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 161533192
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b2d83d5f901eae11145cacd89c72231a010785bff7439a1ba5f4a5e94b8052b4
|
| 3 |
size 161533192
|
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 sha256:1daa0c2c790a56b21b4c6c41ad6a3c0d997b0d94c85b459b9bd582601d994cca
|
| 3 |
+
size 82461044
|
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 sha256:89f940dbe392540bac79082367808c07a597a4f60dff698b2c1ee1c27e1aabed
|
| 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:fd8039035e9f22cc9e9230f66e4f1f1db2add3119ff8edfe450ef63eb16e5439
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 1.
|
| 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": 14.017,
|
| 1432 |
"eval_steps_per_second": 3.505,
|
| 1433 |
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1434 |
}
|
| 1435 |
],
|
| 1436 |
"logging_steps": 1,
|
|
@@ -1459,7 +2167,7 @@
|
|
| 1459 |
"attributes": {}
|
| 1460 |
}
|
| 1461 |
},
|
| 1462 |
-
"total_flos":
|
| 1463 |
"train_batch_size": 4,
|
| 1464 |
"trial_name": null,
|
| 1465 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 1.700640082359314,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-300",
|
| 4 |
+
"epoch": 0.11954274898513187,
|
| 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": 14.017,
|
| 1432 |
"eval_steps_per_second": 3.505,
|
| 1433 |
"step": 200
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 0.08009364182003835,
|
| 1437 |
+
"grad_norm": 0.5120438933372498,
|
| 1438 |
+
"learning_rate": 0.00011964121814957137,
|
| 1439 |
+
"loss": 1.7626,
|
| 1440 |
+
"step": 201
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 0.08049211764998879,
|
| 1444 |
+
"grad_norm": 0.5081213712692261,
|
| 1445 |
+
"learning_rate": 0.00011893581990651848,
|
| 1446 |
+
"loss": 1.7664,
|
| 1447 |
+
"step": 202
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 0.08089059347993924,
|
| 1451 |
+
"grad_norm": 0.45303136110305786,
|
| 1452 |
+
"learning_rate": 0.00011822944303213486,
|
| 1453 |
+
"loss": 1.4845,
|
| 1454 |
+
"step": 203
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 0.08128906930988967,
|
| 1458 |
+
"grad_norm": 0.4521328806877136,
|
| 1459 |
+
"learning_rate": 0.00011752212403302784,
|
| 1460 |
+
"loss": 1.4534,
|
| 1461 |
+
"step": 204
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 0.08168754513984011,
|
| 1465 |
+
"grad_norm": 0.4870966970920563,
|
| 1466 |
+
"learning_rate": 0.00011681389946449504,
|
| 1467 |
+
"loss": 1.568,
|
| 1468 |
+
"step": 205
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 0.08208602096979055,
|
| 1472 |
+
"grad_norm": 0.5694287419319153,
|
| 1473 |
+
"learning_rate": 0.00011610480592863531,
|
| 1474 |
+
"loss": 1.9597,
|
| 1475 |
+
"step": 206
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 0.08248449679974099,
|
| 1479 |
+
"grad_norm": 0.5484179258346558,
|
| 1480 |
+
"learning_rate": 0.00011539488007245702,
|
| 1481 |
+
"loss": 1.8557,
|
| 1482 |
+
"step": 207
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 0.08288297262969144,
|
| 1486 |
+
