Training in progress, step 300, 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 411094576
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c6998651a4f6221c924c5f14c9847c85d7f4ff286da87b492a2c183aab4f8ae3
|
| 3 |
size 411094576
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3c709930fcd73234c8fd0636ff4ab5c31426d68d32e9fb3a01ebf09b79f50d61
|
| 3 |
+
size 209193780
|
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 sha256:5d4c3d45bd5aae62c92c9ff393cc15dabbde2961fe98f16264940444bac07041
|
| 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:ebdb14d51e77eb18f9d6184de19bfac710da5493717593749289db85474b6091
|
| 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": 100,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1431,6 +1431,714 @@
|
|
| 1431 |
"eval_samples_per_second": 19.651,
|
| 1432 |
"eval_steps_per_second": 4.913,
|
| 1433 |
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1434 |
}
|
| 1435 |
],
|
| 1436 |
"logging_steps": 1,
|
|
@@ -1459,7 +2167,7 @@
|
|
| 1459 |
"attributes": {}
|
| 1460 |
}
|
| 1461 |
},
|
| 1462 |
-
"total_flos": 1.
|
| 1463 |
"train_batch_size": 8,
|
| 1464 |
"trial_name": null,
|
| 1465 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.8697348237037659,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-300",
|
| 4 |
+
"epoch": 0.003988022638675179,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 300,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1431 |
"eval_samples_per_second": 19.651,
|
| 1432 |
"eval_steps_per_second": 4.913,
|
| 1433 |
"step": 200
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 0.00267197516791237,
|
| 1437 |
+
"grad_norm": 1.469771385192871,
|
| 1438 |
+
"learning_rate": 6.883142508466054e-05,
|
| 1439 |
+
"loss": 1.1812,
|
| 1440 |
+
"step": 201
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 0.0026852685767079537,
|
| 1444 |
+
"grad_norm": 1.9593838453292847,
|
| 1445 |
+
"learning_rate": 6.852787187549182e-05,
|
| 1446 |
+
"loss": 0.9595,
|
| 1447 |
+
"step": 202
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 0.0026985619855035376,
|
| 1451 |
+
"grad_norm": 1.881966471672058,
|
| 1452 |
+
"learning_rate": 6.82235249939575e-05,
|
| 1453 |
+
"loss": 1.0891,
|
| 1454 |
+
"step": 203
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 0.0027118553942991216,
|
| 1458 |
+
"grad_norm": 1.5613300800323486,
|
| 1459 |
+
"learning_rate": 6.7918397477265e-05,
|
| 1460 |
+
"loss": 1.1484,
|
| 1461 |
+
"step": 204
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 0.0027251488030947055,
|
| 1465 |
+
"grad_norm": 1.6164745092391968,
|
| 1466 |
+
"learning_rate": 6.761250239606169e-05,
|
| 1467 |
+
"loss": 0.8214,
|
| 1468 |
+
"step": 205
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 0.0027384422118902894,
|
| 1472 |
+
"grad_norm": 1.4947420358657837,
|
| 1473 |
+
"learning_rate": 6.730585285387465e-05,
|
| 1474 |
+
"loss": 0.9956,
|
| 1475 |
+
"step": 206
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 0.0027517356206858733,
|
| 1479 |
+
"grad_norm": 1.6902318000793457,
|
| 1480 |
+
"learning_rate": 6.699846198654971e-05,
|
| 1481 |
+
