Training in progress, step 275, 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 63592
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0319b713ae7728f33bdad1072d732c911d942651ce63fc38c8afd1de15515ace
|
| 3 |
size 63592
|
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:e46de7f0fcda3b48c5af54e726d5535f03dfa81a8e6cb480f4a85c7c4a992cf4
|
| 3 |
+
size 132862
|
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:ca4aad791f2499a5a068a1a296bbf0dada64c634ae9dc09f1d85e19cd28a6773
|
| 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 sha256:678cbcbb9a6d696123ac5a18e6d2c2b6e4fd2702514ff58653037b0e5bfa25c4
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 10.98902416229248,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 4 |
-
"epoch":
|
| 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 +1431,531 @@
|
|
| 1431 |
"eval_samples_per_second": 234.205,
|
| 1432 |
"eval_steps_per_second": 58.805,
|
| 1433 |
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1434 |
}
|
| 1435 |
],
|
| 1436 |
"logging_steps": 1,
|
|
@@ -1454,12 +1979,12 @@
|
|
| 1454 |
"should_evaluate": false,
|
| 1455 |
"should_log": false,
|
| 1456 |
"should_save": true,
|
| 1457 |
-
"should_training_stop":
|
| 1458 |
},
|
| 1459 |
"attributes": {}
|
| 1460 |
}
|
| 1461 |
},
|
| 1462 |
-
"total_flos":
|
| 1463 |
"train_batch_size": 4,
|
| 1464 |
"trial_name": null,
|
| 1465 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 10.98902416229248,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 4 |
+
"epoch": 1.0027347310847767,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 275,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1431 |
"eval_samples_per_second": 234.205,
|
| 1432 |
"eval_steps_per_second": 58.805,
|
| 1433 |
"step": 200
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 0.7329079307201458,
|
| 1437 |
+
"grad_norm": 0.47317686676979065,
|
| 1438 |
+
"learning_rate": 3.607509065022101e-05,
|
| 1439 |
+
"loss": 87.9034,
|
| 1440 |
+
"step": 201
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 0.7365542388331814,
|
| 1444 |
+
"grad_norm": 0.5720061659812927,
|
| 1445 |
+
"learning_rate": 3.516794972664455e-05,
|
| 1446 |
+
"loss": 87.9438,
|
| 1447 |
+
"step": 202
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 0.740200546946217,
|
| 1451 |
+
"grad_norm": 0.5105446577072144,
|
| 1452 |
+
"learning_rate": 3.426992036155434e-05,
|
| 1453 |
+
"loss": 87.8961,
|
| 1454 |
+
"step": 203
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 0.7438468550592525,
|
| 1458 |
+
"grad_norm": 0.6989304423332214,
|
| 1459 |
+
"learning_rate": 3.33811287648582e-05,
|
| 1460 |
+
"loss": 87.9163,
|
| 1461 |
+
"step": 204
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 0.7474931631722881,
|
| 1465 |
+
"grad_norm": 0.597347617149353,
|
| 1466 |
+
"learning_rate": 3.250169984817897e-05,
|
| 1467 |
+
"loss": 87.9034,
|
| 1468 |
+
"step": 205
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 0.7511394712853237,
|
| 1472 |
+
"grad_norm": 0.44015979766845703,
|
| 1473 |
+
"learning_rate": 3.163175720729954e-05,
|
| 1474 |
+
"loss": 87.9456,
|
| 1475 |
+
"step": 206
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 0.7547857793983591,
|
| 1479 |
