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 671149168
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:515bd1fd2e800ff785b6479fdd6957fd4bf27e5b1c30a3f1cf8aae527f08a08b
|
| 3 |
size 671149168
|
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 sha256:7bd05b1422ea617e033a87629b7a8d0f6328816ebf140850f114f1f26f4cb7a8
|
| 3 |
+
size 341314644
|
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 sha256:809cadbdaab075c11480d526dd51b8f21d82fd34b84730fb1208ace6f51dd4d6
|
| 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:d5e2a8ca8cbdb222b4fcc6a743fadf9ab6adfaf2459d28805db388e0cbe4b5b9
|
| 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": 12.655,
|
| 1432 |
"eval_steps_per_second": 3.164,
|
| 1433 |
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1434 |
}
|
| 1435 |
],
|
| 1436 |
"logging_steps": 1,
|
|
@@ -1459,7 +2167,7 @@
|
|
| 1459 |
"attributes": {}
|
| 1460 |
}
|
| 1461 |
},
|
| 1462 |
-
"total_flos":
|
| 1463 |
"train_batch_size": 8,
|
| 1464 |
"trial_name": null,
|
| 1465 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.35413047671318054,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-300",
|
| 4 |
+
"epoch": 0.23529411764705882,
|
| 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": 12.655,
|
| 1432 |
"eval_steps_per_second": 3.164,
|
| 1433 |
"step": 200
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 0.15764705882352942,
|
| 1437 |
+
"grad_norm": 2.9938836097717285,
|
| 1438 |
+
"learning_rate": 3.6127556076296055e-05,
|
| 1439 |
+
"loss": 3.0403,
|
| 1440 |
+
"step": 201
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 0.1584313725490196,
|
| 1444 |
+
"grad_norm": 2.702965497970581,
|
| 1445 |
+
"learning_rate": 3.5845731083264626e-05,
|
| 1446 |
+
"loss": 1.9497,
|
| 1447 |
+
"step": 202
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 0.1592156862745098,
|
| 1451 |
+
"grad_norm": 2.5172975063323975,
|
| 1452 |
+
"learning_rate": 3.5563851211955195e-05,
|
| 1453 |
+
"loss": 1.5842,
|
| 1454 |
+
"step": 203
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 0.16,
|
| 1458 |
+
"grad_norm": 2.603217363357544,
|
| 1459 |
+
"learning_rate": 3.528193475314864e-05,
|
| 1460 |
+
"loss": 1.5727,
|
| 1461 |
+
"step": 204
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 0.1607843137254902,
|
| 1465 |
+
"grad_norm": 2.615779399871826,
|
| 1466 |
+
"learning_rate": 3.5e-05,
|
| 1467 |
+
"loss": 1.3275,
|
| 1468 |
+
"step": 205
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 0.16156862745098038,
|
| 1472 |
+
"grad_norm": 2.509752035140991,
|
| 1473 |
+
"learning_rate": 3.4718065246851354e-05,
|
| 1474 |
+
"loss": 1.2942,
|
| 1475 |
+
"step": 206
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 0.1623529411764706,
|
| 1479 |
+
"grad_norm": 2.902695655822754,
|
| 1480 |
+
"learning_rate": 3.4436148788044806e-05,
|
| 1481 |
+
"loss": 1.1889,
|
| 1482 |
+
"step": 207
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 0.1631372549019608,
|
| 1486 |
+
"grad_norm": 3.608950138092041,
|
| 1487 |
+
"learning_rate": 3.4154268916735375e-05,
|
