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 639691872
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:db1af205eff0d1c3981f58af318310bd969f0ccee8e313b79ebead6997c843b5
|
| 3 |
size 639691872
|
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 sha256:8cf0977855e53e7a960c7caf7aa4b8b6575b7d1cea78d007041fe7b74b13d404
|
| 3 |
+
size 325340244
|
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:a30f2d89b79bfb2d9929f2dc3ffeef086eb1500788503382d980f99cbe057e80
|
| 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 sha256:825ef3e3db682363455a0d008a860e7cc4412a53aa533791e6b37fda9dca6312
|
| 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": 6.278,
|
| 1432 |
"eval_steps_per_second": 1.569,
|
| 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": 1.556799054145813,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-300",
|
| 4 |
+
"epoch": 0.8310249307479224,
|
| 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": 6.278,
|
| 1432 |
"eval_steps_per_second": 1.569,
|
| 1433 |
"step": 200
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 0.556786703601108,
|
| 1437 |
+
"grad_norm": 0.4485991597175598,
|
| 1438 |
+
"learning_rate": 4.517077709232411e-05,
|
| 1439 |
+
"loss": 1.7589,
|
| 1440 |
+
"step": 201
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 0.5595567867036011,
|
| 1444 |
+
"grad_norm": 0.4756779372692108,
|
| 1445 |
+
"learning_rate": 4.471249804477758e-05,
|
| 1446 |
+
"loss": 1.7589,
|
| 1447 |
+
"step": 202
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 0.5623268698060941,
|
| 1451 |
+
"grad_norm": 0.445344477891922,
|
| 1452 |
+
"learning_rate": 4.4254667782358924e-05,
|
| 1453 |
+
"loss": 1.6403,
|
| 1454 |
+
"step": 203
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 0.5650969529085873,
|
| 1458 |
+
"grad_norm": 0.448127806186676,
|
| 1459 |
+
"learning_rate": 4.379732516413897e-05,
|
| 1460 |
+
"loss": 1.6286,
|
| 1461 |
+
"step": 204
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 0.5678670360110804,
|
| 1465 |
+
"grad_norm": 0.4261086881160736,
|
| 1466 |
+
"learning_rate": 4.334050900779893e-05,
|
| 1467 |
+
"loss": 1.6222,
|
| 1468 |
+
"step": 205
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 0.5706371191135734,
|
| 1472 |
+
"grad_norm": 0.422758013010025,
|
| 1473 |
+
"learning_rate": 4.288425808633575e-05,
|
| 1474 |
+
"loss": 1.5906,
|
| 1475 |
+
"step": 206
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 0.5734072022160664,
|
| 1479 |
+
"grad_norm": 0.39392560720443726,
|
| 1480 |
+
"learning_rate": 4.2428611124771184e-05,
|
| 1481 |
+
"loss": 1.6972,
|
| 1482 |
+
"step": 207
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 0.5761772853185596,
|
| 1486 |
+
"grad_norm": 0.40653908252716064,
|
| 1487 |
+
"learning_rate": 4.1973606796864884e-05,
|
| 1488 |
+
"loss": 1.5745,
|
| 1489 |
+
"step": 208
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 0.5789473684210527,
|
| 1493 |
+
"grad_norm": 0.4077155888080597,
|
| 1494 |
+
"learning_rate": 4.151928372183198e-05,
|
| 1495 |
+
"loss": 1.5567,
|
| 1496 |
+
"step": 209
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 0.5817174515235457,
|
| 1500 |
+
"grad_norm": 0.4133901596069336,
|
| 1501 |
+
"learning_rate": 4.1065680461065194e-05,
|
| 1502 |
+
"loss": 1.5616,
|
| 1503 |
+
"step": 210
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 0.5844875346260388,
|
| 1507 |
+
"grad_norm": 0.4191035330295563,
|
| 1508 |
+
"learning_rate": 4.061283551486185e-05,