"loss": 0.8679,
|
| 1482 |
+
"step": 207
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 0.0027650290294814572,
|
| 1486 |
+
"grad_norm": 1.6245845556259155,
|
| 1487 |
+
"learning_rate": 6.669034296168855e-05,
|
| 1488 |
+
"loss": 0.9362,
|
| 1489 |
+
"step": 208
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 0.002778322438277041,
|
| 1493 |
+
"grad_norm": 1.687902569770813,
|
| 1494 |
+
"learning_rate": 6.638150897808468e-05,
|
| 1495 |
+
"loss": 0.9086,
|
| 1496 |
+
"step": 209
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 0.002791615847072625,
|
| 1500 |
+
"grad_norm": 1.6738238334655762,
|
| 1501 |
+
"learning_rate": 6.607197326515808e-05,
|
| 1502 |
+
"loss": 1.0715,
|
| 1503 |
+
"step": 210
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 0.002804909255868209,
|
| 1507 |
+
"grad_norm": 1.8981541395187378,
|
| 1508 |
+
"learning_rate": 6.57617490823885e-05,
|
| 1509 |
+
"loss": 1.0964,
|
| 1510 |
+
"step": 211
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 0.002818202664663793,
|
| 1514 |
+
"grad_norm": 1.994768500328064,
|
| 1515 |
+
"learning_rate": 6.545084971874738e-05,
|
| 1516 |
+
"loss": 1.0326,
|
| 1517 |
+
"step": 212
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 0.002831496073459377,
|
| 1521 |
+
"grad_norm": 1.953397274017334,
|
| 1522 |
+
"learning_rate": 6.513928849212873e-05,
|
| 1523 |
+
"loss": 0.9181,
|
| 1524 |
+
"step": 213
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 0.002844789482254961,
|
| 1528 |
+
"grad_norm": 1.5838758945465088,
|
| 1529 |
+
"learning_rate": 6.482707874877854e-05,
|
| 1530 |
+
"loss": 1.0109,
|
| 1531 |
+
"step": 214
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 0.002858082891050545,
|
| 1535 |
+
"grad_norm": 2.204489231109619,
|
| 1536 |
+
"learning_rate": 6.451423386272312e-05,
|
| 1537 |
+
"loss": 0.9583,
|
| 1538 |
+
"step": 215
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 0.002871376299846129,
|
| 1542 |
+
"grad_norm": 1.679416537284851,
|
| 1543 |
+
"learning_rate": 6.420076723519614e-05,
|
| 1544 |
+
"loss": 1.0152,
|
| 1545 |
+
"step": 216
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 0.002884669708641713,
|
| 1549 |
+
"grad_norm": 2.062034845352173,
|
| 1550 |
+
"learning_rate": 6.388669229406462e-05,
|
| 1551 |
+
"loss": 0.9058,
|
| 1552 |
+
"step": 217
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 0.002897963117437297,
|
| 1556 |
+
"grad_norm": 1.9006659984588623,
|
| 1557 |
+
"learning_rate": 6.357202249325371e-05,
|
| 1558 |
+
"loss": 0.8065,
|
| 1559 |
+
"step": 218
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 0.0029112565262328807,
|
| 1563 |
+
"grad_norm": 1.9091770648956299,
|
| 1564 |
+
"learning_rate": 6.32567713121704e-05,
|
| 1565 |
+
"loss": 0.8315,
|
| 1566 |
+
"step": 219
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 0.0029245499350284646,
|
| 1570 |
+
"grad_norm": 2.309170722961426,
|
| 1571 |
+
"learning_rate": 6.294095225512603e-05,
|
| 1572 |
+
"loss": 0.9633,
|
| 1573 |
+
"step": 220
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 0.0029378433438240485,
|
| 1577 |
+
"grad_norm": 2.036591053009033,
