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 35237104
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8badaa52ca519105a5380878a96d36847cb55d906e0d2e705f3717f2655e7067
|
| 3 |
size 35237104
|
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:caec36497ef1581c061792ab4a36c00600adbe0322f14b4febe97fd488c40641
|
| 3 |
+
size 18810356
|
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:d23a005f828083a9add1710a8dc106dbb2334d4fcd07e84318a29c2a3ef22527
|
| 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:ea211bfadd2ed8fd9c0d8cd735af3562d7b6f16823ff7f40e995c8863ea26ad0
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 3.
|
| 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": 58.69,
|
| 1432 |
"eval_steps_per_second": 14.673,
|
| 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": 3.1460793018341064,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-300",
|
| 4 |
+
"epoch": 0.19458407653640344,
|
| 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": 58.69,
|
| 1432 |
"eval_steps_per_second": 14.673,
|
| 1433 |
"step": 200
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 0.1303713312793903,
|
| 1437 |
+
"grad_norm": 1.5361113548278809,
|
| 1438 |
+
"learning_rate": 0.00019810089346802346,
|
| 1439 |
+
"loss": 2.9761,
|
| 1440 |
+
"step": 201
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 0.131019944867845,
|
| 1444 |
+
"grad_norm": 1.7124556303024292,
|
| 1445 |
+
"learning_rate": 0.0001980810194908336,
|
| 1446 |
+
"loss": 3.1084,
|
| 1447 |
+
"step": 202
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 0.13166855845629966,
|
| 1451 |
+
"grad_norm": 1.9076920747756958,
|
| 1452 |
+
"learning_rate": 0.00019806104307190222,
|
| 1453 |
+
"loss": 3.4621,
|
| 1454 |
+
"step": 203
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 0.13231717204475435,
|
| 1458 |
+
"grad_norm": 1.381435751914978,
|
| 1459 |
+
"learning_rate": 0.00019804096423209398,
|
| 1460 |
+
"loss": 2.7652,
|
| 1461 |
+
"step": 204
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 0.13296578563320902,
|
| 1465 |
+
"grad_norm": 1.352182149887085,
|
| 1466 |
+
"learning_rate": 0.00019802078299238044,
|
| 1467 |
+
"loss": 2.717,
|
| 1468 |
+
"step": 205
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 0.13361439922166368,
|
| 1472 |
+
"grad_norm": 1.5664377212524414,
|
| 1473 |
+
"learning_rate": 0.00019800049937384004,
|
| 1474 |
+
"loss": 2.9769,
|
| 1475 |
+
"step": 206
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 0.13426301281011838,
|
| 1479 |
+
"grad_norm": 1.786026120185852,
|
| 1480 |
+
"learning_rate": 0.00019798011339765826,
|
| 1481 |
+
"loss": 3.341,
|
| 1482 |
+
"step": 207
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 0.13491162639857304,
|
| 1486 |
+
"grad_norm": 1.4288690090179443,
|
| 1487 |
+
"learning_rate": 0.00019795962508512742,
|
| 1488 |
+
"loss": 2.676,
|
| 1489 |
+
"step": 208
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 0.13556023998702774,
|
| 1493 |
+
"grad_norm": 1.6291303634643555,
|
| 1494 |
+
"learning_rate": 0.00019793903445764675,
|
| 1495 |
+
"loss": 3.1092,
|
| 1496 |
+
"step": 209
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 0.1362088535754824,
|
| 1500 |
+
"grad_norm": 1.7784898281097412,
|
| 1501 |
+
"learning_rate": 0.0001979183415367224,
|
| 1502 |
+