| 1488 |
+
"loss": 1.4452,
|
| 1489 |
+
"step": 208
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 0.16392156862745097,
|
| 1493 |
+
"grad_norm": 2.8737595081329346,
|
| 1494 |
+
"learning_rate": 3.387244392370395e-05,
|
| 1495 |
+
"loss": 1.2904,
|
| 1496 |
+
"step": 209
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 0.16470588235294117,
|
| 1500 |
+
"grad_norm": 3.235945463180542,
|
| 1501 |
+
"learning_rate": 3.3590692096170476e-05,
|
| 1502 |
+
"loss": 1.2737,
|
| 1503 |
+
"step": 210
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 0.16549019607843138,
|
| 1507 |
+
"grad_norm": 3.6993987560272217,
|
| 1508 |
+
"learning_rate": 3.3309031716607255e-05,
|
| 1509 |
+
"loss": 1.3786,
|
| 1510 |
+
"step": 211
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 0.16627450980392156,
|
| 1514 |
+
"grad_norm": 3.325028419494629,
|
| 1515 |
+
"learning_rate": 3.302748106155268e-05,
|
| 1516 |
+
"loss": 1.1542,
|
| 1517 |
+
"step": 212
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 0.16705882352941176,
|
| 1521 |
+
"grad_norm": 2.829763650894165,
|
| 1522 |
+
"learning_rate": 3.274605840042531e-05,
|
| 1523 |
+
"loss": 0.9143,
|
| 1524 |
+
"step": 213
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 0.16784313725490196,
|
| 1528 |
+
"grad_norm": 4.127065181732178,
|
| 1529 |
+
"learning_rate": 3.2464781994338315e-05,
|
| 1530 |
+
"loss": 1.4021,
|
| 1531 |
+
"step": 214
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 0.16862745098039217,
|
| 1535 |
+
"grad_norm": 3.2039248943328857,
|
| 1536 |
+
"learning_rate": 3.2183670094914595e-05,
|
| 1537 |
+
"loss": 1.1642,
|
| 1538 |
+
"step": 215
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 0.16941176470588235,
|
| 1542 |
+
"grad_norm": 3.7868335247039795,
|
| 1543 |
+
"learning_rate": 3.190274094310245e-05,
|
| 1544 |
+
"loss": 1.3842,
|
| 1545 |
+
"step": 216
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 0.17019607843137255,
|
| 1549 |
+
"grad_norm": 3.718687057495117,
|
| 1550 |
+
"learning_rate": 3.162201276799197e-05,
|
| 1551 |
+
"loss": 1.4267,
|
| 1552 |
+
"step": 217
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 0.17098039215686275,
|
| 1556 |
+
"grad_norm": 3.7910690307617188,
|
| 1557 |
+
"learning_rate": 3.134150378563213e-05,
|
| 1558 |
+
"loss": 1.2455,
|
| 1559 |
+
"step": 218
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 0.17176470588235293,
|
| 1563 |
+
"grad_norm": 3.5929932594299316,
|
| 1564 |
+
"learning_rate": 3.1061232197848805e-05,
|
| 1565 |
+
"loss": 1.2007,
|
| 1566 |
+
"step": 219
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 0.17254901960784313,
|
| 1570 |
+
"grad_norm": 4.848661422729492,
|
| 1571 |
+
"learning_rate": 3.07812161910637e-05,
|
| 1572 |
+
"loss": 1.5243,
|
| 1573 |
+
"step": 220
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 0.17333333333333334,
|
| 1577 |
+
"grad_norm": 4.240635395050049,
|
| 1578 |
+
"learning_rate": 3.050147393511419e-05,
|
| 1579 |
+
"loss": 1.2182,
|
| 1580 |
+
"step": 221
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 0.17411764705882352,