|
| 1578 |
+
"learning_rate": 6.26245788507579e-05,
|
| 1579 |
+
"loss": 0.9114,
|
| 1580 |
+
"step": 221
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 0.0029511367526196325,
|
| 1584 |
+
"grad_norm": 2.540407419204712,
|
| 1585 |
+
"learning_rate": 6.230766465144967e-05,
|
| 1586 |
+
"loss": 0.8928,
|
| 1587 |
+
"step": 222
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 0.0029644301614152164,
|
| 1591 |
+
"grad_norm": 2.447540044784546,
|
| 1592 |
+
"learning_rate": 6.199022323275083e-05,
|
| 1593 |
+
"loss": 0.8958,
|
| 1594 |
+
"step": 223
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 0.0029777235702108003,
|
| 1598 |
+
"grad_norm": 2.1421115398406982,
|
| 1599 |
+
"learning_rate": 6.167226819279528e-05,
|
| 1600 |
+
"loss": 0.9212,
|
| 1601 |
+
"step": 224
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 0.002991016979006384,
|
| 1605 |
+
"grad_norm": 2.2318172454833984,
|
| 1606 |
+
"learning_rate": 6.135381315171867e-05,
|
| 1607 |
+
"loss": 0.7849,
|
| 1608 |
+
"step": 225
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 0.003004310387801968,
|
| 1612 |
+
"grad_norm": 2.107386350631714,
|
| 1613 |
+
"learning_rate": 6.103487175107507e-05,
|
| 1614 |
+
"loss": 0.8173,
|
| 1615 |
+
"step": 226
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 0.003017603796597552,
|
| 1619 |
+
"grad_norm": 2.08105206489563,
|
| 1620 |
+
"learning_rate": 6.071545765325254e-05,
|
| 1621 |
+
"loss": 0.9439,
|
| 1622 |
+
"step": 227
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 0.003030897205393136,
|
| 1626 |
+
"grad_norm": 2.376014471054077,
|
| 1627 |
+
"learning_rate": 6.0395584540887963e-05,
|
| 1628 |
+
"loss": 0.9148,
|
| 1629 |
+
"step": 228
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 0.00304419061418872,
|
| 1633 |
+
"grad_norm": 2.1427810192108154,
|
| 1634 |
+
"learning_rate": 6.007526611628086e-05,
|
| 1635 |
+
"loss": 0.8044,
|
| 1636 |
+
"step": 229
|
| 1637 |
+
},
|
| 1638 |
+
{
|
| 1639 |
+
"epoch": 0.0030574840229843038,
|
| 1640 |
+
"grad_norm": 2.6007277965545654,
|
| 1641 |
+
"learning_rate": 5.9754516100806423e-05,
|
| 1642 |
+
"loss": 0.8667,
|
| 1643 |
+
"step": 230
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 0.0030707774317798877,
|
| 1647 |
+
"grad_norm": 2.6047568321228027,
|
| 1648 |
+
"learning_rate": 5.9433348234327765e-05,
|
| 1649 |
+
"loss": 0.8761,
|
| 1650 |
+
"step": 231
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 0.0030840708405754716,
|
| 1654 |
+
"grad_norm": 2.6537301540374756,
|
| 1655 |
+
"learning_rate": 5.911177627460739e-05,
|
| 1656 |
+
"loss": 1.0008,
|
| 1657 |
+
"step": 232
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 0.0030973642493710555,
|
| 1661 |
+
"grad_norm": 2.5320510864257812,
|
| 1662 |
+
"learning_rate": 5.8789813996717736e-05,
|
| 1663 |
+
"loss": 0.9914,
|
| 1664 |
+
"step": 233
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 0.0031106576581666394,
|
| 1668 |
+
"grad_norm": 2.1913628578186035,
|
| 1669 |
+
"learning_rate": 5.8467475192451226e-05,
|
| 1670 |
+
"loss": 0.8323,
|
| 1671 |
+
"step": 234
