epoch": 2.83,
|
| 1391 |
+
"learning_rate": 3.345e-06,
|
| 1392 |
+
"loss": 0.2559,
|
| 1393 |
"step": 1120
|
| 1394 |
},
|
| 1395 |
{
|
| 1396 |
"epoch": 2.84,
|
| 1397 |
+
"learning_rate": 3.36e-06,
|
| 1398 |
+
"loss": 0.2628,
|
| 1399 |
"step": 1125
|
| 1400 |
},
|
| 1401 |
{
|
| 1402 |
"epoch": 2.85,
|
| 1403 |
+
"learning_rate": 3.3750000000000003e-06,
|
| 1404 |
+
"loss": 0.2455,
|
| 1405 |
"step": 1130
|
| 1406 |
},
|
| 1407 |
{
|
| 1408 |
+
"epoch": 2.87,
|
| 1409 |
+
"learning_rate": 3.39e-06,
|
| 1410 |
+
"loss": 0.2501,
|
| 1411 |
"step": 1135
|
| 1412 |
},
|
| 1413 |
{
|
| 1414 |
"epoch": 2.88,
|
| 1415 |
+
"learning_rate": 3.405e-06,
|
| 1416 |
+
"loss": 0.265,
|
| 1417 |
"step": 1140
|
| 1418 |
},
|
| 1419 |
{
|
| 1420 |
"epoch": 2.89,
|
| 1421 |
+
"learning_rate": 3.417e-06,
|
| 1422 |
+
"loss": 0.2922,
|
| 1423 |
"step": 1145
|
| 1424 |
},
|
| 1425 |
{
|
| 1426 |
"epoch": 2.9,
|
| 1427 |
+
"learning_rate": 3.4320000000000003e-06,
|
| 1428 |
+
"loss": 0.2492,
|
| 1429 |
"step": 1150
|
| 1430 |
},
|
| 1431 |
{
|
| 1432 |
+
"epoch": 2.92,
|
| 1433 |
+
"learning_rate": 3.447e-06,
|
| 1434 |
+
"loss": 0.2417,
|
| 1435 |
"step": 1155
|
| 1436 |
},
|
| 1437 |
{
|
| 1438 |
"epoch": 2.93,
|
| 1439 |
+
"learning_rate": 3.462e-06,
|
| 1440 |
+
"loss": 0.27,
|
| 1441 |
"step": 1160
|
| 1442 |
},
|
| 1443 |
{
|
| 1444 |
"epoch": 2.94,
|
| 1445 |
+
"learning_rate": 3.4770000000000003e-06,
|
| 1446 |
+
"loss": 0.2335,
|
| 1447 |
"step": 1165
|
| 1448 |
},
|
| 1449 |
{
|
| 1450 |
"epoch": 2.95,
|
| 1451 |
+
"learning_rate": 3.492e-06,
|
| 1452 |
+
"loss": 0.2327,
|
| 1453 |
"step": 1170
|
| 1454 |
},
|
| 1455 |
{
|
| 1456 |
+
"epoch": 2.97,
|
| 1457 |
+
"learning_rate": 3.507e-06,
|
| 1458 |
+
"loss": 0.2634,
|
| 1459 |
"step": 1175
|
| 1460 |
},
|
| 1461 |
{
|
| 1462 |
"epoch": 2.98,
|
| 1463 |
+
"learning_rate": 3.5220000000000003e-06,
|
| 1464 |
+
"loss": 0.2389,
|
| 1465 |
"step": 1180
|
| 1466 |
},
|
| 1467 |
{
|
| 1468 |
"epoch": 2.99,
|
| 1469 |
+
"learning_rate": 3.537e-06,
|
| 1470 |
+
"loss": 0.223,
|
| 1471 |
"step": 1185
|
| 1472 |
},
|
| 1473 |
{
|
| 1474 |
"epoch": 3.0,
|
| 1475 |
+
"learning_rate": 3.552e-06,
|
| 1476 |
+
"loss": 0.2773,
|
| 1477 |
+
"step": 1190
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 3.02,
|
| 1481 |
+
"learning_rate": 3.5670000000000003e-06,
|
| 1482 |
+
"loss": 0.2285,
|
| 1483 |
+
"step": 1195
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 3.03,
|
| 1487 |
+
"learning_rate": 3.582e-06,
|
| 1488 |
+
"loss": 0.2183,
|
| 1489 |
+
"step": 1200
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 3.04,
|
| 1493 |
+
"learning_rate": 3.5970000000000005e-06,
|
| 1494 |
+
"loss": 0.2967,
|
| 1495 |
+
"step": 1205
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 3.05,
|
| 1499 |
+
"learning_rate": 3.612e-06,
|
| 1500 |
+
"loss": 0.2179,
|
| 1501 |
+
"step": 1210
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 3.07,
|
| 1505 |
+
"learning_rate": 3.6269999999999997e-06,
|
| 1506 |
+
"loss": 0.231,
|
| 1507 |
+
"step": 1215
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 3.08,
|
| 1511 |
+
"learning_rate": 3.642e-06,
|
| 1512 |
+
"loss": 0.2366,
|
| 1513 |
+
"step": 1220
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 3.09,
|
| 1517 |
+
"learning_rate": 3.657e-06,
|
| 1518 |
+
"loss": 0.2462,
|
| 1519 |
+
"step": 1225
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 3.1,
|
| 1523 |
+
"learning_rate": 3.6719999999999997e-06,
|
| 1524 |
+
"loss": 0.2151,
|
| 1525 |
+
"step": 1230
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 3.12,
|
| 1529 |
+
"learning_rate": 3.687e-06,
|
| 1530 |
+