|
| 1509 |
+
"loss": 1.6254,
|
| 1510 |
+
"step": 211
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 0.5872576177285319,
|
| 1514 |
+
"grad_norm": 0.4195918142795563,
|
| 1515 |
+
"learning_rate": 4.016078731915608e-05,
|
| 1516 |
+
"loss": 1.5852,
|
| 1517 |
+
"step": 212
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 0.590027700831025,
|
| 1521 |
+
"grad_norm": 0.4277445375919342,
|
| 1522 |
+
"learning_rate": 3.970957424225666e-05,
|
| 1523 |
+
"loss": 1.5839,
|
| 1524 |
+
"step": 213
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 0.592797783933518,
|
| 1528 |
+
"grad_norm": 0.4163782596588135,
|
| 1529 |
+
"learning_rate": 3.925923458159023e-05,
|
| 1530 |
+
"loss": 1.6141,
|
| 1531 |
+
"step": 214
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 0.5955678670360111,
|
| 1535 |
+
"grad_norm": 0.42547765374183655,
|
| 1536 |
+
"learning_rate": 3.880980656045087e-05,
|
| 1537 |
+
"loss": 1.5977,
|
| 1538 |
+
"step": 215
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 0.5983379501385041,
|
| 1542 |
+
"grad_norm": 0.4501667320728302,
|
| 1543 |
+
"learning_rate": 3.8361328324755825e-05,
|
| 1544 |
+
"loss": 1.5311,
|
| 1545 |
+
"step": 216
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 0.6011080332409973,
|
| 1549 |
+
"grad_norm": 0.43693989515304565,
|
| 1550 |
+
"learning_rate": 3.791383793980776e-05,
|
| 1551 |
+
"loss": 1.5779,
|
| 1552 |
+
"step": 217
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 0.6038781163434903,
|
| 1556 |
+
"grad_norm": 0.4550263285636902,
|
| 1557 |
+
"learning_rate": 3.746737338706397e-05,
|
| 1558 |
+
"loss": 1.5961,
|
| 1559 |
+
"step": 218
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 0.6066481994459834,
|
| 1563 |
+
"grad_norm": 0.4572041928768158,
|
| 1564 |
+
"learning_rate": 3.70219725609126e-05,
|
| 1565 |
+
"loss": 1.592,
|
| 1566 |
+
"step": 219
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 0.6094182825484764,
|
| 1570 |
+
"grad_norm": 0.4633628726005554,
|
| 1571 |
+
"learning_rate": 3.65776732654563e-05,
|
| 1572 |
+
"loss": 1.5822,
|
| 1573 |
+
"step": 220
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 0.6121883656509696,
|
| 1577 |
+
"grad_norm": 0.4722716808319092,
|
| 1578 |
+
"learning_rate": 3.6134513211303556e-05,
|
| 1579 |
+
"loss": 1.5399,
|
| 1580 |
+
"step": 221
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 0.6149584487534626,
|
| 1584 |
+
"grad_norm": 0.477970689535141,
|
| 1585 |
+
"learning_rate": 3.5692530012367955e-05,
|
| 1586 |
+
"loss": 1.5923,
|
| 1587 |
+
"step": 222
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 0.6177285318559557,
|
| 1591 |
+
"grad_norm": 0.4831191599369049,
|
| 1592 |
+
"learning_rate": 3.5251761182675625e-05,
|
| 1593 |
+
"loss": 1.559,
|
| 1594 |
+
"step": 223
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 0.6204986149584487,
|
| 1598 |
+
"grad_norm": 0.46263793110847473,
|
| 1599 |
+
"learning_rate": 3.481224413318114e-05,
|
| 1600 |
+
"loss": 1.5255,
|
| 1601 |
+
"step": 224
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 0.6232686980609419,
|
| 1605 |
+
"grad_norm": 0.4732118844985962,
|
| 1606 |
+
"learning_rate": 3.4374016168592296e-05,
|
| 1607 |
+
"loss": 1.5777,
|
| 1608 |
+
"step": 225
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 0.6260387811634349,
|
| 1612 |
+
"grad_norm": 0.47649428248405457,
|
| 1613 |
+
"learning_rate": 3.393711448420372e-05,
|
| 1614 |
+
"loss": 1.4918,
|
| 1615 |
+
"step": 226