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 0.0031239510669622233,
|
| 1675 |
+
"grad_norm": 2.716449737548828,
|
| 1676 |
+
"learning_rate": 5.814477366972945e-05,
|
| 1677 |
+
"loss": 0.885,
|
| 1678 |
+
"step": 235
|
| 1679 |
+
},
|
| 1680 |
+
{
|
| 1681 |
+
"epoch": 0.0031372444757578073,
|
| 1682 |
+
"grad_norm": 3.720883846282959,
|
| 1683 |
+
"learning_rate": 5.782172325201155e-05,
|
| 1684 |
+
"loss": 1.0299,
|
| 1685 |
+
"step": 236
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 0.003150537884553391,
|
| 1689 |
+
"grad_norm": 2.6181509494781494,
|
| 1690 |
+
"learning_rate": 5.749833777770225e-05,
|
| 1691 |
+
"loss": 0.8591,
|
| 1692 |
+
"step": 237
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 0.003163831293348975,
|
| 1696 |
+
"grad_norm": 2.382664918899536,
|
| 1697 |
+
"learning_rate": 5.717463109955896e-05,
|
| 1698 |
+
"loss": 0.8851,
|
| 1699 |
+
"step": 238
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 0.003177124702144559,
|
| 1703 |
+
"grad_norm": 2.3294265270233154,
|
| 1704 |
+
"learning_rate": 5.685061708409841e-05,
|
| 1705 |
+
"loss": 0.7961,
|
| 1706 |
+
"step": 239
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 0.003190418110940143,
|
| 1710 |
+
"grad_norm": 3.5375571250915527,
|
| 1711 |
+
"learning_rate": 5.6526309611002594e-05,
|
| 1712 |
+
"loss": 0.9115,
|
| 1713 |
+
"step": 240
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 0.003203711519735727,
|
| 1717 |
+
"grad_norm": 2.9233639240264893,
|
| 1718 |
+
"learning_rate": 5.6201722572524275e-05,
|
| 1719 |
+
"loss": 0.7504,
|
| 1720 |
+
"step": 241
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 0.003217004928531311,
|
| 1724 |
+
"grad_norm": 3.2264790534973145,
|
| 1725 |
+
"learning_rate": 5.587686987289189e-05,
|
| 1726 |
+
"loss": 0.7414,
|
| 1727 |
+
"step": 242
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 0.003230298337326895,
|
| 1731 |
+
"grad_norm": 2.921107530593872,
|
| 1732 |
+
"learning_rate": 5.5551765427713884e-05,
|
| 1733 |
+
"loss": 0.5761,
|
| 1734 |
+
"step": 243
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 0.003243591746122479,
|
| 1738 |
+
"grad_norm": 3.3718862533569336,
|
| 1739 |
+
"learning_rate": 5.522642316338268e-05,
|
| 1740 |
+
"loss": 0.8256,
|
| 1741 |
+
"step": 244
|
| 1742 |
+
},
|
| 1743 |
+
{
|
| 1744 |
+
"epoch": 0.003256885154918063,
|
| 1745 |
+
"grad_norm": 2.911158800125122,
|
| 1746 |
+
"learning_rate": 5.490085701647805e-05,
|
| 1747 |
+
"loss": 0.7299,
|
| 1748 |
+
"step": 245
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 0.003270178563713647,
|
| 1752 |
+
"grad_norm": 3.8199920654296875,
|
| 1753 |
+
"learning_rate": 5.457508093317013e-05,
|
| 1754 |
+
"loss": 0.7782,
|
| 1755 |
+
"step": 246
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 0.0032834719725092307,
|
| 1759 |
+
"grad_norm": 3.6611592769622803,
|
| 1760 |
+
"learning_rate": 5.4249108868622086e-05,
|
| 1761 |
+
"loss": 0.7346,
|
| 1762 |
+
"step": 247
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 0.0032967653813048147,
|
| 1766 |
+
"grad_norm": 3.3983826637268066,
|
| 1767 |
+
"learning_rate": 5.392295478639225e-05,