+
"grad_norm": 0.42753106355667114,
|
| 1480 |
+
"learning_rate": 3.0771423104792454e-05,
|
| 1481 |
+
"loss": 87.9205,
|
| 1482 |
+
"step": 207
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 0.7584320875113947,
|
| 1486 |
+
"grad_norm": 0.48786309361457825,
|
| 1487 |
+
"learning_rate": 2.9920818452837152e-05,
|
| 1488 |
+
"loss": 87.9073,
|
| 1489 |
+
"step": 208
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 0.7620783956244302,
|
| 1493 |
+
"grad_norm": 0.4329456686973572,
|
| 1494 |
+
"learning_rate": 2.908006279622667e-05,
|
| 1495 |
+
"loss": 87.9361,
|
| 1496 |
+
"step": 209
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 0.7657247037374658,
|
| 1500 |
+
"grad_norm": 0.5936325192451477,
|
| 1501 |
+
"learning_rate": 2.8249274295566864e-05,
|
| 1502 |
+
"loss": 87.9373,
|
| 1503 |
+
"step": 210
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 0.7693710118505014,
|
| 1507 |
+
"grad_norm": 0.5534446835517883,
|
| 1508 |
+
"learning_rate": 2.742856971066996e-05,
|
| 1509 |
+
"loss": 87.8996,
|
| 1510 |
+
"step": 211
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 0.773017319963537,
|
| 1514 |
+
"grad_norm": 0.5343015193939209,
|
| 1515 |
+
"learning_rate": 2.6618064384144925e-05,
|
| 1516 |
+
"loss": 87.9545,
|
| 1517 |
+
"step": 212
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 0.7766636280765725,
|
| 1521 |
+
"grad_norm": 0.41181403398513794,
|
| 1522 |
+
"learning_rate": 2.58178722251872e-05,
|
| 1523 |
+
"loss": 87.8978,
|
| 1524 |
+
"step": 213
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 0.780309936189608,
|
| 1528 |
+
"grad_norm": 0.5438745617866516,
|
| 1529 |
+
"learning_rate": 2.502810569356976e-05,
|
| 1530 |
+
"loss": 87.9433,
|
| 1531 |
+
"step": 214
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 0.7839562443026435,
|
| 1535 |
+
"grad_norm": 0.5030822157859802,
|
| 1536 |
+
"learning_rate": 2.4248875783837987e-05,
|
| 1537 |
+
"loss": 87.9129,
|
| 1538 |
+
"step": 215
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 0.7876025524156791,
|
| 1542 |
+
"grad_norm": 0.4455682039260864,
|
| 1543 |
+
"learning_rate": 2.348029200971028e-05,
|
| 1544 |
+
"loss": 87.8729,
|
| 1545 |
+
"step": 216
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 0.7912488605287147,
|
| 1549 |
+
"grad_norm": 0.4018632471561432,
|
| 1550 |
+
"learning_rate": 2.2722462388686873e-05,
|
| 1551 |
+
"loss": 87.928,
|
| 1552 |
+
"step": 217
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 0.7948951686417502,
|
| 1556 |
+
"grad_norm": 0.39684757590293884,
|
| 1557 |
+
"learning_rate": 2.1975493426869153e-05,
|
| 1558 |
+
"loss": 87.9542,
|
| 1559 |
+
"step": 218
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 0.7985414767547858,
|
| 1563 |
+
"grad_norm": 0.7904627919197083,
|
| 1564 |
+
"learning_rate": 2.1239490103990945e-05,
|
| 1565 |
+
"loss": 87.9652,
|
| 1566 |
+
"step": 219
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 0.8021877848678214,
|
| 1570 |
+
"grad_norm": 0.5595088601112366,
|
| 1571 |
+
"learning_rate": 2.0514555858664663e-05,
|
| 1572 |
+
"loss": 87.9473,
|
| 1573 |
+
"step": 220
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 0.8058340929808568,
|
| 1577 |
+