|
| 1584 |
+
"grad_norm": 4.414323806762695,
|
| 1585 |
+
"learning_rate": 3.0222023582074396e-05,
|
| 1586 |
+
"loss": 1.3929,
|
| 1587 |
+
"step": 222
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 0.17490196078431372,
|
| 1591 |
+
"grad_norm": 4.253678321838379,
|
| 1592 |
+
"learning_rate": 2.994288326507726e-05,
|
| 1593 |
+
"loss": 1.0231,
|
| 1594 |
+
"step": 223
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 0.17568627450980392,
|
| 1598 |
+
"grad_norm": 4.7787089347839355,
|
| 1599 |
+
"learning_rate": 2.9664071097137908e-05,
|
| 1600 |
+
"loss": 1.3696,
|
| 1601 |
+
"step": 224
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 0.17647058823529413,
|
| 1605 |
+
"grad_norm": 6.1173248291015625,
|
| 1606 |
+
"learning_rate": 2.9385605169978387e-05,
|
| 1607 |
+
"loss": 1.3486,
|
| 1608 |
+
"step": 225
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 0.1772549019607843,
|
| 1612 |
+
"grad_norm": 5.057071685791016,
|
| 1613 |
+
"learning_rate": 2.9107503552853648e-05,
|
| 1614 |
+
"loss": 1.1463,
|
| 1615 |
+
"step": 226
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 0.1780392156862745,
|
| 1619 |
+
"grad_norm": 4.796481132507324,
|
| 1620 |
+
"learning_rate": 2.8829784291379082e-05,
|
| 1621 |
+
"loss": 1.0947,
|
| 1622 |
+
"step": 227
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 0.17882352941176471,
|
| 1626 |
+
"grad_norm": 4.333375930786133,
|
| 1627 |
+
"learning_rate": 2.8552465406359593e-05,
|
| 1628 |
+
"loss": 1.0557,
|
| 1629 |
+
"step": 228
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 0.1796078431372549,
|
| 1633 |
+
"grad_norm": 4.937939167022705,
|
| 1634 |
+
"learning_rate": 2.8275564892620202e-05,
|
| 1635 |
+
"loss": 0.8979,
|
| 1636 |
+
"step": 229
|
| 1637 |
+
},
|
| 1638 |
+
{
|
| 1639 |
+
"epoch": 0.1803921568627451,
|
| 1640 |
+
"grad_norm": 5.160710334777832,
|
| 1641 |
+
"learning_rate": 2.799910071783845e-05,
|
| 1642 |
+
"loss": 0.9403,
|
| 1643 |
+
"step": 230
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 0.1811764705882353,
|
| 1647 |
+
"grad_norm": 7.218094348907471,
|
| 1648 |
+
"learning_rate": 2.7723090821378425e-05,
|
| 1649 |
+
"loss": 0.899,
|
| 1650 |
+
"step": 231
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 0.18196078431372548,
|
| 1654 |
+
"grad_norm": 4.6421332359313965,
|
| 1655 |
+
"learning_rate": 2.7447553113126786e-05,
|
| 1656 |
+
"loss": 0.934,
|
| 1657 |
+
"step": 232
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 0.18274509803921568,
|
| 1661 |
+
"grad_norm": 7.592020034790039,
|
| 1662 |
+
"learning_rate": 2.7172505472330564e-05,
|
| 1663 |
+
"loss": 0.9639,
|
| 1664 |
+
"step": 233
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 0.18352941176470589,
|
| 1668 |
+
"grad_norm": 6.882318496704102,
|
| 1669 |
+
"learning_rate": 2.6897965746436997e-05,
|
| 1670 |
+
"loss": 1.3508,
|
| 1671 |
+
"step": 234
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 0.1843137254901961,
|
| 1675 |
+
"grad_norm": 6.335398197174072,
|
| 1676 |
+
"learning_rate": 2.6623951749935487e-05,
|