|
| 1768 |
+
"loss": 0.68,
|
| 1769 |
+
"step": 248
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 0.0033100587901003986,
|
| 1773 |
+
"grad_norm": 4.487311840057373,
|
| 1774 |
+
"learning_rate": 5.359663265783598e-05,
|
| 1775 |
+
"loss": 0.7677,
|
| 1776 |
+
"step": 249
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 0.0033233521988959825,
|
| 1780 |
+
"grad_norm": 5.948139667510986,
|
| 1781 |
+
"learning_rate": 5.327015646150716e-05,
|
| 1782 |
+
"loss": 0.7702,
|
| 1783 |
+
"step": 250
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 0.0033366456076915664,
|
| 1787 |
+
"grad_norm": 1.1044285297393799,
|
| 1788 |
+
"learning_rate": 5.294354018255945e-05,
|
| 1789 |
+
"loss": 0.9482,
|
| 1790 |
+
"step": 251
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 0.0033499390164871503,
|
| 1794 |
+
"grad_norm": 1.6621804237365723,
|
| 1795 |
+
"learning_rate": 5.26167978121472e-05,
|
| 1796 |
+
"loss": 0.8978,
|
| 1797 |
+
"step": 252
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 0.0033632324252827342,
|
| 1801 |
+
"grad_norm": 1.7893524169921875,
|
| 1802 |
+
"learning_rate": 5.228994334682604e-05,
|
| 1803 |
+
"loss": 0.9667,
|
| 1804 |
+
"step": 253
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 0.003376525834078318,
|
| 1808 |
+
"grad_norm": 1.6742503643035889,
|
| 1809 |
+
"learning_rate": 5.196299078795344e-05,
|
| 1810 |
+
"loss": 0.8508,
|
| 1811 |
+
"step": 254
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 0.003389819242873902,
|
| 1815 |
+
"grad_norm": 1.6336678266525269,
|
| 1816 |
+
"learning_rate": 5.1635954141088813e-05,
|
| 1817 |
+
"loss": 1.0183,
|
| 1818 |
+
"step": 255
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 0.003403112651669486,
|
| 1822 |
+
"grad_norm": 1.4634300470352173,
|
| 1823 |
+
"learning_rate": 5.1308847415393666e-05,
|
| 1824 |
+
"loss": 0.8325,
|
| 1825 |
+
"step": 256
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 0.00341640606046507,
|
| 1829 |
+
"grad_norm": 1.8752628564834595,
|
| 1830 |
+
"learning_rate": 5.0981684623031415e-05,
|
| 1831 |
+
"loss": 0.8682,
|
| 1832 |
+
"step": 257
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 0.003429699469260654,
|
| 1836 |
+
"grad_norm": 1.516921043395996,
|
| 1837 |
+
"learning_rate": 5.0654479778567223e-05,
|
| 1838 |
+
"loss": 0.87,
|
| 1839 |
+
"step": 258
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 0.0034429928780562377,
|
| 1843 |
+
"grad_norm": 1.953859567642212,
|
| 1844 |
+
"learning_rate": 5.0327246898367597e-05,
|
| 1845 |
+
"loss": 1.0406,
|
| 1846 |
+
"step": 259
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 0.0034562862868518216,
|
| 1850 |
+
"grad_norm": 1.692104697227478,
|
| 1851 |
+
"learning_rate": 5e-05,
|
| 1852 |
+
"loss": 1.0228,
|
| 1853 |
+
"step": 260
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 0.0034695796956474055,
|
| 1857 |
+
"grad_norm": 1.6310532093048096,
|
| 1858 |
+
"learning_rate": 4.9672753101632415e-05,
|
| 1859 |
+
"loss": 1.0342,
|
| 1860 |
+
"step": 261
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 0.0034828731044429895,
|
| 1864 |
+