"loss": 3.4292,
|
| 1503 |
+
"step": 210
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 0.1368574671639371,
|
| 1507 |
+
"grad_norm": 1.7724665403366089,
|
| 1508 |
+
"learning_rate": 0.00019789754634396724,
|
| 1509 |
+
"loss": 3.3831,
|
| 1510 |
+
"step": 211
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 0.13750608075239176,
|
| 1514 |
+
"grad_norm": 1.4297045469284058,
|
| 1515 |
+
"learning_rate": 0.00019787664890110108,
|
| 1516 |
+
"loss": 2.8106,
|
| 1517 |
+
"step": 212
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 0.13815469434084643,
|
| 1521 |
+
"grad_norm": 1.650215983390808,
|
| 1522 |
+
"learning_rate": 0.0001978556492299504,
|
| 1523 |
+
"loss": 3.1152,
|
| 1524 |
+
"step": 213
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 0.13880330792930112,
|
| 1528 |
+
"grad_norm": 1.4097318649291992,
|
| 1529 |
+
"learning_rate": 0.0001978345473524486,
|
| 1530 |
+
"loss": 3.1504,
|
| 1531 |
+
"step": 214
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 0.1394519215177558,
|
| 1535 |
+
"grad_norm": 1.5713872909545898,
|
| 1536 |
+
"learning_rate": 0.00019781334329063572,
|
| 1537 |
+
"loss": 3.5687,
|
| 1538 |
+
"step": 215
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 0.14010053510621048,
|
| 1542 |
+
"grad_norm": 1.2881985902786255,
|
| 1543 |
+
"learning_rate": 0.00019779203706665857,
|
| 1544 |
+
"loss": 3.0303,
|
| 1545 |
+
"step": 216
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 0.14074914869466515,
|
| 1549 |
+
"grad_norm": 1.3899292945861816,
|
| 1550 |
+
"learning_rate": 0.00019777062870277064,
|
| 1551 |
+
"loss": 2.7401,
|
| 1552 |
+
"step": 217
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 0.14139776228311984,
|
| 1556 |
+
"grad_norm": 1.5206934213638306,
|
| 1557 |
+
"learning_rate": 0.00019774911822133216,
|
| 1558 |
+
"loss": 2.8458,
|
| 1559 |
+
"step": 218
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 0.1420463758715745,
|
| 1563 |
+
"grad_norm": 1.777827262878418,
|
| 1564 |
+
"learning_rate": 0.00019772750564480993,
|
| 1565 |
+
"loss": 3.1399,
|
| 1566 |
+
"step": 219
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 0.14269498946002918,
|
| 1570 |
+
"grad_norm": 1.5830134153366089,
|
| 1571 |
+
"learning_rate": 0.00019770579099577743,
|
| 1572 |
+
"loss": 2.9685,
|
| 1573 |
+
"step": 220
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 0.14334360304848387,
|
| 1577 |
+
"grad_norm": 1.619513750076294,
|
| 1578 |
+
"learning_rate": 0.0001976839742969148,
|
| 1579 |
+
"loss": 3.1719,
|
| 1580 |
+
"step": 221
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 0.14399221663693854,
|
| 1584 |
+
"grad_norm": 1.6449589729309082,
|
| 1585 |
+
"learning_rate": 0.00019766205557100868,
|
| 1586 |
+
"loss": 3.1031,
|
| 1587 |
+
"step": 222
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 0.14464083022539323,
|
| 1591 |
+
"grad_norm": 1.5686246156692505,
|
| 1592 |
+
"learning_rate": 0.0001976400348409523,
|
| 1593 |
+
"loss": 3.1232,
|
| 1594 |
+
"step": 223
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 0.1452894438138479,
|
| 1598 |
+
"grad_norm": 1.781656265258789,
|
| 1599 |
+
"learning_rate": 0.0001976179121297455,
|
| 1600 |
+
"loss": 3.3703,
|
| 1601 |
+
"step": 224
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 0.1459380574023026,
|
| 1605 |