"loss": 0.2162,
|
| 1531 |
+
"step": 1235
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 3.13,
|
| 1535 |
+
"learning_rate": 3.702e-06,
|
| 1536 |
+
"loss": 0.2532,
|
| 1537 |
+
"step": 1240
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 3.14,
|
| 1541 |
+
"learning_rate": 3.717e-06,
|
| 1542 |
+
"loss": 0.2035,
|
| 1543 |
+
"step": 1245
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 3.16,
|
| 1547 |
+
"learning_rate": 3.732e-06,
|
| 1548 |
+
"loss": 0.2089,
|
| 1549 |
+
"step": 1250
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 3.16,
|
| 1553 |
+
"eval_cer": 0.044489706264102716,
|
| 1554 |
+
"eval_loss": 0.2078969031572342,
|
| 1555 |
+
"eval_runtime": 113.5687,
|
| 1556 |
+
"eval_samples_per_second": 22.26,
|
| 1557 |
+
"eval_steps_per_second": 0.625,
|
| 1558 |
+
"eval_wer": 0.17657921548650027,
|
| 1559 |
+
"step": 1250
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 3.17,
|
| 1563 |
+
"learning_rate": 3.747e-06,
|
| 1564 |
+
"loss": 0.2531,
|
| 1565 |
+
"step": 1255
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 3.18,
|
| 1569 |
+
"learning_rate": 3.7620000000000006e-06,
|
| 1570 |
+
"loss": 0.2165,
|
| 1571 |
+
"step": 1260
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 3.19,
|
| 1575 |
+
"learning_rate": 3.7770000000000004e-06,
|
| 1576 |
+
"loss": 0.212,
|
| 1577 |
+
"step": 1265
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 3.21,
|
| 1581 |
+
"learning_rate": 3.7920000000000003e-06,
|
| 1582 |
+
"loss": 0.2218,
|
| 1583 |
+
"step": 1270
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 3.22,
|
| 1587 |
+
"learning_rate": 3.8070000000000006e-06,
|
| 1588 |
+
"loss": 0.2297,
|
| 1589 |
+
"step": 1275
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 3.23,
|
| 1593 |
+
"learning_rate": 3.822000000000001e-06,
|
| 1594 |
+
"loss": 0.2095,
|
| 1595 |
+
"step": 1280
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 3.24,
|
| 1599 |
+
"learning_rate": 3.837000000000001e-06,
|
| 1600 |
+
"loss": 0.2028,
|
| 1601 |
+
"step": 1285
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 3.26,
|
| 1605 |
+
"learning_rate": 3.852e-06,
|
| 1606 |
+
"loss": 0.2619,
|
| 1607 |
+
"step": 1290
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 3.27,
|
| 1611 |
+
"learning_rate": 3.8669999999999996e-06,
|
| 1612 |
+
"loss": 0.1918,
|
| 1613 |
+
"step": 1295
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 3.28,
|
| 1617 |
+
"learning_rate": 3.8819999999999994e-06,
|
| 1618 |
+
"loss": 0.1945,
|
| 1619 |
+
"step": 1300
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 3.29,
|
| 1623 |
+
"learning_rate": 3.897e-06,
|
| 1624 |
+
"loss": 0.2347,
|
| 1625 |
+
"step": 1305
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 3.31,
|
| 1629 |
+
"learning_rate": 3.912e-06,
|
| 1630 |
+
"loss": 0.2111,
|
| 1631 |
+
"step": 1310
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 3.32,
|
| 1635 |
+
"learning_rate": 3.927e-06,
|
| 1636 |
+
"loss": 0.207,
|
| 1637 |
+
"step": 1315
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 3.33,
|
| 1641 |
+
"learning_rate": 3.942e-06,
|
| 1642 |
+
"loss": 0.2281,
|
| 1643 |
+
"step": 1320
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 3.34,
|
| 1647 |
+
"learning_rate": 3.9569999999999996e-06,
|
| 1648 |
+
"loss": 0.2334,
|
| 1649 |
+
"step": 1325
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 3.36,
|
| 1653 |
+
"learning_rate": 3.971999999999999e-06,
|
| 1654 |
+
"loss": 0.1804,
|
| 1655 |
+
"step": 1330
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 3.37,
|
| 1659 |
+
"learning_rate": 3.987e-06,
|
| 1660 |
+
"loss": 0.1943,
|
| 1661 |
+