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 0.628808864265928,
|
| 1619 |
+
"grad_norm": 0.4683854579925537,
|
| 1620 |
+
"learning_rate": 3.3501576162739904e-05,
|
| 1621 |
+
"loss": 1.4874,
|
| 1622 |
+
"step": 227
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 0.631578947368421,
|
| 1626 |
+
"grad_norm": 0.4791397750377655,
|
| 1627 |
+
"learning_rate": 3.3067438171207766e-05,
|
| 1628 |
+
"loss": 1.55,
|
| 1629 |
+
"step": 228
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 0.6343490304709142,
|
| 1633 |
+
"grad_norm": 0.474380224943161,
|
| 1634 |
+
"learning_rate": 3.263473735775899e-05,
|
| 1635 |
+
"loss": 1.5044,
|
| 1636 |
+
"step": 229
|
| 1637 |
+
},
|
| 1638 |
+
{
|
| 1639 |
+
"epoch": 0.6371191135734072,
|
| 1640 |
+
"grad_norm": 0.4925787150859833,
|
| 1641 |
+
"learning_rate": 3.220351044856247e-05,
|
| 1642 |
+
"loss": 1.5274,
|
| 1643 |
+
"step": 230
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 0.6398891966759003,
|
| 1647 |
+
"grad_norm": 0.4613766670227051,
|
| 1648 |
+
"learning_rate": 3.177379404468715e-05,
|
| 1649 |
+
"loss": 1.4617,
|
| 1650 |
+
"step": 231
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 0.6426592797783933,
|
| 1654 |
+
"grad_norm": 0.4807930588722229,
|
| 1655 |
+
"learning_rate": 3.134562461899545e-05,
|
| 1656 |
+
"loss": 1.4273,
|
| 1657 |
+
"step": 232
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 0.6454293628808865,
|
| 1661 |
+
"grad_norm": 0.49378934502601624,
|
| 1662 |
+
"learning_rate": 3.091903851304751e-05,
|
| 1663 |
+
"loss": 1.5487,
|
| 1664 |
+
"step": 233
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 0.6481994459833795,
|
| 1668 |
+
"grad_norm": 0.5002757906913757,
|
| 1669 |
+
"learning_rate": 3.0494071934016737e-05,
|
| 1670 |
+
"loss": 1.4499,
|
| 1671 |
+
"step": 234
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 0.6509695290858726,
|
| 1675 |
+
"grad_norm": 0.49733448028564453,
|
| 1676 |
+
"learning_rate": 3.0070760951616618e-05,
|
| 1677 |
+
"loss": 1.4476,
|
| 1678 |
+
"step": 235
|
| 1679 |
+
},
|
| 1680 |
+
{
|
| 1681 |
+
"epoch": 0.6537396121883656,
|
| 1682 |
+
"grad_norm": 0.5034710764884949,
|
| 1683 |
+
"learning_rate": 2.9649141495039223e-05,
|
| 1684 |
+
"loss": 1.5502,
|
| 1685 |
+
"step": 236
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 0.6565096952908587,
|
| 1689 |
+
"grad_norm": 0.49712780117988586,
|
| 1690 |
+
"learning_rate": 2.9229249349905684e-05,
|
| 1691 |
+
"loss": 1.4659,
|
| 1692 |
+
"step": 237
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 0.6592797783933518,
|
| 1696 |
+
"grad_norm": 0.5008156895637512,
|
| 1697 |
+
"learning_rate": 2.8811120155228844e-05,
|
| 1698 |
+
"loss": 1.4596,
|
| 1699 |
+
"step": 238
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 0.6620498614958449,
|
| 1703 |
+
"grad_norm": 0.5352665185928345,
|
| 1704 |
+
"learning_rate": 2.8394789400388328e-05,
|
| 1705 |
+
"loss": 1.5186,
|
| 1706 |
+
"step": 239
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 0.6648199445983379,
|
| 1710 |
+
"grad_norm": 0.5389817357063293,
|
| 1711 |
+
"learning_rate": 2.798029242211828e-05,
|
| 1712 |
+
"loss": 1.5449,
|
| 1713 |
+
"step": 240
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 0.667590027700831,
|
| 1717 |
+
"grad_norm": 0.5315467715263367,
|
| 1718 |
+
"learning_rate": 2.7567664401508225e-05,
|
| 1719 |
+
"loss": 1.4991,
|
| 1720 |
+
"step": 241
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 0.6703601108033241,