+
"grad_norm": 1.6864426136016846,
|
| 1606 |
+
"learning_rate": 0.00019759568746049452,
|
| 1607 |
+
"loss": 3.3688,
|
| 1608 |
+
"step": 225
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 0.14658667099075726,
|
| 1612 |
+
"grad_norm": 1.545371651649475,
|
| 1613 |
+
"learning_rate": 0.00019757336085641218,
|
| 1614 |
+
"loss": 3.3151,
|
| 1615 |
+
"step": 226
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 0.14723528457921192,
|
| 1619 |
+
"grad_norm": 1.3052035570144653,
|
| 1620 |
+
"learning_rate": 0.0001975509323408177,
|
| 1621 |
+
"loss": 2.8913,
|
| 1622 |
+
"step": 227
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 0.14788389816766662,
|
| 1626 |
+
"grad_norm": 1.2720938920974731,
|
| 1627 |
+
"learning_rate": 0.0001975284019371368,
|
| 1628 |
+
"loss": 3.0601,
|
| 1629 |
+
"step": 228
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 0.14853251175612128,
|
| 1633 |
+
"grad_norm": 1.3967403173446655,
|
| 1634 |
+
"learning_rate": 0.00019750576966890158,
|
| 1635 |
+
"loss": 3.1093,
|
| 1636 |
+
"step": 229
|
| 1637 |
+
},
|
| 1638 |
+
{
|
| 1639 |
+
"epoch": 0.14918112534457598,
|
| 1640 |
+
"grad_norm": 1.4218616485595703,
|
| 1641 |
+
"learning_rate": 0.00019748303555975057,
|
| 1642 |
+
"loss": 2.8957,
|
| 1643 |
+
"step": 230
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 0.14982973893303064,
|
| 1647 |
+
"grad_norm": 1.502661943435669,
|
| 1648 |
+
"learning_rate": 0.0001974601996334286,
|
| 1649 |
+
"loss": 3.1115,
|
| 1650 |
+
"step": 231
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 0.15047835252148534,
|
| 1654 |
+
"grad_norm": 1.4609853029251099,
|
| 1655 |
+
"learning_rate": 0.00019743726191378698,
|
| 1656 |
+
"loss": 2.9707,
|
| 1657 |
+
"step": 232
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 0.15112696610994,
|
| 1661 |
+
"grad_norm": 1.4173173904418945,
|
| 1662 |
+
"learning_rate": 0.00019741422242478316,
|
| 1663 |
+
"loss": 2.7085,
|
| 1664 |
+
"step": 233
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 0.15177557969839467,
|
| 1668 |
+
"grad_norm": 1.6199264526367188,
|
| 1669 |
+
"learning_rate": 0.000197391081190481,
|
| 1670 |
+
"loss": 3.1486,
|
| 1671 |
+
"step": 234
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 0.15242419328684936,
|
| 1675 |
+
"grad_norm": 1.5446836948394775,
|
| 1676 |
+
"learning_rate": 0.00019736783823505065,
|
| 1677 |
+
"loss": 2.884,
|
| 1678 |
+
"step": 235
|
| 1679 |
+
},
|
| 1680 |
+
{
|
| 1681 |
+
"epoch": 0.15307280687530403,
|
| 1682 |
+
"grad_norm": 1.59959876537323,
|
| 1683 |
+
"learning_rate": 0.00019734449358276842,
|
| 1684 |
+
"loss": 3.1943,
|
| 1685 |
+
"step": 236
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 0.15372142046375872,
|
| 1689 |
+
"grad_norm": 1.547925353050232,
|
| 1690 |
+
"learning_rate": 0.00019732104725801684,
|
| 1691 |
+
"loss": 3.1038,
|
| 1692 |
+
"step": 237
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 0.1543700340522134,
|
| 1696 |
+
"grad_norm": 1.5789741277694702,
|
| 1697 |
+
"learning_rate": 0.0001972974992852847,
|
| 1698 |
+
"loss": 3.0996,
|
| 1699 |
+
"step": 238
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 0.15501864764066808,
|
| 1703 |
+
"grad_norm": 1.675233244895935,
|
| 1704 |
+
"learning_rate": 0.00019727384968916693,
|
| 1705 |
+
"loss": 3.3453,
|
| 1706 |