| 1677 |
+
"loss": 1.2642,
|
| 1678 |
+
"step": 235
|
| 1679 |
+
},
|
| 1680 |
+
{
|
| 1681 |
+
"epoch": 0.18509803921568627,
|
| 1682 |
+
"grad_norm": 5.449366569519043,
|
| 1683 |
+
"learning_rate": 2.635048126320156e-05,
|
| 1684 |
+
"loss": 1.4236,
|
| 1685 |
+
"step": 236
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 0.18588235294117647,
|
| 1689 |
+
"grad_norm": 8.11070728302002,
|
| 1690 |
+
"learning_rate": 2.6077572031343227e-05,
|
| 1691 |
+
"loss": 1.3364,
|
| 1692 |
+
"step": 237
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 0.18666666666666668,
|
| 1696 |
+
"grad_norm": 6.608270168304443,
|
| 1697 |
+
"learning_rate": 2.5805241763049455e-05,
|
| 1698 |
+
"loss": 1.4484,
|
| 1699 |
+
"step": 238
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 0.18745098039215685,
|
| 1703 |
+
"grad_norm": 7.942669868469238,
|
| 1704 |
+
"learning_rate": 2.553350812944107e-05,
|
| 1705 |
+
"loss": 1.4957,
|
| 1706 |
+
"step": 239
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 0.18823529411764706,
|
| 1710 |
+
"grad_norm": 7.074437618255615,
|
| 1711 |
+
"learning_rate": 2.5262388762924156e-05,
|
| 1712 |
+
"loss": 1.3078,
|
| 1713 |
+
"step": 240
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 0.18901960784313726,
|
| 1717 |
+
"grad_norm": 7.799503326416016,
|
| 1718 |
+
"learning_rate": 2.499190125604588e-05,
|
| 1719 |
+
"loss": 1.5845,
|
| 1720 |
+
"step": 241
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 0.18980392156862744,
|
| 1724 |
+
"grad_norm": 9.548646926879883,
|
| 1725 |
+
"learning_rate": 2.4722063160352926e-05,
|
| 1726 |
+
"loss": 2.075,
|
| 1727 |
+
"step": 242
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 0.19058823529411764,
|
| 1731 |
+
"grad_norm": 9.75729751586914,
|
| 1732 |
+
"learning_rate": 2.4452891985252647e-05,
|
| 1733 |
+
"loss": 2.0484,
|
| 1734 |
+
"step": 243
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 0.19137254901960785,
|
| 1738 |
+
"grad_norm": 8.452042579650879,
|
| 1739 |
+
"learning_rate": 2.4184405196876842e-05,
|
| 1740 |
+
"loss": 1.6378,
|
| 1741 |
+
"step": 244
|
| 1742 |
+
},
|
| 1743 |
+
{
|
| 1744 |
+
"epoch": 0.19215686274509805,
|
| 1745 |
+
"grad_norm": 7.993230819702148,
|
| 1746 |
+
"learning_rate": 2.3916620216948467e-05,
|
| 1747 |
+
"loss": 1.8914,
|
| 1748 |
+
"step": 245
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 0.19294117647058823,
|
| 1752 |
+
"grad_norm": 20.330001831054688,
|
| 1753 |
+
"learning_rate": 2.3649554421651106e-05,
|
| 1754 |
+
"loss": 2.7448,
|
| 1755 |
+
"step": 246
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 0.19372549019607843,
|
| 1759 |
+
"grad_norm": 11.52759838104248,
|
| 1760 |
+
"learning_rate": 2.338322514050151e-05,
|
| 1761 |
+
"loss": 1.4114,
|
| 1762 |
+
"step": 247
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 0.19450980392156864,
|
| 1766 |
+
"grad_norm": 12.445416450500488,
|
| 1767 |
+
"learning_rate": 2.3117649655225048e-05,
|
| 1768 |
+
"loss": 2.3627,
|
| 1769 |
+
"step": 248
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 0.1952941176470588,