|
| 1724 |
+
"grad_norm": 0.5424023270606995,
|
| 1725 |
+
"learning_rate": 2.7156940361016864e-05,
|
| 1726 |
+
"loss": 1.528,
|
| 1727 |
+
"step": 242
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 0.6731301939058172,
|
| 1731 |
+
"grad_norm": 0.5434011816978455,
|
| 1732 |
+
"learning_rate": 2.6748155161499567e-05,
|
| 1733 |
+
"loss": 1.4446,
|
| 1734 |
+
"step": 243
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 0.6759002770083102,
|
| 1738 |
+
"grad_norm": 0.5452317595481873,
|
| 1739 |
+
"learning_rate": 2.634134349924956e-05,
|
| 1740 |
+
"loss": 1.36,
|
| 1741 |
+
"step": 244
|
| 1742 |
+
},
|
| 1743 |
+
{
|
| 1744 |
+
"epoch": 0.6786703601108033,
|
| 1745 |
+
"grad_norm": 0.5486112833023071,
|
| 1746 |
+
"learning_rate": 2.5936539903052892e-05,
|
| 1747 |
+
"loss": 1.4738,
|
| 1748 |
+
"step": 245
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 0.6814404432132964,
|
| 1752 |
+
"grad_norm": 0.582291841506958,
|
| 1753 |
+
"learning_rate": 2.5533778731257824e-05,
|
| 1754 |
+
"loss": 1.4956,
|
| 1755 |
+
"step": 246
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 0.6842105263157895,
|
| 1759 |
+
"grad_norm": 0.5929666757583618,
|
| 1760 |
+
"learning_rate": 2.513309416885865e-05,
|
| 1761 |
+
"loss": 1.5308,
|
| 1762 |
+
"step": 247
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 0.6869806094182825,
|
| 1766 |
+
"grad_norm": 0.5847265124320984,
|
| 1767 |
+
"learning_rate": 2.4734520224594093e-05,
|
| 1768 |
+
"loss": 1.5132,
|
| 1769 |
+
"step": 248
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 0.6897506925207756,
|
| 1773 |
+
"grad_norm": 0.7136350274085999,
|
| 1774 |
+
"learning_rate": 2.433809072806081e-05,
|
| 1775 |
+
"loss": 1.8617,
|
| 1776 |
+
"step": 249
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 0.6925207756232687,
|
| 1780 |
+
"grad_norm": 1.0075485706329346,
|
| 1781 |
+
"learning_rate": 2.3943839326842092e-05,
|
| 1782 |
+
"loss": 2.1937,
|
| 1783 |
+
"step": 250
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 0.6952908587257618,
|
| 1787 |
+
"grad_norm": 0.42398756742477417,
|
| 1788 |
+
"learning_rate": 2.3551799483651894e-05,
|
| 1789 |
+
"loss": 1.7047,
|
| 1790 |
+
"step": 251
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 0.6980609418282548,
|
| 1794 |
+
"grad_norm": 0.4284769296646118,
|
| 1795 |
+
"learning_rate": 2.3162004473494657e-05,
|
| 1796 |
+
"loss": 1.6634,
|
| 1797 |
+
"step": 252
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 0.7008310249307479,
|
| 1801 |
+
"grad_norm": 0.4206549823284149,
|
| 1802 |
+
"learning_rate": 2.2774487380841115e-05,
|
| 1803 |
+
"loss": 1.6303,
|
| 1804 |
+
"step": 253
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 0.703601108033241,
|
| 1808 |
+
"grad_norm": 0.41631007194519043,
|
| 1809 |
+
"learning_rate": 2.2389281096820075e-05,
|
| 1810 |
+
"loss": 1.6223,
|
| 1811 |
+
"step": 254
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 0.7063711911357341,
|
| 1815 |
+
"grad_norm": 0.4172951579093933,
|
| 1816 |
+
"learning_rate": 2.2006418316426775e-05,
|
| 1817 |
+
"loss": 1.6146,
|
| 1818 |
+
"step": 255
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 0.7091412742382271,
|
| 1822 |
+
"grad_norm": 0.4219145178794861,
|
| 1823 |
+
"learning_rate": 2.1625931535747964e-05,
|
| 1824 |
+
"loss": 1.5774,
|
| 1825 |
+
"step": 256
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 0.7119113573407202,
|
| 1829 |
+
"grad_norm": 0.41025567054748535,