"grad_norm": 0.44166100025177,
|
| 1487 |
+
"learning_rate": 0.00011468415858598411,
|
| 1488 |
+
"loss": 1.5371,
|
| 1489 |
+
"step": 208
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 0.08328144845964187,
|
| 1493 |
+
"grad_norm": 0.5365661978721619,
|
| 1494 |
+
"learning_rate": 0.00011397267820035986,
|
| 1495 |
+
"loss": 1.7778,
|
| 1496 |
+
"step": 209
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 0.08367992428959231,
|
| 1500 |
+
"grad_norm": 0.45911017060279846,
|
| 1501 |
+
"learning_rate": 0.00011326047568594851,
|
| 1502 |
+
"loss": 1.5729,
|
| 1503 |
+
"step": 210
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 0.08407840011954275,
|
| 1507 |
+
"grad_norm": 0.5417599678039551,
|
| 1508 |
+
"learning_rate": 0.00011254758785043515,
|
| 1509 |
+
"loss": 1.8296,
|
| 1510 |
+
"step": 211
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 0.08447687594949319,
|
| 1514 |
+
"grad_norm": 0.48594942688941956,
|
| 1515 |
+
"learning_rate": 0.0001118340515369232,
|
| 1516 |
+
"loss": 1.7837,
|
| 1517 |
+
"step": 212
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 0.08487535177944362,
|
| 1521 |
+
"grad_norm": 0.4888298511505127,
|
| 1522 |
+
"learning_rate": 0.00011111990362203033,
|
| 1523 |
+
"loss": 1.6575,
|
| 1524 |
+
"step": 213
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 0.08527382760939407,
|
| 1528 |
+
"grad_norm": 0.5313907265663147,
|
| 1529 |
+
"learning_rate": 0.00011040518101398276,
|
| 1530 |
+
"loss": 1.7803,
|
| 1531 |
+
"step": 214
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 0.08567230343934451,
|
| 1535 |
+
"grad_norm": 0.5065906643867493,
|
| 1536 |
+
"learning_rate": 0.00010968992065070769,
|
| 1537 |
+
"loss": 1.6539,
|
| 1538 |
+
"step": 215
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 0.08607077926929495,
|
| 1542 |
+
"grad_norm": 0.46294957399368286,
|
| 1543 |
+
"learning_rate": 0.00010897415949792427,
|
| 1544 |
+
"loss": 1.6412,
|
| 1545 |
+
"step": 216
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 0.08646925509924538,
|
| 1549 |
+
"grad_norm": 0.5068647861480713,
|
| 1550 |
+
"learning_rate": 0.00010825793454723325,
|
| 1551 |
+
"loss": 1.7173,
|
| 1552 |
+
"step": 217
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 0.08686773092919582,
|
| 1556 |
+
"grad_norm": 0.45219966769218445,
|
| 1557 |
+
"learning_rate": 0.0001075412828142051,
|
| 1558 |
+
"loss": 1.4531,
|
| 1559 |
+
"step": 218
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 0.08726620675914626,
|
| 1563 |
+
"grad_norm": 0.48035022616386414,
|
| 1564 |
+
"learning_rate": 0.0001068242413364671,
|
| 1565 |
+
"loss": 1.6187,
|
| 1566 |
+
"step": 219
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 0.08766468258909671,
|
| 1570 |
+
"grad_norm": 0.5463985204696655,
|
| 1571 |
+
"learning_rate": 0.00010610684717178905,
|
| 1572 |
+
"loss": 1.749,
|
| 1573 |
+
"step": 220
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 0.08806315841904715,
|
| 1577 |
+
"grad_norm": 0.4818764626979828,
|
| 1578 |
+
"learning_rate": 0.00010538913739616816,
|
| 1579 |
+
"loss": 1.4508,
|
| 1580 |