|
| 1773 |
+
"grad_norm": 16.47171401977539,
|
| 1774 |
+
"learning_rate": 2.2852845198634395e-05,
|
| 1775 |
+
"loss": 3.235,
|
| 1776 |
+
"step": 249
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 0.19607843137254902,
|
| 1780 |
+
"grad_norm": 14.980457305908203,
|
| 1781 |
+
"learning_rate": 2.258882895351125e-05,
|
| 1782 |
+
"loss": 2.4173,
|
| 1783 |
+
"step": 250
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 0.19686274509803922,
|
| 1787 |
+
"grad_norm": 2.359992742538452,
|
| 1788 |
+
"learning_rate": 2.2325618051491415e-05,
|
| 1789 |
+
"loss": 3.2603,
|
| 1790 |
+
"step": 251
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 0.1976470588235294,
|
| 1794 |
+
"grad_norm": 1.96599280834198,
|
| 1795 |
+
"learning_rate": 2.2063229571953114e-05,
|
| 1796 |
+
"loss": 1.6396,
|
| 1797 |
+
"step": 252
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 0.1984313725490196,
|
| 1801 |
+
"grad_norm": 2.0115535259246826,
|
| 1802 |
+
"learning_rate": 2.1801680540908783e-05,
|
| 1803 |
+
"loss": 1.2482,
|
| 1804 |
+
"step": 253
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 0.1992156862745098,
|
| 1808 |
+
"grad_norm": 2.1881866455078125,
|
| 1809 |
+
"learning_rate": 2.154098792990022e-05,
|
| 1810 |
+
"loss": 1.1678,
|
| 1811 |
+
"step": 254
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 0.2,
|
| 1815 |
+
"grad_norm": 2.3712048530578613,
|
| 1816 |
+
"learning_rate": 2.1281168654897375e-05,
|
| 1817 |
+
"loss": 1.134,
|
| 1818 |
+
"step": 255
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 0.2007843137254902,
|
| 1822 |
+
"grad_norm": 2.554553985595703,
|
| 1823 |
+
"learning_rate": 2.102223957520066e-05,
|
| 1824 |
+
"loss": 1.2021,
|
| 1825 |
+
"step": 256
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 0.2015686274509804,
|
| 1829 |
+
"grad_norm": 2.6500658988952637,
|
| 1830 |
+
"learning_rate": 2.0764217492346998e-05,
|
| 1831 |
+
"loss": 1.0142,
|
| 1832 |
+
"step": 257
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 0.2023529411764706,
|
| 1836 |
+
"grad_norm": 3.106962203979492,
|
| 1837 |
+
"learning_rate": 2.0507119149019585e-05,
|
| 1838 |
+
"loss": 1.3034,
|
| 1839 |
+
"step": 258
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 0.20313725490196077,
|
| 1843 |
+
"grad_norm": 2.9345974922180176,
|
| 1844 |
+
"learning_rate": 2.0250961227961486e-05,
|
| 1845 |
+
"loss": 0.9994,
|
| 1846 |
+
"step": 259
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 0.20392156862745098,
|
| 1850 |
+
"grad_norm": 3.326606035232544,
|
| 1851 |
+
"learning_rate": 1.9995760350893095e-05,
|
| 1852 |
+
"loss": 1.2394,
|
| 1853 |
+
"step": 260
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 0.20470588235294118,
|
| 1857 |
+
"grad_norm": 2.8550965785980225,
|
| 1858 |
+
"learning_rate": 1.9741533077433642e-05,
|
| 1859 |
+
"loss": 1.1,
|
| 1860 |
+
"step": 261
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 0.2054901960784314,
|
| 1864 |
+
"grad_norm": 3.312087297439575,
|
| 1865 |
+
"learning_rate": 1.9488295904026556e-05,
|
| 1866 |
+
"loss": 1.1038,