"step": 1335
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 3.38,
|
| 1665 |
+
"learning_rate": 4.002e-06,
|
| 1666 |
+
"loss": 0.2298,
|
| 1667 |
+
"step": 1340
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 3.39,
|
| 1671 |
+
"learning_rate": 4.017e-06,
|
| 1672 |
+
"loss": 0.1911,
|
| 1673 |
+
"step": 1345
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 3.41,
|
| 1677 |
+
"learning_rate": 4.032e-06,
|
| 1678 |
+
"loss": 0.1967,
|
| 1679 |
+
"step": 1350
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"epoch": 3.42,
|
| 1683 |
+
"learning_rate": 4.0469999999999995e-06,
|
| 1684 |
+
"loss": 0.2311,
|
| 1685 |
+
"step": 1355
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 3.43,
|
| 1689 |
+
"learning_rate": 4.062e-06,
|
| 1690 |
+
"loss": 0.201,
|
| 1691 |
+
"step": 1360
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 3.45,
|
| 1695 |
+
"learning_rate": 4.077e-06,
|
| 1696 |
+
"loss": 0.1956,
|
| 1697 |
+
"step": 1365
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 3.46,
|
| 1701 |
+
"learning_rate": 4.092e-06,
|
| 1702 |
+
"loss": 0.2068,
|
| 1703 |
+
"step": 1370
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 3.47,
|
| 1707 |
+
"learning_rate": 4.107e-06,
|
| 1708 |
+
"loss": 0.2211,
|
| 1709 |
+
"step": 1375
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 3.48,
|
| 1713 |
+
"learning_rate": 4.122e-06,
|
| 1714 |
+
"loss": 0.187,
|
| 1715 |
+
"step": 1380
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 3.5,
|
| 1719 |
+
"learning_rate": 4.137e-06,
|
| 1720 |
+
"loss": 0.1901,
|
| 1721 |
+
"step": 1385
|
| 1722 |
+
},
|
| 1723 |
+
{
|
| 1724 |
+
"epoch": 3.51,
|
| 1725 |
+
"learning_rate": 4.152e-06,
|
| 1726 |
+
"loss": 0.2077,
|
| 1727 |
+
"step": 1390
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 3.52,
|
| 1731 |
+
"learning_rate": 4.167e-06,
|
| 1732 |
+
"loss": 0.1781,
|
| 1733 |
+
"step": 1395
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 3.53,
|
| 1737 |
+
"learning_rate": 4.182e-06,
|
| 1738 |
+
"loss": 0.1806,
|
| 1739 |
+
"step": 1400
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 3.55,
|
| 1743 |
+
"learning_rate": 4.197e-06,
|
| 1744 |
+
"loss": 0.2346,
|
| 1745 |
+
"step": 1405
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 3.56,
|
| 1749 |
+
"learning_rate": 4.212e-06,
|
| 1750 |
+
"loss": 0.1864,
|
| 1751 |
+
"step": 1410
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 3.57,
|
| 1755 |
+
"learning_rate": 4.227e-06,
|
| 1756 |
+
"loss": 0.1682,
|
| 1757 |
+
"step": 1415
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 3.58,
|
| 1761 |
+
"learning_rate": 4.242e-06,
|
| 1762 |
+
"loss": 0.1944,
|
| 1763 |
+
"step": 1420
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 3.6,
|
| 1767 |
+
"learning_rate": 4.257e-06,
|
| 1768 |
+
"loss": 0.2108,
|
| 1769 |
+
"step": 1425
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 3.61,
|
| 1773 |
+
"learning_rate": 4.272e-06,
|
| 1774 |
+
"loss": 0.1731,
|
| 1775 |
+
"step": 1430
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 3.62,
|
| 1779 |
+
"learning_rate": 4.287e-06,
|
| 1780 |
+
"loss": 0.1734,
|
| 1781 |
+
"step": 1435
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 3.63,
|
| 1785 |
+
"learning_rate": 4.3020000000000005e-06,
|
| 1786 |
+
"loss": 0.2366,
|
| 1787 |
+
"step": 1440
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 3.65,
|
| 1791 |
+
"learning_rate": 4.317e-06,
|
| 1792 |
+
"loss": 0.1858,
|
| 1793 |
+
"step": 1445
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 3.66,
|
| 1797 |
+
"learning_rate": 4.332e-06,
|
| 1798 |
+
"loss": 0.1837,
|
| 1799 |
+
"step": 1450