+
"step": 221
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 0.08846163424899758,
|
| 1584 |
+
"grad_norm": 0.5018213987350464,
|
| 1585 |
+
"learning_rate": 0.00010467114910191289,
|
| 1586 |
+
"loss": 1.6853,
|
| 1587 |
+
"step": 222
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 0.08886011007894802,
|
| 1591 |
+
"grad_norm": 0.5122075080871582,
|
| 1592 |
+
"learning_rate": 0.00010395291939572593,
|
| 1593 |
+
"loss": 1.6991,
|
| 1594 |
+
"step": 223
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 0.08925858590889846,
|
| 1598 |
+
"grad_norm": 0.48191148042678833,
|
| 1599 |
+
"learning_rate": 0.00010323448539678653,
|
| 1600 |
+
"loss": 1.6428,
|
| 1601 |
+
"step": 224
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 0.08965706173884891,
|
| 1605 |
+
"grad_norm": 0.4748276472091675,
|
| 1606 |
+
"learning_rate": 0.00010251588423483205,
|
| 1607 |
+
"loss": 1.7059,
|
| 1608 |
+
"step": 225
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 0.09005553756879935,
|
| 1612 |
+
"grad_norm": 0.5150067806243896,
|
| 1613 |
+
"learning_rate": 0.0001017971530482392,
|
| 1614 |
+
"loss": 1.7409,
|
| 1615 |
+
"step": 226
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 0.09045401339874978,
|
| 1619 |
+
"grad_norm": 0.5893855094909668,
|
| 1620 |
+
"learning_rate": 0.00010107832898210439,
|
| 1621 |
+
"loss": 1.7183,
|
| 1622 |
+
"step": 227
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 0.09085248922870022,
|
| 1626 |
+
"grad_norm": 0.5195055603981018,
|
| 1627 |
+
"learning_rate": 0.00010035944918632429,
|
| 1628 |
+
"loss": 1.8396,
|
| 1629 |
+
"step": 228
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 0.09125096505865066,
|
| 1633 |
+
"grad_norm": 0.5996953845024109,
|
| 1634 |
+
"learning_rate": 9.96405508136757e-05,
|
| 1635 |
+
"loss": 1.9944,
|
| 1636 |
+
"step": 229
|
| 1637 |
+
},
|
| 1638 |
+
{
|
| 1639 |
+
"epoch": 0.0916494408886011,
|
| 1640 |
+
"grad_norm": 0.5057780146598816,
|
| 1641 |
+
"learning_rate": 9.892167101789564e-05,
|
| 1642 |
+
"loss": 1.6473,
|
| 1643 |
+
"step": 230
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 0.09204791671855155,
|
| 1647 |
+
"grad_norm": 0.46774283051490784,
|
| 1648 |
+
"learning_rate": 9.820284695176082e-05,
|
| 1649 |
+
"loss": 1.5973,
|
| 1650 |
+
"step": 231
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 0.09244639254850198,
|
| 1654 |
+
"grad_norm": 0.46982142329216003,
|
| 1655 |
+
"learning_rate": 9.748411576516794e-05,
|
| 1656 |
+
"loss": 1.6464,
|
| 1657 |
+
"step": 232
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 0.09284486837845242,
|
| 1661 |
+
"grad_norm": 0.4873621165752411,
|
| 1662 |
+
"learning_rate": 9.676551460321349e-05,
|
| 1663 |
+
"loss": 1.6629,
|
| 1664 |
+
"step": 233
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 0.09324334420840286,
|
| 1668 |
+
"grad_norm": 0.4866909682750702,
|
| 1669 |
+
"learning_rate": 9.60470806042741e-05,
|
| 1670 |
+
"loss": 1.6262,
|
| 1671 |
+
"step": 234
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 0.0936418200383533,
|
| 1675 |
+
"grad_norm": 0.5320809483528137,