+
"step": 239
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 0.15566726122912275,
|
| 1710 |
+
"grad_norm": 1.5575461387634277,
|
| 1711 |
+
"learning_rate": 0.00019725009849436463,
|
| 1712 |
+
"loss": 3.0918,
|
| 1713 |
+
"step": 240
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 0.15631587481757742,
|
| 1717 |
+
"grad_norm": 1.6142653226852417,
|
| 1718 |
+
"learning_rate": 0.00019722624572568492,
|
| 1719 |
+
"loss": 3.0376,
|
| 1720 |
+
"step": 241
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 0.1569644884060321,
|
| 1724 |
+
"grad_norm": 1.6989195346832275,
|
| 1725 |
+
"learning_rate": 0.0001972022914080411,
|
| 1726 |
+
"loss": 3.4667,
|
| 1727 |
+
"step": 242
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 0.15761310199448678,
|
| 1731 |
+
"grad_norm": 1.5104671716690063,
|
| 1732 |
+
"learning_rate": 0.0001971782355664525,
|
| 1733 |
+
"loss": 2.9191,
|
| 1734 |
+
"step": 243
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 0.15826171558294147,
|
| 1738 |
+
"grad_norm": 1.792704463005066,
|
| 1739 |
+
"learning_rate": 0.00019715407822604451,
|
| 1740 |
+
"loss": 3.2042,
|
| 1741 |
+
"step": 244
|
| 1742 |
+
},
|
| 1743 |
+
{
|
| 1744 |
+
"epoch": 0.15891032917139614,
|
| 1745 |
+
"grad_norm": 1.798724889755249,
|
| 1746 |
+
"learning_rate": 0.00019712981941204848,
|
| 1747 |
+
"loss": 3.1007,
|
| 1748 |
+
"step": 245
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 0.15955894275985083,
|
| 1752 |
+
"grad_norm": 1.3790630102157593,
|
| 1753 |
+
"learning_rate": 0.00019710545914980183,
|
| 1754 |
+
"loss": 3.0556,
|
| 1755 |
+
"step": 246
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 0.1602075563483055,
|
| 1759 |
+
"grad_norm": 1.8706490993499756,
|
| 1760 |
+
"learning_rate": 0.00019708099746474785,
|
| 1761 |
+
"loss": 3.3396,
|
| 1762 |
+
"step": 247
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 0.16085616993676016,
|
| 1766 |
+
"grad_norm": 1.6114487648010254,
|
| 1767 |
+
"learning_rate": 0.00019705643438243584,
|
| 1768 |
+
"loss": 2.6075,
|
| 1769 |
+
"step": 248
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 0.16150478352521486,
|
| 1773 |
+
"grad_norm": 1.5196325778961182,
|
| 1774 |
+
"learning_rate": 0.0001970317699285209,
|
| 1775 |
+
"loss": 3.0033,
|
| 1776 |
+
"step": 249
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 0.16215339711366952,
|
| 1780 |
+
"grad_norm": 1.6828662157058716,
|
| 1781 |
+
"learning_rate": 0.00019700700412876416,
|
| 1782 |
+
"loss": 3.0495,
|
| 1783 |
+
"step": 250
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 0.16280201070212422,
|
| 1787 |
+
"grad_norm": 1.5269091129302979,
|
| 1788 |
+
"learning_rate": 0.00019698213700903246,
|
| 1789 |
+
"loss": 3.0902,
|
| 1790 |
+
"step": 251
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 0.16345062429057888,
|
| 1794 |
+
"grad_norm": 1.3918476104736328,
|
| 1795 |
+
"learning_rate": 0.00019695716859529855,
|
| 1796 |
+
"loss": 2.9452,
|
| 1797 |
+
"step": 252
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 0.16409923787903358,
|
| 1801 |
+
"grad_norm": 1.6424914598464966,
|
| 1802 |
+
"learning_rate": 0.00019693209891364093,
|
| 1803 |
+
"loss": 3.2129,
|
| 1804 |
+
"step": 253
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 0.16474785146748824,
|
| 1808 |
+