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 3.67,
|
| 1803 |
+
"learning_rate": 4.347e-06,
|
| 1804 |
+
"loss": 0.2047,
|
| 1805 |
+
"step": 1455
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 3.68,
|
| 1809 |
+
"learning_rate": 4.362e-06,
|
| 1810 |
+
"loss": 0.1838,
|
| 1811 |
+
"step": 1460
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 3.7,
|
| 1815 |
+
"learning_rate": 4.377e-06,
|
| 1816 |
+
"loss": 0.1684,
|
| 1817 |
+
"step": 1465
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 3.71,
|
| 1821 |
+
"learning_rate": 4.3920000000000005e-06,
|
| 1822 |
+
"loss": 0.1965,
|
| 1823 |
+
"step": 1470
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 3.72,
|
| 1827 |
+
"learning_rate": 4.407e-06,
|
| 1828 |
+
"loss": 0.2069,
|
| 1829 |
+
"step": 1475
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 3.74,
|
| 1833 |
+
"learning_rate": 4.422e-06,
|
| 1834 |
+
"loss": 0.1751,
|
| 1835 |
+
"step": 1480
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 3.75,
|
| 1839 |
+
"learning_rate": 4.437e-06,
|
| 1840 |
+
"loss": 0.1724,
|
| 1841 |
+
"step": 1485
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 3.76,
|
| 1845 |
+
"learning_rate": 4.452e-06,
|
| 1846 |
+
"loss": 0.2224,
|
| 1847 |
+
"step": 1490
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 3.77,
|
| 1851 |
+
"learning_rate": 4.467000000000001e-06,
|
| 1852 |
+
"loss": 0.1684,
|
| 1853 |
+
"step": 1495
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 3.79,
|
| 1857 |
+
"learning_rate": 4.4820000000000005e-06,
|
| 1858 |
+
"loss": 0.1634,
|
| 1859 |
+
"step": 1500
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 3.79,
|
| 1863 |
+
"eval_cer": 0.03657820643383454,
|
| 1864 |
+
"eval_loss": 0.16873595118522644,
|
| 1865 |
+
"eval_runtime": 60.2617,
|
| 1866 |
+
"eval_samples_per_second": 41.95,
|
| 1867 |
+
"eval_steps_per_second": 1.178,
|
| 1868 |
+
"eval_wer": 0.14106809305484802,
|
| 1869 |
+
"step": 1500
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 3.8,
|
| 1873 |
+
"learning_rate": 4.497e-06,
|
| 1874 |
+
"loss": 0.2137,
|
| 1875 |
+
"step": 1505
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 3.81,
|
| 1879 |
+
"learning_rate": 4.5089999999999995e-06,
|
| 1880 |
+
"loss": 0.1759,
|
| 1881 |
+
"step": 1510
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 3.82,
|
| 1885 |
+
"learning_rate": 4.524e-06,
|
| 1886 |
+
"loss": 0.1698,
|
| 1887 |
+
"step": 1515
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 3.84,
|
| 1891 |
+
"learning_rate": 4.539e-06,
|
| 1892 |
+
"loss": 0.1703,
|
| 1893 |
+
"step": 1520
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 3.85,
|
| 1897 |
+
"learning_rate": 4.554e-06,
|
| 1898 |
+
"loss": 0.1841,
|
| 1899 |
+
"step": 1525
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 3.86,
|
| 1903 |
+
"learning_rate": 4.569e-06,
|
| 1904 |
+
"loss": 0.1783,
|
| 1905 |
+
"step": 1530
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 3.87,
|
| 1909 |
+
"learning_rate": 4.584e-06,
|
| 1910 |
+
"loss": 0.1633,
|
| 1911 |
+
"step": 1535
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 3.89,
|
| 1915 |
+
"learning_rate": 4.5989999999999995e-06,
|
| 1916 |
+
"loss": 0.1839,
|
| 1917 |
+
"step": 1540
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 3.9,
|
| 1921 |
+
"learning_rate": 4.614e-06,
|
| 1922 |
+
"loss": 0.153,
|
| 1923 |
+
"step": 1545
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 3.91,
|
| 1927 |
+
"learning_rate": 4.629e-06,
|
| 1928 |
+
"loss": 0.1553,
|
| 1929 |
+
"step": 1550
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 3.92,
|
| 1933 |
+
"learning_rate": 4.644e-06,