"grad_norm": 0.4692498743534088,
|
| 1578 |
+
"learning_rate": 1.9800792573844053e-05,
|
| 1579 |
+
"loss": 87.9208,
|
| 1580 |
+
"step": 221
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 0.8094804010938924,
|
| 1584 |
+
"grad_norm": 0.6133750677108765,
|
| 1585 |
+
"learning_rate": 1.9098300562505266e-05,
|
| 1586 |
+
"loss": 87.8955,
|
| 1587 |
+
"step": 222
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 0.813126709206928,
|
| 1591 |
+
"grad_norm": 0.5072586536407471,
|
| 1592 |
+
"learning_rate": 1.8407178553548878e-05,
|
| 1593 |
+
"loss": 87.943,
|
| 1594 |
+
"step": 223
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 0.8167730173199635,
|
| 1598 |
+
"grad_norm": 0.424775630235672,
|
| 1599 |
+
"learning_rate": 1.772752367792452e-05,
|
| 1600 |
+
"loss": 87.9526,
|
| 1601 |
+
"step": 224
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 0.8204193254329991,
|
| 1605 |
+
"grad_norm": 0.5902768969535828,
|
| 1606 |
+
"learning_rate": 1.7059431454979824e-05,
|
| 1607 |
+
"loss": 87.9597,
|
| 1608 |
+
"step": 225
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 0.8240656335460347,
|
| 1612 |
+
"grad_norm": 0.5081129670143127,
|
| 1613 |
+
"learning_rate": 1.6402995779036146e-05,
|
| 1614 |
+
"loss": 87.8714,
|
| 1615 |
+
"step": 226
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 0.8277119416590701,
|
| 1619 |
+
"grad_norm": 0.47168031334877014,
|
| 1620 |
+
"learning_rate": 1.5758308906192608e-05,
|
| 1621 |
+
"loss": 87.9489,
|
| 1622 |
+
"step": 227
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 0.8313582497721057,
|
| 1626 |
+
"grad_norm": 0.642856776714325,
|
| 1627 |
+
"learning_rate": 1.5125461441360223e-05,
|
| 1628 |
+
"loss": 87.8958,
|
| 1629 |
+
"step": 228
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 0.8350045578851413,
|
| 1633 |
+
"grad_norm": 0.5361236929893494,
|
| 1634 |
+
"learning_rate": 1.4504542325528258e-05,
|
| 1635 |
+
"loss": 87.938,
|
| 1636 |
+
"step": 229
|
| 1637 |
+
},
|
| 1638 |
+
{
|
| 1639 |
+
"epoch": 0.8386508659981768,
|
| 1640 |
+
"grad_norm": 0.6032053232192993,
|
| 1641 |
+
"learning_rate": 1.3895638823264446e-05,
|
| 1642 |
+
"loss": 87.9137,
|
| 1643 |
+
"step": 230
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 0.8422971741112124,
|
| 1647 |
+
"grad_norm": 0.5546380877494812,
|
| 1648 |
+
"learning_rate": 1.3298836510450596e-05,
|
| 1649 |
+
"loss": 87.9028,
|
| 1650 |
+
"step": 231
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 0.845943482224248,
|
| 1654 |
+
"grad_norm": 0.5335825681686401,
|
| 1655 |
+
"learning_rate": 1.2714219262255778e-05,
|
| 1656 |
+
"loss": 87.9484,
|
| 1657 |
+
"step": 232
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 0.8495897903372835,
|
| 1661 |
+
"grad_norm": 0.44430670142173767,
|
| 1662 |
+
"learning_rate": 1.214186924134838e-05,
|
| 1663 |
+
"loss": 87.8892,
|
| 1664 |
+
"step": 233
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 0.853236098450319,
|
| 1668 |
+
"grad_norm": 0.4880811274051666,
|
| 1669 |
+
"learning_rate": 1.158186688634898e-05,
|
| 1670 |
+
"loss": 87.9078,
|
| 1671 |
+
"step": 234
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 0.8568824065633546,
|
| 1675 |
+
"grad_norm": 0.8008484840393066,