"grad_norm": 1.8367486000061035,
|
| 1865 |
+
"learning_rate": 4.934552022143279e-05,
|
| 1866 |
+
"loss": 0.8968,
|
| 1867 |
+
"step": 262
|
| 1868 |
+
},
|
| 1869 |
+
{
|
| 1870 |
+
"epoch": 0.0034961665132385734,
|
| 1871 |
+
"grad_norm": 1.6751683950424194,
|
| 1872 |
+
"learning_rate": 4.901831537696859e-05,
|
| 1873 |
+
"loss": 0.9189,
|
| 1874 |
+
"step": 263
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.0035094599220341573,
|
| 1878 |
+
"grad_norm": 1.8059558868408203,
|
| 1879 |
+
"learning_rate": 4.869115258460635e-05,
|
| 1880 |
+
"loss": 0.9239,
|
| 1881 |
+
"step": 264
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 0.003522753330829741,
|
| 1885 |
+
"grad_norm": 1.7002429962158203,
|
| 1886 |
+
"learning_rate": 4.83640458589112e-05,
|
| 1887 |
+
"loss": 0.8102,
|
| 1888 |
+
"step": 265
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 0.003536046739625325,
|
| 1892 |
+
"grad_norm": 1.738466501235962,
|
| 1893 |
+
"learning_rate": 4.8037009212046586e-05,
|
| 1894 |
+
"loss": 1.0022,
|
| 1895 |
+
"step": 266
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 0.003549340148420909,
|
| 1899 |
+
"grad_norm": 2.1397879123687744,
|
| 1900 |
+
"learning_rate": 4.7710056653173976e-05,
|
| 1901 |
+
"loss": 0.8882,
|
| 1902 |
+
"step": 267
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 0.003562633557216493,
|
| 1906 |
+
"grad_norm": 2.1561543941497803,
|
| 1907 |
+
"learning_rate": 4.738320218785281e-05,
|
| 1908 |
+
"loss": 0.9589,
|
| 1909 |
+
"step": 268
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 0.003575926966012077,
|
| 1913 |
+
"grad_norm": 2.116396427154541,
|
| 1914 |
+
"learning_rate": 4.7056459817440544e-05,
|
| 1915 |
+
"loss": 0.8827,
|
| 1916 |
+
"step": 269
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 0.003589220374807661,
|
| 1920 |
+
"grad_norm": 2.172356367111206,
|
| 1921 |
+
"learning_rate": 4.6729843538492847e-05,
|
| 1922 |
+
"loss": 0.9613,
|
| 1923 |
+
"step": 270
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 0.003602513783603245,
|
| 1927 |
+
"grad_norm": 1.7960728406906128,
|
| 1928 |
+
"learning_rate": 4.640336734216403e-05,
|
| 1929 |
+
"loss": 0.8781,
|
| 1930 |
+
"step": 271
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 0.003615807192398829,
|
| 1934 |
+
"grad_norm": 1.8511698246002197,
|
| 1935 |
+
"learning_rate": 4.607704521360776e-05,
|
| 1936 |
+
"loss": 0.8647,
|
| 1937 |
+
"step": 272
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 0.003629100601194413,
|
| 1941 |
+
"grad_norm": 2.0229597091674805,
|
| 1942 |
+
"learning_rate": 4.575089113137792e-05,
|
| 1943 |
+
"loss": 1.004,
|
| 1944 |
+
"step": 273
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 0.003642394009989997,
|
| 1948 |
+
"grad_norm": 2.3480939865112305,
|
| 1949 |
+
"learning_rate": 4.542491906682989e-05,
|
| 1950 |
+
"loss": 1.0119,
|
| 1951 |
+
"step": 274
|
| 1952 |
+
},
|
| 1953 |
+
{
|
| 1954 |
+
"epoch": 0.0036556874187855808,
|
| 1955 |
+
"grad_norm": 1.9140769243240356,
|
| 1956 |
+
"learning_rate": 4.509914298352197e-05,
|
| 1957 |
+
"loss": 0.8089,
|
| 1958 |
+
"step": 275