|
| 1934 |
+
"loss": 0.1834,
|
| 1935 |
+
"step": 1555
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 3.94,
|
| 1939 |
+
"learning_rate": 4.659e-06,
|
| 1940 |
+
"loss": 0.1582,
|
| 1941 |
+
"step": 1560
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 3.95,
|
| 1945 |
+
"learning_rate": 4.674e-06,
|
| 1946 |
+
"loss": 0.1592,
|
| 1947 |
+
"step": 1565
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 3.96,
|
| 1951 |
+
"learning_rate": 4.689e-06,
|
| 1952 |
+
"loss": 0.1823,
|
| 1953 |
+
"step": 1570
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 3.97,
|
| 1957 |
+
"learning_rate": 4.704e-06,
|
| 1958 |
+
"loss": 0.1782,
|
| 1959 |
+
"step": 1575
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 3.99,
|
| 1963 |
+
"learning_rate": 4.719e-06,
|
| 1964 |
+
"loss": 0.1543,
|
| 1965 |
+
"step": 1580
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 4.0,
|
| 1969 |
+
"learning_rate": 4.734e-06,
|
| 1970 |
+
"loss": 0.1943,
|
| 1971 |
+
"step": 1585
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 4.01,
|
| 1975 |
+
"learning_rate": 4.749e-06,
|
| 1976 |
+
"loss": 0.1491,
|
| 1977 |
+
"step": 1590
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 4.03,
|
| 1981 |
+
"learning_rate": 4.764e-06,
|
| 1982 |
+
"loss": 0.1656,
|
| 1983 |
+
"step": 1595
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 4.04,
|
| 1987 |
+
"learning_rate": 4.779e-06,
|
| 1988 |
+
"loss": 0.1507,
|
| 1989 |
+
"step": 1600
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 4.05,
|
| 1993 |
+
"learning_rate": 4.794e-06,
|
| 1994 |
+
"loss": 0.1589,
|
| 1995 |
+
"step": 1605
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 4.06,
|
| 1999 |
+
"learning_rate": 4.809e-06,
|
| 2000 |
+
"loss": 0.1443,
|
| 2001 |
+
"step": 1610
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 4.08,
|
| 2005 |
+
"learning_rate": 4.824e-06,
|
| 2006 |
+
"loss": 0.1487,
|
| 2007 |
+
"step": 1615
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 4.09,
|
| 2011 |
+
"learning_rate": 4.839e-06,
|
| 2012 |
+
"loss": 0.1766,
|
| 2013 |
+
"step": 1620
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 4.1,
|
| 2017 |
+
"learning_rate": 4.8540000000000005e-06,
|
| 2018 |
+
"loss": 0.1548,
|
| 2019 |
+
"step": 1625
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 4.11,
|
| 2023 |
+
"learning_rate": 4.869e-06,
|
| 2024 |
+
"loss": 0.1494,
|
| 2025 |
+
"step": 1630
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 4.13,
|
| 2029 |
+
"learning_rate": 4.884e-06,
|
| 2030 |
+
"loss": 0.1793,
|
| 2031 |
+
"step": 1635
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 4.14,
|
| 2035 |
+
"learning_rate": 4.899e-06,
|
| 2036 |
+
"loss": 0.1417,
|
| 2037 |
+
"step": 1640
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 4.15,
|
| 2041 |
+
"learning_rate": 4.914e-06,
|
| 2042 |
+
"loss": 0.1399,
|
| 2043 |
+
"step": 1645
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 4.16,
|
| 2047 |
+
"learning_rate": 4.929000000000001e-06,
|
| 2048 |
+
"loss": 0.166,
|
| 2049 |
+
"step": 1650
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 4.18,
|
| 2053 |
+
"learning_rate": 4.9440000000000004e-06,
|
| 2054 |
+
"loss": 0.1748,
|
| 2055 |
+
"step": 1655
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 4.19,
|
| 2059 |
+
"learning_rate": 4.959e-06,
|
| 2060 |
+
"loss": 0.1378,
|
| 2061 |
+
"step": 1660
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 4.2,
|
| 2065 |
+
"learning_rate": 4.974e-06,
|
| 2066 |
+
"loss": 0.144,
|
| 2067 |
+
"step": 1665
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 4.21,
|
| 2071 |
+
"learning_rate": 4.989e-06,
|
| 2072 |