|
| 1676 |
+
"learning_rate": 1.103429090052528e-05,
|
| 1677 |
+
"loss": 87.9225,
|
| 1678 |
+
"step": 235
|
| 1679 |
+
},
|
| 1680 |
+
{
|
| 1681 |
+
"epoch": 0.8605287146763901,
|
| 1682 |
+
"grad_norm": 0.5064822435379028,
|
| 1683 |
+
"learning_rate": 1.0499218240731157e-05,
|
| 1684 |
+
"loss": 87.8992,
|
| 1685 |
+
"step": 236
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 0.8641750227894257,
|
| 1689 |
+
"grad_norm": 0.45253366231918335,
|
| 1690 |
+
"learning_rate": 9.976724106591129e-06,
|
| 1691 |
+
"loss": 87.9188,
|
| 1692 |
+
"step": 237
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 0.8678213309024613,
|
| 1696 |
+
"grad_norm": 0.5466088056564331,
|
| 1697 |
+
"learning_rate": 9.466881929931581e-06,
|
| 1698 |
+
"loss": 87.8964,
|
| 1699 |
+
"step": 238
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 0.8714676390154968,
|
| 1703 |
+
"grad_norm": 0.4804587662220001,
|
| 1704 |
+
"learning_rate": 8.969763364460681e-06,
|
| 1705 |
+
"loss": 87.9137,
|
| 1706 |
+
"step": 239
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 0.8751139471285324,
|
| 1710 |
+
"grad_norm": 0.5322223901748657,
|
| 1711 |
+
"learning_rate": 8.485438275698154e-06,
|
| 1712 |
+
"loss": 87.9524,
|
| 1713 |
+
"step": 240
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 0.8787602552415679,
|
| 1717 |
+
"grad_norm": 0.42037254571914673,
|
| 1718 |
+
"learning_rate": 8.01397473115616e-06,
|
| 1719 |
+
"loss": 87.8973,
|
| 1720 |
+
"step": 241
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 0.8824065633546034,
|
| 1724 |
+
"grad_norm": 0.5107778906822205,
|
| 1725 |
+
"learning_rate": 7.5554389907731334e-06,
|
| 1726 |
+
"loss": 87.9181,
|
| 1727 |
+
"step": 242
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 0.886052871467639,
|
| 1731 |
+
"grad_norm": 0.48512643575668335,
|
| 1732 |
+
"learning_rate": 7.109895497601571e-06,
|
| 1733 |
+
"loss": 87.9424,
|
| 1734 |
+
"step": 243
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 0.8896991795806746,
|
| 1738 |
+
"grad_norm": 0.5565751791000366,
|
| 1739 |
+
"learning_rate": 6.677406868751013e-06,
|
| 1740 |
+
"loss": 87.9308,
|
| 1741 |
+
"step": 244
|
| 1742 |
+
},
|
| 1743 |
+
{
|
| 1744 |
+
"epoch": 0.8933454876937101,
|
| 1745 |
+
"grad_norm": 0.4575342833995819,
|
| 1746 |
+
"learning_rate": 6.258033886587911e-06,
|
| 1747 |
+
"loss": 87.9269,
|
| 1748 |
+
"step": 245
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 0.8969917958067457,
|
| 1752 |
+
"grad_norm": 0.47321030497550964,
|
| 1753 |
+
"learning_rate": 5.8518354901931364e-06,
|
| 1754 |
+
"loss": 87.9444,
|
| 1755 |
+
"step": 246
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 0.9006381039197813,
|
| 1759 |
+
"grad_norm": 0.5037922859191895,
|
| 1760 |
+
"learning_rate": 5.4588687670786735e-06,
|
| 1761 |
+
"loss": 87.9298,
|
| 1762 |
+
"step": 247
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 0.9042844120328167,
|
| 1766 |
+
"grad_norm": 0.4619918465614319,
|
| 1767 |
+
"learning_rate": 5.079188945164426e-06,
|
| 1768 |
+
"loss": 87.9265,
|
| 1769 |
+
"step": 248
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 0.9079307201458523,
|
| 1773 |
+
"grad_norm": 0.7055258750915527,
|
| 1774 |