"grad_norm": 2.0256524085998535,
|
| 1809 |
+
"learning_rate": 0.0001969069279902439,
|
| 1810 |
+
"loss": 3.6181,
|
| 1811 |
+
"step": 254
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 0.1653964650559429,
|
| 1815 |
+
"grad_norm": 1.7617406845092773,
|
| 1816 |
+
"learning_rate": 0.00019688165585139748,
|
| 1817 |
+
"loss": 3.2462,
|
| 1818 |
+
"step": 255
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 0.1660450786443976,
|
| 1822 |
+
"grad_norm": 1.5007916688919067,
|
| 1823 |
+
"learning_rate": 0.00019685628252349744,
|
| 1824 |
+
"loss": 3.0225,
|
| 1825 |
+
"step": 256
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 0.16669369223285227,
|
| 1829 |
+
"grad_norm": 1.6564241647720337,
|
| 1830 |
+
"learning_rate": 0.0001968308080330452,
|
| 1831 |
+
"loss": 3.0276,
|
| 1832 |
+
"step": 257
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 0.16734230582130696,
|
| 1836 |
+
"grad_norm": 1.571022868156433,
|
| 1837 |
+
"learning_rate": 0.00019680523240664786,
|
| 1838 |
+
"loss": 2.853,
|
| 1839 |
+
"step": 258
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 0.16799091940976163,
|
| 1843 |
+
"grad_norm": 1.3106783628463745,
|
| 1844 |
+
"learning_rate": 0.00019677955567101813,
|
| 1845 |
+
"loss": 2.539,
|
| 1846 |
+
"step": 259
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 0.16863953299821632,
|
| 1850 |
+
"grad_norm": 1.4485841989517212,
|
| 1851 |
+
"learning_rate": 0.0001967537778529744,
|
| 1852 |
+
"loss": 2.9644,
|
| 1853 |
+
"step": 260
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 0.169288146586671,
|
| 1857 |
+
"grad_norm": 1.4725185632705688,
|
| 1858 |
+
"learning_rate": 0.00019672789897944056,
|
| 1859 |
+
"loss": 2.8638,
|
| 1860 |
+
"step": 261
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 0.16993676017512566,
|
| 1864 |
+
"grad_norm": 1.7089518308639526,
|
| 1865 |
+
"learning_rate": 0.00019670191907744598,
|
| 1866 |
+
"loss": 3.2726,
|
| 1867 |
+
"step": 262
|
| 1868 |
+
},
|
| 1869 |
+
{
|
| 1870 |
+
"epoch": 0.17058537376358035,
|
| 1871 |
+
"grad_norm": 2.0805907249450684,
|
| 1872 |
+
"learning_rate": 0.00019667583817412578,
|
| 1873 |
+
"loss": 3.5152,
|
| 1874 |
+
"step": 263
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.17123398735203502,
|
| 1878 |
+
"grad_norm": 1.42578125,
|
| 1879 |
+
"learning_rate": 0.00019664965629672033,
|
| 1880 |
+
"loss": 2.7068,
|
| 1881 |
+
"step": 264
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 0.1718826009404897,
|
| 1885 |
+
"grad_norm": 1.70866858959198,
|
| 1886 |
+
"learning_rate": 0.0001966233734725756,
|
| 1887 |
+
"loss": 2.9573,
|
| 1888 |
+
"step": 265
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 0.17253121452894438,
|
| 1892 |
+
"grad_norm": 1.55576491355896,
|
| 1893 |
+
"learning_rate": 0.000196596989729143,
|
| 1894 |
+
"loss": 3.1079,
|
| 1895 |
+
"step": 266
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 0.17317982811739907,
|
| 1899 |
+
"grad_norm": 1.4486907720565796,
|
| 1900 |
+
"learning_rate": 0.00019657050509397923,
|
| 1901 |
+
"loss": 2.9568,
|
| 1902 |
+
"step": 267
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 0.17382844170585374,
|
| 1906 |
+
"grad_norm": 2.7671091556549072,
|
| 1907 |
+
"learning_rate": 0.00019654391959474647,
|
| 1908 |
+
"loss": 3.3971,
|
| 1909 |
+
"step": 268