+
"loss": 0.1794,
|
| 2073 |
+
"step": 1670
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 4.23,
|
| 2077 |
+
"learning_rate": 5.004e-06,
|
| 2078 |
+
"loss": 0.146,
|
| 2079 |
+
"step": 1675
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 4.24,
|
| 2083 |
+
"learning_rate": 5.0190000000000006e-06,
|
| 2084 |
+
"loss": 0.141,
|
| 2085 |
+
"step": 1680
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 4.25,
|
| 2089 |
+
"learning_rate": 5.034e-06,
|
| 2090 |
+
"loss": 0.1757,
|
| 2091 |
+
"step": 1685
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 4.26,
|
| 2095 |
+
"learning_rate": 5.049e-06,
|
| 2096 |
+
"loss": 0.1423,
|
| 2097 |
+
"step": 1690
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 4.28,
|
| 2101 |
+
"learning_rate": 5.064e-06,
|
| 2102 |
+
"loss": 0.1414,
|
| 2103 |
+
"step": 1695
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 4.29,
|
| 2107 |
+
"learning_rate": 5.079e-06,
|
| 2108 |
+
"loss": 0.1556,
|
| 2109 |
+
"step": 1700
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 4.3,
|
| 2113 |
+
"learning_rate": 5.094000000000001e-06,
|
| 2114 |
+
"loss": 0.1599,
|
| 2115 |
+
"step": 1705
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 4.32,
|
| 2119 |
+
"learning_rate": 5.1090000000000006e-06,
|
| 2120 |
+
"loss": 0.1436,
|
| 2121 |
+
"step": 1710
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 4.33,
|
| 2125 |
+
"learning_rate": 5.124e-06,
|
| 2126 |
+
"loss": 0.1346,
|
| 2127 |
+
"step": 1715
|
| 2128 |
+
},
|
| 2129 |
+
{
|
| 2130 |
+
"epoch": 4.34,
|
| 2131 |
+
"learning_rate": 5.139e-06,
|
| 2132 |
+
"loss": 0.1702,
|
| 2133 |
+
"step": 1720
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 4.35,
|
| 2137 |
+
"learning_rate": 5.154e-06,
|
| 2138 |
+
"loss": 0.1397,
|
| 2139 |
+
"step": 1725
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 4.37,
|
| 2143 |
+
"learning_rate": 5.169e-06,
|
| 2144 |
+
"loss": 0.1373,
|
| 2145 |
+
"step": 1730
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 4.38,
|
| 2149 |
+
"learning_rate": 5.184000000000001e-06,
|
| 2150 |
+
"loss": 0.1929,
|
| 2151 |
+
"step": 1735
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 4.39,
|
| 2155 |
+
"learning_rate": 5.1990000000000005e-06,
|
| 2156 |
+
"loss": 0.1426,
|
| 2157 |
+
"step": 1740
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 4.4,
|
| 2161 |
+
"learning_rate": 5.214e-06,
|
| 2162 |
+
"loss": 0.1481,
|
| 2163 |
+
"step": 1745
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 4.42,
|
| 2167 |
+
"learning_rate": 5.229e-06,
|
| 2168 |
+
"loss": 0.163,
|
| 2169 |
+
"step": 1750
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 4.42,
|
| 2173 |
+
"eval_cer": 0.029768965035244312,
|
| 2174 |
+
"eval_loss": 0.1489591896533966,
|
| 2175 |
+
"eval_runtime": 60.6618,
|
| 2176 |
+
"eval_samples_per_second": 41.674,
|
| 2177 |
+
"eval_steps_per_second": 1.17,
|
| 2178 |
+
"eval_wer": 0.1163397860417728,
|
| 2179 |
+
"step": 1750
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 4.43,
|
| 2183 |
+
"learning_rate": 5.244e-06,
|
| 2184 |
+
"loss": 0.1816,
|
| 2185 |
+
"step": 1755
|
| 2186 |
+
},
|
| 2187 |
+
{
|
| 2188 |
+
"epoch": 4.44,
|
| 2189 |
+
"learning_rate": 5.259000000000001e-06,
|
| 2190 |
+
"loss": 0.1444,
|
| 2191 |
+
"step": 1760
|
| 2192 |
+
},
|
| 2193 |
+
{
|
| 2194 |
+
"epoch": 4.45,
|
| 2195 |
+
"learning_rate": 5.274000000000001e-06,
|
| 2196 |
+
"loss": 0.1423,
|
| 2197 |
+
"step": 1765
|
| 2198 |
+
},
|
| 2199 |
+
{
|
| 2200 |
+
"epoch": 4.47,
|
| 2201 |
+
"learning_rate": 5.2890000000000005e-06,
|
| 2202 |
+
"loss": 0.1605,
|
| 2203 |
+
"step": 1770
|
| 2204 |
+
},
|
| 2205 |