|
| 1830 |
+
"learning_rate": 2.1247853049203543e-05,
|
| 1831 |
+
"loss": 1.6285,
|
| 1832 |
+
"step": 257
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 0.7146814404432132,
|
| 1836 |
+
"grad_norm": 0.4165142774581909,
|
| 1837 |
+
"learning_rate": 2.087221494680563e-05,
|
| 1838 |
+
"loss": 1.6049,
|
| 1839 |
+
"step": 258
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 0.7174515235457064,
|
| 1843 |
+
"grad_norm": 0.4020419418811798,
|
| 1844 |
+
"learning_rate": 2.049904911143492e-05,
|
| 1845 |
+
"loss": 1.5822,
|
| 1846 |
+
"step": 259
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 0.7202216066481995,
|
| 1850 |
+
"grad_norm": 0.39877504110336304,
|
| 1851 |
+
"learning_rate": 2.012838721613447e-05,
|
| 1852 |
+
"loss": 1.5962,
|
| 1853 |
+
"step": 260
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 0.7229916897506925,
|
| 1857 |
+
"grad_norm": 0.404995858669281,
|
| 1858 |
+
"learning_rate": 1.9760260721421426e-05,
|
| 1859 |
+
"loss": 1.61,
|
| 1860 |
+
"step": 261
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 0.7257617728531855,
|
| 1864 |
+
"grad_norm": 0.41909393668174744,
|
| 1865 |
+
"learning_rate": 1.9394700872616855e-05,
|
| 1866 |
+
"loss": 1.5494,
|
| 1867 |
+
"step": 262
|
| 1868 |
+
},
|
| 1869 |
+
{
|
| 1870 |
+
"epoch": 0.7285318559556787,
|
| 1871 |
+
"grad_norm": 0.42204612493515015,
|
| 1872 |
+
"learning_rate": 1.903173869719362e-05,
|
| 1873 |
+
"loss": 1.673,
|
| 1874 |
+
"step": 263
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.7313019390581718,
|
| 1878 |
+
"grad_norm": 0.4202198088169098,
|
| 1879 |
+
"learning_rate": 1.8671405002142918e-05,
|
| 1880 |
+
"loss": 1.5661,
|
| 1881 |
+
"step": 264
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 0.7340720221606648,
|
| 1885 |
+
"grad_norm": 0.42738965153694153,
|
| 1886 |
+
"learning_rate": 1.831373037135955e-05,
|
| 1887 |
+
"loss": 1.5709,
|
| 1888 |
+
"step": 265
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 0.7368421052631579,
|
| 1892 |
+
"grad_norm": 0.4346977174282074,
|
| 1893 |
+
"learning_rate": 1.7958745163045986e-05,
|
| 1894 |
+
"loss": 1.5803,
|
| 1895 |
+
"step": 266
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 0.739612188365651,
|
| 1899 |
+
"grad_norm": 0.4478447437286377,
|
| 1900 |
+
"learning_rate": 1.760647950713566e-05,
|
| 1901 |
+
"loss": 1.584,
|
| 1902 |
+
"step": 267
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 0.7423822714681441,
|
| 1906 |
+
"grad_norm": 0.43284082412719727,
|
| 1907 |
+
"learning_rate": 1.725696330273575e-05,
|
| 1908 |
+
"loss": 1.5911,
|
| 1909 |
+
"step": 268
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 0.7451523545706371,
|
| 1913 |
+
"grad_norm": 0.4424854815006256,
|
| 1914 |
+
"learning_rate": 1.6910226215589303e-05,
|
| 1915 |
+
"loss": 1.5178,
|
| 1916 |
+
"step": 269
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 0.7479224376731302,
|
| 1920 |
+
"grad_norm": 0.4423474669456482,
|
| 1921 |
+
"learning_rate": 1.656629767555739e-05,
|
| 1922 |
+
"loss": 1.5337,
|
| 1923 |
+
"step": 270
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 0.7506925207756233,
|
| 1927 |
+
"grad_norm": 0.450488418340683,
|
| 1928 |
+
"learning_rate": 1.6225206874121218e-05,
|
| 1929 |
+
"loss": 1.5476,
|
| 1930 |
+
"step": 271
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 0.7534626038781164,
|
| 1934 |
+
"grad_norm": 0.4583964943885803,
|
| 1935 |
+
"learning_rate": 1.5886982761904377e-05,
|