|
| 1676 |
+
"learning_rate": 9.532885089808713e-05,
|
| 1677 |
+
"loss": 1.7158,
|
| 1678 |
+
"step": 235
|
| 1679 |
+
},
|
| 1680 |
+
{
|
| 1681 |
+
"epoch": 0.09404029586830374,
|
| 1682 |
+
"grad_norm": 0.47346270084381104,
|
| 1683 |
+
"learning_rate": 9.461086260383187e-05,
|
| 1684 |
+
"loss": 1.6044,
|
| 1685 |
+
"step": 236
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 0.09443877169825418,
|
| 1689 |
+
"grad_norm": 0.5696609616279602,
|
| 1690 |
+
"learning_rate": 9.389315282821097e-05,
|
| 1691 |
+
"loss": 1.7883,
|
| 1692 |
+
"step": 237
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 0.09483724752820462,
|
| 1696 |
+
"grad_norm": 0.4926949441432953,
|
| 1697 |
+
"learning_rate": 9.317575866353292e-05,
|
| 1698 |
+
"loss": 1.7306,
|
| 1699 |
+
"step": 238
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 0.09523572335815506,
|
| 1703 |
+
"grad_norm": 0.5241943001747131,
|
| 1704 |
+
"learning_rate": 9.245871718579491e-05,
|
| 1705 |
+
"loss": 1.732,
|
| 1706 |
+
"step": 239
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 0.09563419918810549,
|
| 1710 |
+
"grad_norm": 0.5425236225128174,
|
| 1711 |
+
"learning_rate": 9.174206545276677e-05,
|
| 1712 |
+
"loss": 1.6209,
|
| 1713 |
+
"step": 240
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 0.09603267501805593,
|
| 1717 |
+
"grad_norm": 0.5216458439826965,
|
| 1718 |
+
"learning_rate": 9.102584050207578e-05,
|
| 1719 |
+
"loss": 1.74,
|
| 1720 |
+
"step": 241
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 0.09643115084800638,
|
| 1724 |
+
"grad_norm": 0.5082316994667053,
|
| 1725 |
+
"learning_rate": 9.031007934929236e-05,
|
| 1726 |
+
"loss": 1.6836,
|
| 1727 |
+
"step": 242
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 0.09682962667795682,
|
| 1731 |
+
"grad_norm": 0.48965132236480713,
|
| 1732 |
+
"learning_rate": 8.959481898601728e-05,
|
| 1733 |
+
"loss": 1.7055,
|
| 1734 |
+
"step": 243
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 0.09722810250790725,
|
| 1738 |
+
"grad_norm": 0.514946699142456,
|
| 1739 |
+
"learning_rate": 8.888009637796968e-05,
|
| 1740 |
+
"loss": 1.684,
|
| 1741 |
+
"step": 244
|
| 1742 |
+
},
|
| 1743 |
+
{
|
| 1744 |
+
"epoch": 0.09762657833785769,
|
| 1745 |
+
"grad_norm": 0.551802396774292,
|
| 1746 |
+
"learning_rate": 8.81659484630768e-05,
|
| 1747 |
+
"loss": 1.8566,
|
| 1748 |
+
"step": 245
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 0.09802505416780813,
|
| 1752 |
+
"grad_norm": 0.4790934920310974,
|
| 1753 |
+
"learning_rate": 8.745241214956483e-05,
|
| 1754 |
+
"loss": 1.6461,
|
| 1755 |
+
"step": 246
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 0.09842352999775858,
|
| 1759 |
+
"grad_norm": 0.5450412631034851,
|
| 1760 |
+
"learning_rate": 8.673952431405148e-05,
|
| 1761 |
+
"loss": 1.7215,
|
| 1762 |
+
"step": 247
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 0.09882200582770902,
|
| 1766 |
+
"grad_norm": 0.5299497842788696,
|
| 1767 |
+
"learning_rate": 8.602732179964017e-05,
|
| 1768 |
+
"loss": 1.7454,
|
| 1769 |
+
"step": 248