|
| 1867 |
+
"step": 262
|
| 1868 |
+
},
|
| 1869 |
+
{
|
| 1870 |
+
"epoch": 0.20627450980392156,
|
| 1871 |
+
"grad_norm": 3.6784210205078125,
|
| 1872 |
+
"learning_rate": 1.9236065262869107e-05,
|
| 1873 |
+
"loss": 1.1669,
|
| 1874 |
+
"step": 263
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.20705882352941177,
|
| 1878 |
+
"grad_norm": 4.164744853973389,
|
| 1879 |
+
"learning_rate": 1.8984857520846164e-05,
|
| 1880 |
+
"loss": 1.5447,
|
| 1881 |
+
"step": 264
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 0.20784313725490197,
|
| 1885 |
+
"grad_norm": 3.8596198558807373,
|
| 1886 |
+
"learning_rate": 1.87346889784681e-05,
|
| 1887 |
+
"loss": 1.3421,
|
| 1888 |
+
"step": 265
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 0.20862745098039215,
|
| 1892 |
+
"grad_norm": 4.981409549713135,
|
| 1893 |
+
"learning_rate": 1.848557586881313e-05,
|
| 1894 |
+
"loss": 1.5625,
|
| 1895 |
+
"step": 266
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 0.20941176470588235,
|
| 1899 |
+
"grad_norm": 5.043278694152832,
|
| 1900 |
+
"learning_rate": 1.823753435647393e-05,
|
| 1901 |
+
"loss": 1.1419,
|
| 1902 |
+
"step": 267
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 0.21019607843137256,
|
| 1906 |
+
"grad_norm": 5.383882999420166,
|
| 1907 |
+
"learning_rate": 1.7990580536508787e-05,
|
| 1908 |
+
"loss": 1.6431,
|
| 1909 |
+
"step": 268
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 0.21098039215686273,
|
| 1913 |
+
"grad_norm": 3.967205286026001,
|
| 1914 |
+
"learning_rate": 1.774473043339719e-05,
|
| 1915 |
+
"loss": 1.053,
|
| 1916 |
+
"step": 269
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 0.21176470588235294,
|
| 1920 |
+
"grad_norm": 4.5731072425842285,
|
| 1921 |
+
"learning_rate": 1.7500000000000005e-05,
|
| 1922 |
+
"loss": 1.435,
|
| 1923 |
+
"step": 270
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 0.21254901960784314,
|
| 1927 |
+
"grad_norm": 4.236204147338867,
|
| 1928 |
+
"learning_rate": 1.7256405116524344e-05,
|
| 1929 |
+
"loss": 0.9911,
|
| 1930 |
+
"step": 271
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 0.21333333333333335,
|
| 1934 |
+
"grad_norm": 4.109267711639404,
|
| 1935 |
+
"learning_rate": 1.7013961589493098e-05,
|
| 1936 |
+
"loss": 0.8052,
|
| 1937 |
+
"step": 272
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 0.21411764705882352,
|
| 1941 |
+
"grad_norm": 4.672120571136475,
|
| 1942 |
+
"learning_rate": 1.6772685150719313e-05,
|
| 1943 |
+
"loss": 1.013,
|
| 1944 |
+
"step": 273
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 0.21490196078431373,
|
| 1948 |
+
"grad_norm": 6.016241550445557,
|
| 1949 |
+
"learning_rate": 1.653259145628535e-05,
|
| 1950 |
+
"loss": 1.5547,
|
| 1951 |
+
"step": 274
|
| 1952 |
+
},
|
| 1953 |
+
{
|
| 1954 |
+
"epoch": 0.21568627450980393,
|
| 1955 |
+
"grad_norm": 4.318469524383545,
|
| 1956 |
+
"learning_rate": 1.6293696085526958e-05,
|
| 1957 |
+
"loss": 1.0291,
|
| 1958 |
+
"step": 275
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 0.2164705882352941,
|
| 1962 |