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 0.0036689808275811647,
|
| 1962 |
+
"grad_norm": 1.9026778936386108,
|
| 1963 |
+
"learning_rate": 4.477357683661734e-05,
|
| 1964 |
+
"loss": 0.7569,
|
| 1965 |
+
"step": 276
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 0.0036822742363767486,
|
| 1969 |
+
"grad_norm": 2.2026915550231934,
|
| 1970 |
+
"learning_rate": 4.444823457228612e-05,
|
| 1971 |
+
"loss": 0.8855,
|
| 1972 |
+
"step": 277
|
| 1973 |
+
},
|
| 1974 |
+
{
|
| 1975 |
+
"epoch": 0.0036955676451723325,
|
| 1976 |
+
"grad_norm": 2.2045536041259766,
|
| 1977 |
+
"learning_rate": 4.412313012710813e-05,
|
| 1978 |
+
"loss": 0.8306,
|
| 1979 |
+
"step": 278
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 0.0037088610539679164,
|
| 1983 |
+
"grad_norm": 2.3194525241851807,
|
| 1984 |
+
"learning_rate": 4.379827742747575e-05,
|
| 1985 |
+
"loss": 0.8154,
|
| 1986 |
+
"step": 279
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 0.0037221544627635004,
|
| 1990 |
+
"grad_norm": 2.6129322052001953,
|
| 1991 |
+
"learning_rate": 4.347369038899744e-05,
|
| 1992 |
+
"loss": 0.9187,
|
| 1993 |
+
"step": 280
|
| 1994 |
+
},
|
| 1995 |
+
{
|
| 1996 |
+
"epoch": 0.0037354478715590843,
|
| 1997 |
+
"grad_norm": 2.2725565433502197,
|
| 1998 |
+
"learning_rate": 4.3149382915901606e-05,
|
| 1999 |
+
"loss": 0.793,
|
| 2000 |
+
"step": 281
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 0.003748741280354668,
|
| 2004 |
+
"grad_norm": 2.23557186126709,
|
| 2005 |
+
"learning_rate": 4.282536890044104e-05,
|
| 2006 |
+
"loss": 0.8578,
|
| 2007 |
+
"step": 282
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 0.003762034689150252,
|
| 2011 |
+
"grad_norm": 2.394350290298462,
|
| 2012 |
+
"learning_rate": 4.250166222229774e-05,
|
| 2013 |
+
"loss": 0.8555,
|
| 2014 |
+
"step": 283
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 0.003775328097945836,
|
| 2018 |
+
"grad_norm": 2.935340642929077,
|
| 2019 |
+
"learning_rate": 4.2178276747988446e-05,
|
| 2020 |
+
"loss": 0.852,
|
| 2021 |
+
"step": 284
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 0.00378862150674142,
|
| 2025 |
+
"grad_norm": 3.061005115509033,
|
| 2026 |
+
"learning_rate": 4.185522633027057e-05,
|
| 2027 |
+
"loss": 0.8782,
|
| 2028 |
+
"step": 285
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 0.003801914915537004,
|
| 2032 |
+
"grad_norm": 2.7278201580047607,
|
| 2033 |
+
"learning_rate": 4.153252480754877e-05,
|
| 2034 |
+
"loss": 0.8451,
|
| 2035 |
+
"step": 286
|
| 2036 |
+
},
|
| 2037 |
+
{
|
| 2038 |
+
"epoch": 0.0038152083243325878,
|
| 2039 |
+
"grad_norm": 2.643934488296509,
|
| 2040 |
+
"learning_rate": 4.1210186003282275e-05,
|
| 2041 |
+
"loss": 0.7744,
|
| 2042 |
+
"step": 287
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 0.0038285017331281717,
|
| 2046 |
+
"grad_norm": 3.218041181564331,
|
| 2047 |
+
"learning_rate": 4.088822372539263e-05,
|
| 2048 |
+
"loss": 0.8261,
|
| 2049 |
+
"step": 288
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 0.0038417951419237556,
|
| 2053 |
+
"grad_norm": 2.650805950164795,
|
| 2054 |
+
"learning_rate": 4.0566651765672246e-05,