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 0.09922048165765945,
|
| 1773 |
+
"grad_norm": 0.5010784268379211,
|
| 1774 |
+
"learning_rate": 8.531584141401591e-05,
|
| 1775 |
+
"loss": 1.6028,
|
| 1776 |
+
"step": 249
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 0.09961895748760989,
|
| 1780 |
+
"grad_norm": 0.4926188886165619,
|
| 1781 |
+
"learning_rate": 8.4605119927543e-05,
|
| 1782 |
+
"loss": 1.6837,
|
| 1783 |
+
"step": 250
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 0.10001743331756033,
|
| 1787 |
+
"grad_norm": 0.5703017115592957,
|
| 1788 |
+
"learning_rate": 8.38951940713647e-05,
|
| 1789 |
+
"loss": 1.8639,
|
| 1790 |
+
"step": 251
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 0.10041590914751077,
|
| 1794 |
+
"grad_norm": 0.5429261326789856,
|
| 1795 |
+
"learning_rate": 8.318610053550497e-05,
|
| 1796 |
+
"loss": 1.7258,
|
| 1797 |
+
"step": 252
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 0.10081438497746122,
|
| 1801 |
+
"grad_norm": 0.48338782787323,
|
| 1802 |
+
"learning_rate": 8.247787596697218e-05,
|
| 1803 |
+
"loss": 1.5873,
|
| 1804 |
+
"step": 253
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 0.10121286080741165,
|
| 1808 |
+
"grad_norm": 0.506877601146698,
|
| 1809 |
+
"learning_rate": 8.177055696786516e-05,
|
| 1810 |
+
"loss": 1.6736,
|
| 1811 |
+
"step": 254
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 0.10161133663736209,
|
| 1815 |
+
"grad_norm": 0.537820041179657,
|
| 1816 |
+
"learning_rate": 8.106418009348157e-05,
|
| 1817 |
+
"loss": 1.9075,
|
| 1818 |
+
"step": 255
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 0.10200981246731253,
|
| 1822 |
+
"grad_norm": 0.4729152023792267,
|
| 1823 |
+
"learning_rate": 8.035878185042868e-05,
|
| 1824 |
+
"loss": 1.5359,
|
| 1825 |
+
"step": 256
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 0.10240828829726296,
|
| 1829 |
+
"grad_norm": 0.4413747191429138,
|
| 1830 |
+
"learning_rate": 7.965439869473664e-05,
|
| 1831 |
+
"loss": 1.6245,
|
| 1832 |
+
"step": 257
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 0.10280676412721342,
|
| 1836 |
+
"grad_norm": 0.5398510694503784,
|
| 1837 |
+
"learning_rate": 7.895106702997437e-05,
|
| 1838 |
+
"loss": 1.6318,
|
| 1839 |
+
"step": 258
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 0.10320523995716385,
|
| 1843 |
+
"grad_norm": 0.5172785520553589,
|
| 1844 |
+
"learning_rate": 7.824882320536814e-05,
|
| 1845 |
+
"loss": 1.6601,
|
| 1846 |
+
"step": 259
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 0.10360371578711429,
|
| 1850 |
+
"grad_norm": 0.4824993908405304,
|
| 1851 |
+
"learning_rate": 7.754770351392311e-05,
|
| 1852 |
+
"loss": 1.5672,
|
| 1853 |
+
"step": 260
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 0.10400219161706473,
|
| 1857 |
+
"grad_norm": 0.4745709300041199,
|
| 1858 |
+
"learning_rate": 7.684774419054747e-05,
|
| 1859 |
+
"loss": 1.7128,
|
| 1860 |
+
"step": 261
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 0.10440066744701516,
|
| 1864 |
+
"grad_norm": 0.5071855783462524,
|
| 1865 |
+