+
"grad_norm": 5.033075332641602,
|
| 1963 |
+
"learning_rate": 1.6056014540022405e-05,
|
| 1964 |
+
"loss": 0.8338,
|
| 1965 |
+
"step": 276
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 0.2172549019607843,
|
| 1969 |
+
"grad_norm": 5.945756912231445,
|
| 1970 |
+
"learning_rate": 1.5819562242586556e-05,
|
| 1971 |
+
"loss": 1.2347,
|
| 1972 |
+
"step": 277
|
| 1973 |
+
},
|
| 1974 |
+
{
|
| 1975 |
+
"epoch": 0.21803921568627452,
|
| 1976 |
+
"grad_norm": 5.846539497375488,
|
| 1977 |
+
"learning_rate": 1.5584354536270135e-05,
|
| 1978 |
+
"loss": 0.9788,
|
| 1979 |
+
"step": 278
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 0.2188235294117647,
|
| 1983 |
+
"grad_norm": 4.702358245849609,
|
| 1984 |
+
"learning_rate": 1.535040668336417e-05,
|
| 1985 |
+
"loss": 0.9638,
|
| 1986 |
+
"step": 279
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 0.2196078431372549,
|
| 1990 |
+
"grad_norm": 5.319607734680176,
|
| 1991 |
+
"learning_rate": 1.511773386440955e-05,
|
| 1992 |
+
"loss": 1.1117,
|
| 1993 |
+
"step": 280
|
| 1994 |
+
},
|
| 1995 |
+
{
|
| 1996 |
+
"epoch": 0.2203921568627451,
|
| 1997 |
+
"grad_norm": 4.019665241241455,
|
| 1998 |
+
"learning_rate": 1.4886351177212074e-05,
|
| 1999 |
+
"loss": 0.6134,
|
| 2000 |
+
"step": 281
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 0.2211764705882353,
|
| 2004 |
+
"grad_norm": 5.152523994445801,
|
| 2005 |
+
"learning_rate": 1.4656273635862728e-05,
|
| 2006 |
+
"loss": 1.053,
|
| 2007 |
+
"step": 282
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 0.22196078431372548,
|
| 2011 |
+
"grad_norm": 6.206692218780518,
|
| 2012 |
+
"learning_rate": 1.4427516169763444e-05,
|
| 2013 |
+
"loss": 1.6169,
|
| 2014 |
+
"step": 283
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 0.2227450980392157,
|
| 2018 |
+
"grad_norm": 5.860468864440918,
|
| 2019 |
+
"learning_rate": 1.4200093622658394e-05,
|
| 2020 |
+
"loss": 1.1937,
|
| 2021 |
+
"step": 284
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 0.2235294117647059,
|
| 2025 |
+
"grad_norm": 5.6560773849487305,
|
| 2026 |
+
"learning_rate": 1.3974020751670732e-05,
|
| 2027 |
+
"loss": 1.4664,
|
| 2028 |
+
"step": 285
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 0.22431372549019607,
|
| 2032 |
+
"grad_norm": 7.365577220916748,
|
| 2033 |
+
"learning_rate": 1.3749312226345108e-05,
|
| 2034 |
+
"loss": 1.4524,
|
| 2035 |
+
"step": 286
|
| 2036 |
+
},
|
| 2037 |
+
{
|
| 2038 |
+
"epoch": 0.22509803921568627,
|
| 2039 |
+
"grad_norm": 6.791868686676025,
|
| 2040 |
+
"learning_rate": 1.3525982627695691e-05,
|
| 2041 |
+
"loss": 0.9192,
|
| 2042 |
+
"step": 287
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 0.22588235294117648,
|
| 2046 |
+
"grad_norm": 7.1987080574035645,
|
| 2047 |
+
"learning_rate": 1.3304046447260085e-05,
|
| 2048 |
+
"loss": 1.5558,
|
| 2049 |
+
"step": 288
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 0.22666666666666666,
|
| 2053 |
+
"grad_norm": 8.44493293762207,
|
| 2054 |
+
"learning_rate": 1.308351808615902e-05,
|
| 2055 |
+
"loss": 1.6203,
|