+
"learning_rate": 4.712849385016471e-06,
|
| 1775 |
+
"loss": 87.8928,
|
| 1776 |
+
"step": 249
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 0.9115770282588879,
|
| 1780 |
+
"grad_norm": 0.49585390090942383,
|
| 1781 |
+
"learning_rate": 4.359901572347758e-06,
|
| 1782 |
+
"loss": 87.9267,
|
| 1783 |
+
"step": 250
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 0.9152233363719234,
|
| 1787 |
+
"grad_norm": 0.4524700939655304,
|
| 1788 |
+
"learning_rate": 4.02039511078216e-06,
|
| 1789 |
+
"loss": 87.9438,
|
| 1790 |
+
"step": 251
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 0.918869644484959,
|
| 1794 |
+
"grad_norm": 0.45400986075401306,
|
| 1795 |
+
"learning_rate": 3.6943777148831902e-06,
|
| 1796 |
+
"loss": 87.9397,
|
| 1797 |
+
"step": 252
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 0.9225159525979946,
|
| 1801 |
+
"grad_norm": 0.45972177386283875,
|
| 1802 |
+
"learning_rate": 3.3818952034481823e-06,
|
| 1803 |
+
"loss": 87.9405,
|
| 1804 |
+
"step": 253
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 0.92616226071103,
|
| 1808 |
+
"grad_norm": 0.4140474498271942,
|
| 1809 |
+
"learning_rate": 3.0829914930687764e-06,
|
| 1810 |
+
"loss": 87.8955,
|
| 1811 |
+
"step": 254
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 0.9298085688240656,
|
| 1815 |
+
"grad_norm": 0.5047031044960022,
|
| 1816 |
+
"learning_rate": 2.7977085919589254e-06,
|
| 1817 |
+
"loss": 87.956,
|
| 1818 |
+
"step": 255
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 0.9334548769371012,
|
| 1822 |
+
"grad_norm": 0.5136326551437378,
|
| 1823 |
+
"learning_rate": 2.5260865940510025e-06,
|
| 1824 |
+
"loss": 87.9139,
|
| 1825 |
+
"step": 256
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 0.9371011850501367,
|
| 1829 |
+
"grad_norm": 0.4639621675014496,
|
| 1830 |
+
"learning_rate": 2.2681636733609457e-06,
|
| 1831 |
+
"loss": 87.9264,
|
| 1832 |
+
"step": 257
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 0.9407474931631723,
|
| 1836 |
+
"grad_norm": 0.6435884833335876,
|
| 1837 |
+
"learning_rate": 2.0239760786232353e-06,
|
| 1838 |
+
"loss": 87.8972,
|
| 1839 |
+
"step": 258
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 0.9443938012762079,
|
| 1843 |
+
"grad_norm": 0.516819179058075,
|
| 1844 |
+
"learning_rate": 1.793558128196493e-06,
|
| 1845 |
+
"loss": 87.887,
|
| 1846 |
+
"step": 259
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 0.9480401093892434,
|
| 1850 |
+
"grad_norm": 0.606723964214325,
|
| 1851 |
+
"learning_rate": 1.576942205240317e-06,
|
| 1852 |
+
"loss": 87.8938,
|
| 1853 |
+
"step": 260
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 0.9516864175022789,
|
| 1857 |
+
"grad_norm": 0.7339270114898682,
|
| 1858 |
+
"learning_rate": 1.3741587531641564e-06,
|
| 1859 |
+
"loss": 87.8665,
|
| 1860 |
+
"step": 261
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 0.9553327256153145,
|
| 1864 |
+
"grad_norm": 0.5948337912559509,
|
| 1865 |
+
"learning_rate": 1.185236271348722e-06,
|
| 1866 |
+
"loss": 87.8634,
|
| 1867 |
+
"step": 262
|
| 1868 |
+
},
|
| 1869 |
+
{
|
| 1870 |
+
"epoch": 0.95897903372835,
|
| 1871 |
+
"grad_norm": 0.5366274118423462,
|
| 1872 |
+