"learning_rate": 7.614898141017996e-05,
|
| 1866 |
+
"loss": 1.7368,
|
| 1867 |
+
"step": 262
|
| 1868 |
+
},
|
| 1869 |
+
{
|
| 1870 |
+
"epoch": 0.1047991432769656,
|
| 1871 |
+
"grad_norm": 0.5377690196037292,
|
| 1872 |
+
"learning_rate": 7.54514512859201e-05,
|
| 1873 |
+
"loss": 1.8659,
|
| 1874 |
+
"step": 263
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.10519761910691605,
|
| 1878 |
+
"grad_norm": 0.4762866199016571,
|
| 1879 |
+
"learning_rate": 7.475518986716194e-05,
|
| 1880 |
+
"loss": 1.6012,
|
| 1881 |
+
"step": 264
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 0.10559609493686649,
|
| 1885 |
+
"grad_norm": 0.46296924352645874,
|
| 1886 |
+
"learning_rate": 7.406023313773097e-05,
|
| 1887 |
+
"loss": 1.5484,
|
| 1888 |
+
"step": 265
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 0.10599457076681693,
|
| 1892 |
+
"grad_norm": 0.47845426201820374,
|
| 1893 |
+
"learning_rate": 7.336661701402439e-05,
|
| 1894 |
+
"loss": 1.6248,
|
| 1895 |
+
"step": 266
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 0.10639304659676736,
|
| 1899 |
+
"grad_norm": 0.48351001739501953,
|
| 1900 |
+
"learning_rate": 7.267437734315492e-05,
|
| 1901 |
+
"loss": 1.5549,
|
| 1902 |
+
"step": 267
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 0.1067915224267178,
|
| 1906 |
+
"grad_norm": 0.48554375767707825,
|
| 1907 |
+
"learning_rate": 7.198354990109805e-05,
|
| 1908 |
+
"loss": 1.5708,
|
| 1909 |
+
"step": 268
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 0.10718999825666825,
|
| 1913 |
+
"grad_norm": 0.47755077481269836,
|
| 1914 |
+
"learning_rate": 7.129417039084333e-05,
|
| 1915 |
+
"loss": 1.5864,
|
| 1916 |
+
"step": 269
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 0.10758847408661869,
|
| 1920 |
+
"grad_norm": 0.4970269799232483,
|
| 1921 |
+
"learning_rate": 7.060627444054893e-05,
|
| 1922 |
+
"loss": 1.6373,
|
| 1923 |
+
"step": 270
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 0.10798694991656912,
|
| 1927 |
+
"grad_norm": 0.47547978162765503,
|
| 1928 |
+
"learning_rate": 6.99198976017005e-05,
|
| 1929 |
+
"loss": 1.7433,
|
| 1930 |
+
"step": 271
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 0.10838542574651956,
|
| 1934 |
+
"grad_norm": 0.5408848524093628,
|
| 1935 |
+
"learning_rate": 6.923507534727373e-05,
|
| 1936 |
+
"loss": 1.77,
|
| 1937 |
+
"step": 272
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 0.10878390157647,
|
| 1941 |
+
"grad_norm": 0.49777430295944214,
|
| 1942 |
+
"learning_rate": 6.855184306990106e-05,
|
| 1943 |
+
"loss": 1.6071,
|
| 1944 |
+
"step": 273
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 0.10918237740642044,
|
| 1948 |
+
"grad_norm": 0.4691534638404846,
|
| 1949 |
+
"learning_rate": 6.78702360800425e-05,
|
| 1950 |
+
"loss": 1.5913,
|
| 1951 |
+
"step": 274
|
| 1952 |
+
},
|
| 1953 |
+
{
|
| 1954 |
+
"epoch": 0.10958085323637089,
|
| 1955 |
+
"grad_norm": 0.5284269452095032,
|
| 1956 |
+
"learning_rate": 6.719028960416098e-05,
|
| 1957 |
+
"loss": 1.8038,
|
| 1958 |
+
"step": 275
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 0.10997932906632132,