| 2056 |
+
"step": 289
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 0.22745098039215686,
|
| 2060 |
+
"grad_norm": 7.552881240844727,
|
| 2061 |
+
"learning_rate": 1.2864411854161799e-05,
|
| 2062 |
+
"loss": 1.6813,
|
| 2063 |
+
"step": 290
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 0.22823529411764706,
|
| 2067 |
+
"grad_norm": 9.73569393157959,
|
| 2068 |
+
"learning_rate": 1.2646741968757827e-05,
|
| 2069 |
+
"loss": 1.6688,
|
| 2070 |
+
"step": 291
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 0.22901960784313727,
|
| 2074 |
+
"grad_norm": 8.195075035095215,
|
| 2075 |
+
"learning_rate": 1.2430522554234036e-05,
|
| 2076 |
+
"loss": 1.8058,
|
| 2077 |
+
"step": 292
|
| 2078 |
+
},
|
| 2079 |
+
{
|
| 2080 |
+
"epoch": 0.22980392156862745,
|
| 2081 |
+
"grad_norm": 10.075333595275879,
|
| 2082 |
+
"learning_rate": 1.2215767640758367e-05,
|
| 2083 |
+
"loss": 1.983,
|
| 2084 |
+
"step": 293
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 0.23058823529411765,
|
| 2088 |
+
"grad_norm": 9.949265480041504,
|
| 2089 |
+
"learning_rate": 1.2002491163469428e-05,
|
| 2090 |
+
"loss": 1.9262,
|
| 2091 |
+
"step": 294
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 0.23137254901960785,
|
| 2095 |
+
"grad_norm": 9.431818962097168,
|
| 2096 |
+
"learning_rate": 1.1790706961572175e-05,
|
| 2097 |
+
"loss": 2.0562,
|
| 2098 |
+
"step": 295
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 0.23215686274509803,
|
| 2102 |
+
"grad_norm": 21.495634078979492,
|
| 2103 |
+
"learning_rate": 1.1580428777439972e-05,
|
| 2104 |
+
"loss": 1.9353,
|
| 2105 |
+
"step": 296
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 0.23294117647058823,
|
| 2109 |
+
"grad_norm": 9.102306365966797,
|
| 2110 |
+
"learning_rate": 1.1371670255722838e-05,
|
| 2111 |
+
"loss": 1.6257,
|
| 2112 |
+
"step": 297
|
| 2113 |
+
},
|
| 2114 |
+
{
|
| 2115 |
+
"epoch": 0.23372549019607844,
|
| 2116 |
+
"grad_norm": 14.28601360321045,
|
| 2117 |
+
"learning_rate": 1.1164444942462095e-05,
|
| 2118 |
+
"loss": 2.8148,
|
| 2119 |
+
"step": 298
|
| 2120 |
+
},
|
| 2121 |
+
{
|
| 2122 |
+
"epoch": 0.23450980392156862,
|
| 2123 |
+
"grad_norm": 18.514936447143555,
|
| 2124 |
+
"learning_rate": 1.0958766284211363e-05,
|
| 2125 |
+
"loss": 2.4219,
|
| 2126 |
+
"step": 299
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 0.23529411764705882,
|
| 2130 |
+
"grad_norm": 17.440731048583984,
|
| 2131 |
+
"learning_rate": 1.0754647627164022e-05,
|
| 2132 |
+
"loss": 2.8099,
|
| 2133 |
+
"step": 300
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.23529411764705882,
|
| 2137 |
+
"eval_loss": 0.35413047671318054,
|
| 2138 |
+
"eval_runtime": 168.8442,
|
| 2139 |
+
"eval_samples_per_second": 12.722,
|
| 2140 |
+
"eval_steps_per_second": 3.18,
|
| 2141 |
+
"step": 300
|
| 2142 |
}
|
| 2143 |
],
|
| 2144 |
"logging_steps": 1,
|
|
|
|
| 2167 |
"attributes": {}
|
| 2168 |
}
|
| 2169 |
},
|
| 2170 |
+
"total_flos": 4.536268794133217e+17,
|
| 2171 |
"train_batch_size": 8,
|
| 2172 |
"trial_name": null,
|
| 2173 |
"trial_params": null
|