+
{
|
| 2206 |
+
"epoch": 4.48,
|
| 2207 |
+
"learning_rate": 5.304e-06,
|
| 2208 |
+
"loss": 0.1355,
|
| 2209 |
+
"step": 1775
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 4.49,
|
| 2213 |
+
"learning_rate": 5.319e-06,
|
| 2214 |
+
"loss": 0.1401,
|
| 2215 |
+
"step": 1780
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 4.5,
|
| 2219 |
+
"learning_rate": 5.334000000000001e-06,
|
| 2220 |
+
"loss": 0.1593,
|
| 2221 |
+
"step": 1785
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 4.52,
|
| 2225 |
+
"learning_rate": 5.349e-06,
|
| 2226 |
+
"loss": 0.1361,
|
| 2227 |
+
"step": 1790
|
| 2228 |
+
},
|
| 2229 |
+
{
|
| 2230 |
+
"epoch": 4.53,
|
| 2231 |
+
"learning_rate": 5.364e-06,
|
| 2232 |
+
"loss": 0.1363,
|
| 2233 |
+
"step": 1795
|
| 2234 |
+
},
|
| 2235 |
+
{
|
| 2236 |
+
"epoch": 4.54,
|
| 2237 |
+
"learning_rate": 5.379e-06,
|
| 2238 |
+
"loss": 0.1527,
|
| 2239 |
+
"step": 1800
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 4.55,
|
| 2243 |
+
"learning_rate": 5.3939999999999995e-06,
|
| 2244 |
+
"loss": 0.1645,
|
| 2245 |
+
"step": 1805
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 4.57,
|
| 2249 |
+
"learning_rate": 5.408999999999999e-06,
|
| 2250 |
+
"loss": 0.1285,
|
| 2251 |
+
"step": 1810
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 4.58,
|
| 2255 |
+
"learning_rate": 5.424e-06,
|
| 2256 |
+
"loss": 0.1366,
|
| 2257 |
+
"step": 1815
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 4.59,
|
| 2261 |
+
"learning_rate": 5.439e-06,
|
| 2262 |
+
"loss": 0.1808,
|
| 2263 |
+
"step": 1820
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 4.61,
|
| 2267 |
+
"learning_rate": 5.454e-06,
|
| 2268 |
+
"loss": 0.1428,
|
| 2269 |
+
"step": 1825
|
| 2270 |
+
},
|
| 2271 |
+
{
|
| 2272 |
+
"epoch": 4.62,
|
| 2273 |
+
"learning_rate": 5.469e-06,
|
| 2274 |
+
"loss": 0.1314,
|
| 2275 |
+
"step": 1830
|
| 2276 |
+
},
|
| 2277 |
+
{
|
| 2278 |
+
"epoch": 4.63,
|
| 2279 |
+
"learning_rate": 5.4839999999999995e-06,
|
| 2280 |
+
"loss": 0.1479,
|
| 2281 |
+
"step": 1835
|
| 2282 |
+
},
|
| 2283 |
+
{
|
| 2284 |
+
"epoch": 4.64,
|
| 2285 |
+
"learning_rate": 5.499e-06,
|
| 2286 |
+
"loss": 0.135,
|
| 2287 |
+
"step": 1840
|
| 2288 |
+
},
|
| 2289 |
+
{
|
| 2290 |
+
"epoch": 4.66,
|
| 2291 |
+
"learning_rate": 5.514e-06,
|
| 2292 |
+
"loss": 0.1253,
|
| 2293 |
+
"step": 1845
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 4.67,
|
| 2297 |
+
"learning_rate": 5.529e-06,
|
| 2298 |
+
"loss": 0.13,
|
| 2299 |
+
"step": 1850
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 4.68,
|
| 2303 |
+
"learning_rate": 5.544e-06,
|
| 2304 |
+
"loss": 0.1484,
|
| 2305 |
+
"step": 1855
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 4.69,
|
| 2309 |
+
"learning_rate": 5.559e-06,
|
| 2310 |
+
"loss": 0.1201,
|
| 2311 |
+
"step": 1860
|
| 2312 |
+
},
|
| 2313 |
+
{
|
| 2314 |
+
"epoch": 4.71,
|
| 2315 |
+
"learning_rate": 5.574e-06,
|
| 2316 |
+
"loss": 0.1263,
|
| 2317 |
+
"step": 1865
|
| 2318 |
+
},
|
| 2319 |
+
{
|
| 2320 |
+
"epoch": 4.72,
|
| 2321 |
+
"learning_rate": 5.589e-06,
|
| 2322 |
+
"loss": 0.1493,
|
| 2323 |
+
"step": 1870
|
| 2324 |
+
},
|
| 2325 |
+
{
|
| 2326 |
+
"epoch": 4.73,
|
| 2327 |
+
"learning_rate": 5.604e-06,
|
| 2328 |
+
"loss": 0.1276,
|
| 2329 |
+
"step": 1875
|
| 2330 |
+
},
|
| 2331 |
+
{
|
| 2332 |
+
"epoch": 4.74,
|
| 2333 |
+
"learning_rate": 5.619e-06,
|
| 2334 |
+
"loss": 0.1236,
|
| 2335 |
+
"step": 1880
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 4.76,
|
| 2339 |
+
"learning_rate": 5.634e-06,
|
| 2340 |
+
"loss": 0.1718,
|
| 2341 |
+
"step": 1885
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 4.77,