|
| 1962 |
+
"grad_norm": 0.49061042070388794,
|
| 1963 |
+
"learning_rate": 6.651203878290139e-05,
|
| 1964 |
+
"loss": 1.5991,
|
| 1965 |
+
"step": 276
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 0.11037780489627176,
|
| 1969 |
+
"grad_norm": 0.5676330327987671,
|
| 1970 |
+
"learning_rate": 6.583551866927475e-05,
|
| 1971 |
+
"loss": 1.8924,
|
| 1972 |
+
"step": 277
|
| 1973 |
+
},
|
| 1974 |
+
{
|
| 1975 |
+
"epoch": 0.1107762807262222,
|
| 1976 |
+
"grad_norm": 0.5392544865608215,
|
| 1977 |
+
"learning_rate": 6.516076422684654e-05,
|
| 1978 |
+
"loss": 1.7611,
|
| 1979 |
+
"step": 278
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 0.11117475655617264,
|
| 1983 |
+
"grad_norm": 0.5719506740570068,
|
| 1984 |
+
"learning_rate": 6.448781032792972e-05,
|
| 1985 |
+
"loss": 1.756,
|
| 1986 |
+
"step": 279
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 0.11157323238612307,
|
| 1990 |
+
"grad_norm": 0.4809233248233795,
|
| 1991 |
+
"learning_rate": 6.381669175178248e-05,
|
| 1992 |
+
"loss": 1.641,
|
| 1993 |
+
"step": 280
|
| 1994 |
+
},
|
| 1995 |
+
{
|
| 1996 |
+
"epoch": 0.11197170821607352,
|
| 1997 |
+
"grad_norm": 0.48434188961982727,
|
| 1998 |
+
"learning_rate": 6.31474431828108e-05,
|
| 1999 |
+
"loss": 1.579,
|
| 2000 |
+
"step": 281
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 0.11237018404602396,
|
| 2004 |
+
"grad_norm": 0.5024405717849731,
|
| 2005 |
+
"learning_rate": 6.248009920877592e-05,
|
| 2006 |
+
"loss": 1.6653,
|
| 2007 |
+
"step": 282
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 0.1127686598759744,
|
| 2011 |
+
"grad_norm": 0.441279798746109,
|
| 2012 |
+
"learning_rate": 6.181469431900672e-05,
|
| 2013 |
+
"loss": 1.5105,
|
| 2014 |
+
"step": 283
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 0.11316713570592483,
|
| 2018 |
+
"grad_norm": 0.5233234763145447,
|
| 2019 |
+
"learning_rate": 6.115126290261745e-05,
|
| 2020 |
+
"loss": 1.7695,
|
| 2021 |
+
"step": 284
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 0.11356561153587527,
|
| 2025 |
+
"grad_norm": 0.5281261801719666,
|
| 2026 |
+
"learning_rate": 6.048983924673022e-05,
|
| 2027 |
+
"loss": 1.76,
|
| 2028 |
+
"step": 285
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 0.11396408736582572,
|
| 2032 |
+
"grad_norm": 0.534590482711792,
|
| 2033 |
+
"learning_rate": 5.983045753470308e-05,
|
| 2034 |
+
"loss": 1.8155,
|
| 2035 |
+
"step": 286
|
| 2036 |
+
},
|
| 2037 |
+
{
|
| 2038 |
+
"epoch": 0.11436256319577616,
|
| 2039 |
+
"grad_norm": 0.5247072577476501,
|
| 2040 |
+
"learning_rate": 5.917315184436345e-05,
|
| 2041 |
+
"loss": 1.6073,
|
| 2042 |
+
"step": 287
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 0.1147610390257266,
|
| 2046 |
+
"grad_norm": 0.4829355776309967,
|
| 2047 |
+
"learning_rate": 5.851795614624682e-05,
|
| 2048 |
+
"loss": 1.5224,
|
| 2049 |
+
"step": 288
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 0.11515951485567703,
|
| 2053 |
+
"grad_norm": 0.516015887260437,
|
| 2054 |
+
"learning_rate": 5.786490430184115e-05,
|
| 2055 |
+