"learning_rate": 1.0102013111406906e-06,
|
| 1873 |
+
"loss": 87.9242,
|
| 1874 |
+
"step": 263
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.9626253418413856,
|
| 1878 |
+
"grad_norm": 0.6018470525741577,
|
| 1879 |
+
"learning_rate": 8.490784721211453e-07,
|
| 1880 |
+
"loss": 87.9057,
|
| 1881 |
+
"step": 264
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 0.9662716499544212,
|
| 1885 |
+
"grad_norm": 0.47167351841926575,
|
| 1886 |
+
"learning_rate": 7.018903986483083e-07,
|
| 1887 |
+
"loss": 87.9259,
|
| 1888 |
+
"step": 265
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 0.9699179580674567,
|
| 1892 |
+
"grad_norm": 0.5208801627159119,
|
| 1893 |
+
"learning_rate": 5.686577766751078e-07,
|
| 1894 |
+
"loss": 87.9223,
|
| 1895 |
+
"step": 266
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 0.9735642661804923,
|
| 1899 |
+
"grad_norm": 0.4630851447582245,
|
| 1900 |
+
"learning_rate": 4.4939933084192644e-07,
|
| 1901 |
+
"loss": 87.9368,
|
| 1902 |
+
"step": 267
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 0.9772105742935278,
|
| 1906 |
+
"grad_norm": 0.46978411078453064,
|
| 1907 |
+
"learning_rate": 3.4413182184507285e-07,
|
| 1908 |
+
"loss": 87.9259,
|
| 1909 |
+
"step": 268
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 0.9808568824065633,
|
| 1913 |
+
"grad_norm": 0.5887516736984253,
|
| 1914 |
+
"learning_rate": 2.528700440811438e-07,
|
| 1915 |
+
"loss": 87.9486,
|
| 1916 |
+
"step": 269
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 0.9845031905195989,
|
| 1920 |
+
"grad_norm": 0.5563957095146179,
|
| 1921 |
+
"learning_rate": 1.7562682356786487e-07,
|
| 1922 |
+
"loss": 87.9239,
|
| 1923 |
+
"step": 270
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 0.9881494986326345,
|
| 1927 |
+
"grad_norm": 0.5462055206298828,
|
| 1928 |
+
"learning_rate": 1.1241301614147714e-07,
|
| 1929 |
+
"loss": 87.9323,
|
| 1930 |
+
"step": 271
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 0.99179580674567,
|
| 1934 |
+
"grad_norm": 0.4110185503959656,
|
| 1935 |
+
"learning_rate": 6.323750593106859e-08,
|
| 1936 |
+
"loss": 87.916,
|
| 1937 |
+
"step": 272
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 0.9954421148587056,
|
| 1941 |
+
"grad_norm": 0.4329504668712616,
|
| 1942 |
+
"learning_rate": 2.810720410998391e-08,
|
| 1943 |
+
"loss": 87.9155,
|
| 1944 |
+
"step": 273
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 0.9990884229717412,
|
| 1948 |
+
"grad_norm": 0.5021648406982422,
|
| 1949 |
+
"learning_rate": 7.027047924512698e-09,
|
| 1950 |
+
"loss": 87.9112,
|
| 1951 |
+
"step": 274
|
| 1952 |
+
},
|
| 1953 |
+
{
|
| 1954 |
+
"epoch": 1.0027347310847767,
|
| 1955 |
+
"grad_norm": 0.5477859377861023,
|
| 1956 |
+
"learning_rate": 0.0,
|
| 1957 |
+
"loss": 87.9211,
|
| 1958 |
+
"step": 275
|
| 1959 |
}
|
| 1960 |
],
|
| 1961 |
"logging_steps": 1,
|
|
|
|
| 1979 |
"should_evaluate": false,
|
| 1980 |
"should_log": false,
|
| 1981 |
"should_save": true,
|
| 1982 |
+
"should_training_stop": true
|
| 1983 |
},
|
| 1984 |
"attributes": {}
|
| 1985 |
}
|
| 1986 |
},
|
| 1987 |
+
"total_flos": 4163614801920.0,
|
| 1988 |
"train_batch_size": 4,
|
| 1989 |
"trial_name": null,
|
| 1990 |
"trial_params": null
|