|
| 2345 |
+
"learning_rate": 5.649e-06,
|
| 2346 |
+
"loss": 0.1292,
|
| 2347 |
+
"step": 1890
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 4.78,
|
| 2351 |
+
"learning_rate": 5.664e-06,
|
| 2352 |
+
"loss": 0.1175,
|
| 2353 |
+
"step": 1895
|
| 2354 |
+
},
|
| 2355 |
+
{
|
| 2356 |
+
"epoch": 4.79,
|
| 2357 |
+
"learning_rate": 5.679e-06,
|
| 2358 |
+
"loss": 0.1317,
|
| 2359 |
+
"step": 1900
|
| 2360 |
+
},
|
| 2361 |
+
{
|
| 2362 |
+
"epoch": 4.81,
|
| 2363 |
+
"learning_rate": 5.694e-06,
|
| 2364 |
+
"loss": 0.1412,
|
| 2365 |
+
"step": 1905
|
| 2366 |
+
},
|
| 2367 |
+
{
|
| 2368 |
+
"epoch": 4.82,
|
| 2369 |
+
"learning_rate": 5.709e-06,
|
| 2370 |
+
"loss": 0.1222,
|
| 2371 |
+
"step": 1910
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 4.83,
|
| 2375 |
+
"learning_rate": 5.724e-06,
|
| 2376 |
+
"loss": 0.1193,
|
| 2377 |
+
"step": 1915
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 4.84,
|
| 2381 |
+
"learning_rate": 5.7390000000000004e-06,
|
| 2382 |
+
"loss": 0.16,
|
| 2383 |
+
"step": 1920
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 4.86,
|
| 2387 |
+
"learning_rate": 5.754e-06,
|
| 2388 |
+
"loss": 0.1228,
|
| 2389 |
+
"step": 1925
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 4.87,
|
| 2393 |
+
"learning_rate": 5.769e-06,
|
| 2394 |
+
"loss": 0.1195,
|
| 2395 |
+
"step": 1930
|
| 2396 |
+
},
|
| 2397 |
+
{
|
| 2398 |
+
"epoch": 4.88,
|
| 2399 |
+
"learning_rate": 5.784e-06,
|
| 2400 |
+
"loss": 0.1608,
|
| 2401 |
+
"step": 1935
|
| 2402 |
+
},
|
| 2403 |
+
{
|
| 2404 |
+
"epoch": 4.9,
|
| 2405 |
+
"learning_rate": 5.799e-06,
|
| 2406 |
+
"loss": 0.1234,
|
| 2407 |
+
"step": 1940
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 4.91,
|
| 2411 |
+
"learning_rate": 5.814e-06,
|
| 2412 |
+
"loss": 0.1265,
|
| 2413 |
+
"step": 1945
|
| 2414 |
+
},
|
| 2415 |
+
{
|
| 2416 |
+
"epoch": 4.92,
|
| 2417 |
+
"learning_rate": 5.8290000000000004e-06,
|
| 2418 |
+
"loss": 0.1431,
|
| 2419 |
+
"step": 1950
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 4.93,
|
| 2423 |
+
"learning_rate": 5.844e-06,
|
| 2424 |
+
"loss": 0.1366,
|
| 2425 |
+
"step": 1955
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 4.95,
|
| 2429 |
+
"learning_rate": 5.859e-06,
|
| 2430 |
+
"loss": 0.1091,
|
| 2431 |
+
"step": 1960
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 4.96,
|
| 2435 |
+
"learning_rate": 5.874e-06,
|
| 2436 |
+
"loss": 0.139,
|
| 2437 |
+
"step": 1965
|
| 2438 |
+
},
|
| 2439 |
+
{
|
| 2440 |
+
"epoch": 4.97,
|
| 2441 |
+
"learning_rate": 5.886000000000001e-06,
|
| 2442 |
+
"loss": 0.1726,
|
| 2443 |
+
"step": 1970
|
| 2444 |
+
},
|
| 2445 |
+
{
|
| 2446 |
+
"epoch": 4.98,
|
| 2447 |
+
"learning_rate": 5.901000000000001e-06,
|
| 2448 |
+
"loss": 0.1148,
|
| 2449 |
+
"step": 1975
|
| 2450 |
+
},
|
| 2451 |
+
{
|
| 2452 |
+
"epoch": 5.0,
|
| 2453 |
+
"learning_rate": 5.916e-06,
|
| 2454 |
+
"loss": 0.1296,
|
| 2455 |
+
"step": 1980
|
| 2456 |
+
},
|
| 2457 |
+
{
|
| 2458 |
+
"epoch": 5.0,
|
| 2459 |
+
"step": 1980,
|
| 2460 |
+
"total_flos": 5.2867996741439324e+19,
|
| 2461 |
+
"train_loss": 0.09581804365822763,
|
| 2462 |
+
"train_runtime": 4900.8909,
|
| 2463 |
+
"train_samples_per_second": 29.115,
|
| 2464 |
+
"train_steps_per_second": 0.404
|
| 2465 |
}
|
| 2466 |
],
|
| 2467 |
"logging_steps": 5,
|
| 2468 |
+
"max_steps": 1980,
|
| 2469 |
"num_input_tokens_seen": 0,
|
| 2470 |
+
"num_train_epochs": 5,
|
| 2471 |
"save_steps": 500,
|
| 2472 |
+
"total_flos": 5.2867996741439324e+19,
|
| 2473 |
"train_batch_size": 12,
|
| 2474 |
"trial_name": null,
|
| 2475 |
"trial_params": null
|