|
| 2055 |
+
"loss": 0.8542,
|
| 2056 |
+
"step": 289
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 0.0038550885507193395,
|
| 2060 |
+
"grad_norm": 2.679455041885376,
|
| 2061 |
+
"learning_rate": 4.0245483899193595e-05,
|
| 2062 |
+
"loss": 0.8374,
|
| 2063 |
+
"step": 290
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 0.0038683819595149234,
|
| 2067 |
+
"grad_norm": 3.0015034675598145,
|
| 2068 |
+
"learning_rate": 3.992473388371915e-05,
|
| 2069 |
+
"loss": 0.9829,
|
| 2070 |
+
"step": 291
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 0.0038816753683105073,
|
| 2074 |
+
"grad_norm": 3.549671173095703,
|
| 2075 |
+
"learning_rate": 3.960441545911204e-05,
|
| 2076 |
+
"loss": 0.8356,
|
| 2077 |
+
"step": 292
|
| 2078 |
+
},
|
| 2079 |
+
{
|
| 2080 |
+
"epoch": 0.0038949687771060912,
|
| 2081 |
+
"grad_norm": 3.1017885208129883,
|
| 2082 |
+
"learning_rate": 3.928454234674747e-05,
|
| 2083 |
+
"loss": 0.8398,
|
| 2084 |
+
"step": 293
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 0.003908262185901675,
|
| 2088 |
+
"grad_norm": 3.0585556030273438,
|
| 2089 |
+
"learning_rate": 3.896512824892495e-05,
|
| 2090 |
+
"loss": 0.7647,
|
| 2091 |
+
"step": 294
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 0.0039215555946972595,
|
| 2095 |
+
"grad_norm": 3.385533571243286,
|
| 2096 |
+
"learning_rate": 3.864618684828134e-05,
|
| 2097 |
+
"loss": 0.6782,
|
| 2098 |
+
"step": 295
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 0.003934849003492843,
|
| 2102 |
+
"grad_norm": 5.169503211975098,
|
| 2103 |
+
"learning_rate": 3.832773180720475e-05,
|
| 2104 |
+
"loss": 0.7389,
|
| 2105 |
+
"step": 296
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 0.003948142412288427,
|
| 2109 |
+
"grad_norm": 2.9151158332824707,
|
| 2110 |
+
"learning_rate": 3.800977676724919e-05,
|
| 2111 |
+
"loss": 0.6878,
|
| 2112 |
+
"step": 297
|
| 2113 |
+
},
|
| 2114 |
+
{
|
| 2115 |
+
"epoch": 0.003961435821084011,
|
| 2116 |
+
"grad_norm": 3.42378568649292,
|
| 2117 |
+
"learning_rate": 3.769233534855035e-05,
|
| 2118 |
+
"loss": 0.6473,
|
| 2119 |
+
"step": 298
|
| 2120 |
+
},
|
| 2121 |
+
{
|
| 2122 |
+
"epoch": 0.003974729229879595,
|
| 2123 |
+
"grad_norm": 4.071504592895508,
|
| 2124 |
+
"learning_rate": 3.73754211492421e-05,
|
| 2125 |
+
"loss": 0.8104,
|
| 2126 |
+
"step": 299
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 0.003988022638675179,
|
| 2130 |
+
"grad_norm": 5.344113349914551,
|
| 2131 |
+
"learning_rate": 3.705904774487396e-05,
|
| 2132 |
+
"loss": 0.9578,
|
| 2133 |
+
"step": 300
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.003988022638675179,
|
| 2137 |
+
"eval_loss": 0.8697348237037659,
|
| 2138 |
+
"eval_runtime": 6447.6628,
|
| 2139 |
+
"eval_samples_per_second": 19.65,
|
| 2140 |
+
"eval_steps_per_second": 4.912,
|
| 2141 |
+
"step": 300
|
| 2142 |
}
|
| 2143 |
],
|
| 2144 |
"logging_steps": 1,
|
|
|
|
| 2167 |
"attributes": {}
|
| 2168 |
}
|
| 2169 |
},
|
| 2170 |
+
"total_flos": 1.880597072801956e+17,
|
| 2171 |
"train_batch_size": 8,
|
| 2172 |
"trial_name": null,
|
| 2173 |
"trial_params": null
|