| 1936 |
+
"loss": 1.559,
|
| 1937 |
+
"step": 272
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 0.7562326869806094,
|
| 1941 |
+
"grad_norm": 0.45635777711868286,
|
| 1942 |
+
"learning_rate": 1.555165404621567e-05,
|
| 1943 |
+
"loss": 1.5798,
|
| 1944 |
+
"step": 273
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 0.7590027700831025,
|
| 1948 |
+
"grad_norm": 0.4589531123638153,
|
| 1949 |
+
"learning_rate": 1.5219249188612556e-05,
|
| 1950 |
+
"loss": 1.5855,
|
| 1951 |
+
"step": 274
|
| 1952 |
+
},
|
| 1953 |
+
{
|
| 1954 |
+
"epoch": 0.7617728531855956,
|
| 1955 |
+
"grad_norm": 0.47425007820129395,
|
| 1956 |
+
"learning_rate": 1.488979640248534e-05,
|
| 1957 |
+
"loss": 1.565,
|
| 1958 |
+
"step": 275
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 0.7645429362880887,
|
| 1962 |
+
"grad_norm": 0.4853668808937073,
|
| 1963 |
+
"learning_rate": 1.4563323650662586e-05,
|
| 1964 |
+
"loss": 1.5671,
|
| 1965 |
+
"step": 276
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 0.7673130193905817,
|
| 1969 |
+
"grad_norm": 0.4773860275745392,
|
| 1970 |
+
"learning_rate": 1.4239858643037751e-05,
|
| 1971 |
+
"loss": 1.562,
|
| 1972 |
+
"step": 277
|
| 1973 |
+
},
|
| 1974 |
+
{
|
| 1975 |
+
"epoch": 0.7700831024930748,
|
| 1976 |
+
"grad_norm": 0.5074953436851501,
|
| 1977 |
+
"learning_rate": 1.3919428834217163e-05,
|
| 1978 |
+
"loss": 1.5222,
|
| 1979 |
+
"step": 278
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 0.7728531855955678,
|
| 1983 |
+
"grad_norm": 0.4980280101299286,
|
| 1984 |
+
"learning_rate": 1.36020614211899e-05,
|
| 1985 |
+
"loss": 1.5357,
|
| 1986 |
+
"step": 279
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 0.775623268698061,
|
| 1990 |
+
"grad_norm": 0.4977880120277405,
|
| 1991 |
+
"learning_rate": 1.3287783341019278e-05,
|
| 1992 |
+
"loss": 1.5099,
|
| 1993 |
+
"step": 280
|
| 1994 |
+
},
|
| 1995 |
+
{
|
| 1996 |
+
"epoch": 0.778393351800554,
|
| 1997 |
+
"grad_norm": 0.490857869386673,
|
| 1998 |
+
"learning_rate": 1.2976621268556571e-05,
|
| 1999 |
+
"loss": 1.4592,
|
| 2000 |
+
"step": 281
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 0.7811634349030471,
|
| 2004 |
+
"grad_norm": 0.4895660877227783,
|
| 2005 |
+
"learning_rate": 1.2668601614177017e-05,
|
| 2006 |
+
"loss": 1.5262,
|
| 2007 |
+
"step": 282
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 0.7839335180055401,
|
| 2011 |
+
"grad_norm": 0.5010288953781128,
|
| 2012 |
+
"learning_rate": 1.2363750521538064e-05,
|
| 2013 |
+
"loss": 1.4536,
|
| 2014 |
+
"step": 283
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 0.7867036011080333,
|
| 2018 |
+
"grad_norm": 0.5021648406982422,
|
| 2019 |
+
"learning_rate": 1.2062093865360458e-05,
|
| 2020 |
+
"loss": 1.4938,
|
| 2021 |
+
"step": 284
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 0.7894736842105263,
|
| 2025 |
+
"grad_norm": 0.5206820368766785,
|
| 2026 |
+
"learning_rate": 1.1763657249232107e-05,
|
| 2027 |
+
"loss": 1.5172,
|
| 2028 |
+
"step": 285
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 0.7922437673130194,
|
| 2032 |
+
"grad_norm": 0.5139452815055847,
|
| 2033 |
+
"learning_rate": 1.146846600343488e-05,
|
| 2034 |
+
"loss": 1.4832,
|
| 2035 |
+
"step": 286
|
| 2036 |
+
},
|
| 2037 |
+
{
|
| 2038 |
+
"epoch": 0.7950138504155124,
|
| 2039 |
+
"grad_norm": 0.5119965672492981,
|
| 2040 |
+
"learning_rate": 1.1176545182794674e-05,
|
| 2041 |
+
"loss": 1.5057,
|
| 2042 |
+
"step": 287