"loss": 1.6813,
|
| 2056 |
+
"step": 289
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 0.11555799068562747,
|
| 2060 |
+
"grad_norm": 0.48894891142845154,
|
| 2061 |
+
"learning_rate": 5.72140300618369e-05,
|
| 2062 |
+
"loss": 1.7965,
|
| 2063 |
+
"step": 290
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 0.11595646651557791,
|
| 2067 |
+
"grad_norm": 0.49149996042251587,
|
| 2068 |
+
"learning_rate": 5.656536706438267e-05,
|
| 2069 |
+
"loss": 1.6388,
|
| 2070 |
+
"step": 291
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 0.11635494234552836,
|
| 2074 |
+
"grad_norm": 0.4835774898529053,
|
| 2075 |
+
"learning_rate": 5.591894883334667e-05,
|
| 2076 |
+
"loss": 1.6856,
|
| 2077 |
+
"step": 292
|
| 2078 |
+
},
|
| 2079 |
+
{
|
| 2080 |
+
"epoch": 0.1167534181754788,
|
| 2081 |
+
"grad_norm": 0.5278857946395874,
|
| 2082 |
+
"learning_rate": 5.5274808776584367e-05,
|
| 2083 |
+
"loss": 1.6883,
|
| 2084 |
+
"step": 293
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 0.11715189400542923,
|
| 2088 |
+
"grad_norm": 0.4995588958263397,
|
| 2089 |
+
"learning_rate": 5.463298018421171e-05,
|
| 2090 |
+
"loss": 1.519,
|
| 2091 |
+
"step": 294
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 0.11755036983537967,
|
| 2095 |
+
"grad_norm": 0.5236543416976929,
|
| 2096 |
+
"learning_rate": 5.399349622688479e-05,
|
| 2097 |
+
"loss": 1.7372,
|
| 2098 |
+
"step": 295
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 0.11794884566533011,
|
| 2102 |
+
"grad_norm": 0.45699524879455566,
|
| 2103 |
+
"learning_rate": 5.335638995408545e-05,
|
| 2104 |
+
"loss": 1.6082,
|
| 2105 |
+
"step": 296
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 0.11834732149528056,
|
| 2109 |
+
"grad_norm": 0.5191316604614258,
|
| 2110 |
+
"learning_rate": 5.272169429241325e-05,
|
| 2111 |
+
"loss": 1.7123,
|
| 2112 |
+
"step": 297
|
| 2113 |
+
},
|
| 2114 |
+
{
|
| 2115 |
+
"epoch": 0.118745797325231,
|
| 2116 |
+
"grad_norm": 0.42880895733833313,
|
| 2117 |
+
"learning_rate": 5.208944204388377e-05,
|
| 2118 |
+
"loss": 1.4809,
|
| 2119 |
+
"step": 298
|
| 2120 |
+
},
|
| 2121 |
+
{
|
| 2122 |
+
"epoch": 0.11914427315518143,
|
| 2123 |
+
"grad_norm": 0.5574065446853638,
|
| 2124 |
+
"learning_rate": 5.145966588423341e-05,
|
| 2125 |
+
"loss": 1.8128,
|
| 2126 |
+
"step": 299
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 0.11954274898513187,
|
| 2130 |
+
"grad_norm": 0.47847244143486023,
|
| 2131 |
+
"learning_rate": 5.0832398361230596e-05,
|
| 2132 |
+
"loss": 1.5699,
|
| 2133 |
+
"step": 300
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.11954274898513187,
|
| 2137 |
+
"eval_loss": 1.700640082359314,
|
| 2138 |
+
"eval_runtime": 603.0633,
|
| 2139 |
+
"eval_samples_per_second": 14.018,
|
| 2140 |
+
"eval_steps_per_second": 3.505,
|
| 2141 |
+
"step": 300
|
| 2142 |
}
|
| 2143 |
],
|
| 2144 |
"logging_steps": 1,
|
|
|
|
| 2167 |
"attributes": {}
|
| 2168 |
}
|
| 2169 |
},
|
| 2170 |
+
"total_flos": 8.388464319922176e+17,
|
| 2171 |
"train_batch_size": 4,
|
| 2172 |
"trial_name": null,
|
| 2173 |
"trial_params": null
|