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 0.7977839335180056,
|
| 2046 |
+
"grad_norm": 0.5294527411460876,
|
| 2047 |
+
"learning_rate": 1.0887919564554894e-05,
|
| 2048 |
+
"loss": 1.4764,
|
| 2049 |
+
"step": 288
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 0.8005540166204986,
|
| 2053 |
+
"grad_norm": 0.5366680026054382,
|
| 2054 |
+
"learning_rate": 1.0602613646273374e-05,
|
| 2055 |
+
"loss": 1.4931,
|
| 2056 |
+
"step": 289
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 0.8033240997229917,
|
| 2060 |
+
"grad_norm": 0.5404147505760193,
|
| 2061 |
+
"learning_rate": 1.032065164374313e-05,
|
| 2062 |
+
"loss": 1.4244,
|
| 2063 |
+
"step": 290
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 0.8060941828254847,
|
| 2067 |
+
"grad_norm": 0.520129919052124,
|
| 2068 |
+
"learning_rate": 1.0042057488937067e-05,
|
| 2069 |
+
"loss": 1.461,
|
| 2070 |
+
"step": 291
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 0.8088642659279779,
|
| 2074 |
+
"grad_norm": 0.5262720584869385,
|
| 2075 |
+
"learning_rate": 9.766854827976617e-06,
|
| 2076 |
+
"loss": 1.4333,
|
| 2077 |
+
"step": 292
|
| 2078 |
+
},
|
| 2079 |
+
{
|
| 2080 |
+
"epoch": 0.8116343490304709,
|
| 2081 |
+
"grad_norm": 0.5492231249809265,
|
| 2082 |
+
"learning_rate": 9.495067019124792e-06,
|
| 2083 |
+
"loss": 1.5142,
|
| 2084 |
+
"step": 293
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 0.814404432132964,
|
| 2088 |
+
"grad_norm": 0.5467971563339233,
|
| 2089 |
+
"learning_rate": 9.226717130803636e-06,
|
| 2090 |
+
"loss": 1.4606,
|
| 2091 |
+
"step": 294
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 0.817174515235457,
|
| 2095 |
+
"grad_norm": 0.5795299410820007,
|
| 2096 |
+
"learning_rate": 8.961827939636196e-06,
|
| 2097 |
+
"loss": 1.4897,
|
| 2098 |
+
"step": 295
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 0.8199445983379502,
|
| 2102 |
+
"grad_norm": 0.5784875154495239,
|
| 2103 |
+
"learning_rate": 8.700421928513352e-06,
|
| 2104 |
+
"loss": 1.4938,
|
| 2105 |
+
"step": 296
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 0.8227146814404432,
|
| 2109 |
+
"grad_norm": 0.5875300765037537,
|
| 2110 |
+
"learning_rate": 8.442521284685573e-06,
|
| 2111 |
+
"loss": 1.4066,
|
| 2112 |
+
"step": 297
|
| 2113 |
+
},
|
| 2114 |
+
{
|
| 2115 |
+
"epoch": 0.8254847645429363,
|
| 2116 |
+
"grad_norm": 0.6010243892669678,
|
| 2117 |
+
"learning_rate": 8.188147897879667e-06,
|
| 2118 |
+
"loss": 1.482,
|
| 2119 |
+
"step": 298
|
| 2120 |
+
},
|
| 2121 |
+
{
|
| 2122 |
+
"epoch": 0.8282548476454293,
|
| 2123 |
+
"grad_norm": 0.6733506321907043,
|
| 2124 |
+
"learning_rate": 7.937323358440935e-06,
|
| 2125 |
+
"loss": 1.6978,
|
| 2126 |
+
"step": 299
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 0.8310249307479224,
|
| 2130 |
+
"grad_norm": 0.987700879573822,
|
| 2131 |
+
"learning_rate": 7.690068955500624e-06,
|
| 2132 |
+
"loss": 2.2272,
|
| 2133 |
+
"step": 300
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.8310249307479224,
|
| 2137 |
+
"eval_loss": 1.556799054145813,
|
| 2138 |
+
"eval_runtime": 95.6603,
|
| 2139 |
+
"eval_samples_per_second": 6.356,
|
| 2140 |
+
"eval_steps_per_second": 1.589,
|
| 2141 |
+
"step": 300
|
| 2142 |
}
|
| 2143 |
],
|
| 2144 |
"logging_steps": 1,
|
|
|
|
| 2167 |
"attributes": {}
|
| 2168 |
}
|
| 2169 |
},
|
| 2170 |
+
"total_flos": 4.281474924675072e+17,
|
| 2171 |
"train_batch_size": 8,
|
| 2172 |
"trial_name": null,
|
| 2173 |
"trial_params": null
|