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 0.1744770552943084,
|
| 1913 |
+
"grad_norm": 1.6282739639282227,
|
| 1914 |
+
"learning_rate": 0.00019651723325921224,
|
| 1915 |
+
"loss": 3.3782,
|
| 1916 |
+
"step": 269
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 0.1751256688827631,
|
| 1920 |
+
"grad_norm": 1.6156924962997437,
|
| 1921 |
+
"learning_rate": 0.00019649044611524933,
|
| 1922 |
+
"loss": 3.1748,
|
| 1923 |
+
"step": 270
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 0.17577428247121776,
|
| 1927 |
+
"grad_norm": 1.4323394298553467,
|
| 1928 |
+
"learning_rate": 0.00019646355819083589,
|
| 1929 |
+
"loss": 2.9028,
|
| 1930 |
+
"step": 271
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 0.17642289605967246,
|
| 1934 |
+
"grad_norm": 1.5577207803726196,
|
| 1935 |
+
"learning_rate": 0.00019643656951405525,
|
| 1936 |
+
"loss": 3.1325,
|
| 1937 |
+
"step": 272
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 0.17707150964812712,
|
| 1941 |
+
"grad_norm": 1.5941506624221802,
|
| 1942 |
+
"learning_rate": 0.00019640948011309604,
|
| 1943 |
+
"loss": 3.0872,
|
| 1944 |
+
"step": 273
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 0.17772012323658182,
|
| 1948 |
+
"grad_norm": 1.3285024166107178,
|
| 1949 |
+
"learning_rate": 0.00019638229001625205,
|
| 1950 |
+
"loss": 2.6622,
|
| 1951 |
+
"step": 274
|
| 1952 |
+
},
|
| 1953 |
+
{
|
| 1954 |
+
"epoch": 0.17836873682503648,
|
| 1955 |
+
"grad_norm": 1.4673588275909424,
|
| 1956 |
+
"learning_rate": 0.0001963549992519223,
|
| 1957 |
+
"loss": 3.1656,
|
| 1958 |
+
"step": 275
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 0.17901735041349115,
|
| 1962 |
+
"grad_norm": 1.9262672662734985,
|
| 1963 |
+
"learning_rate": 0.00019632760784861087,
|
| 1964 |
+
"loss": 2.8955,
|
| 1965 |
+
"step": 276
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 0.17966596400194584,
|
| 1969 |
+
"grad_norm": 1.5531195402145386,
|
| 1970 |
+
"learning_rate": 0.00019630011583492702,
|
| 1971 |
+
"loss": 2.9895,
|
| 1972 |
+
"step": 277
|
| 1973 |
+
},
|
| 1974 |
+
{
|
| 1975 |
+
"epoch": 0.1803145775904005,
|
| 1976 |
+
"grad_norm": 1.259700059890747,
|
| 1977 |
+
"learning_rate": 0.00019627252323958504,
|
| 1978 |
+
"loss": 2.4601,
|
| 1979 |
+
"step": 278
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 0.1809631911788552,
|
| 1983 |
+
"grad_norm": 1.4137221574783325,
|
| 1984 |
+
"learning_rate": 0.00019624483009140435,
|
| 1985 |
+
"loss": 2.7425,
|
| 1986 |
+
"step": 279
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 0.18161180476730987,
|
| 1990 |
+
"grad_norm": 1.2611486911773682,
|
| 1991 |
+
"learning_rate": 0.0001962170364193093,
|
| 1992 |
+
"loss": 2.7371,
|
| 1993 |
+
"step": 280
|
| 1994 |
+
},
|
| 1995 |
+
{
|
| 1996 |
+
"epoch": 0.18226041835576456,
|
| 1997 |
+
"grad_norm": 1.7357769012451172,
|
| 1998 |
+
"learning_rate": 0.00019618914225232934,
|
| 1999 |
+
"loss": 3.0614,
|
| 2000 |
+
"step": 281
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 0.18290903194421923,
|
| 2004 |
+
"grad_norm": 1.706430435180664,
|
| 2005 |
+
"learning_rate": 0.00019616114761959874,
|
| 2006 |
+
"loss": 2.9595,
|
| 2007 |
+
"step": 282
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 0.1835576455326739,
|
| 2011 |
+
"grad_norm": 1.6613085269927979,
|
| 2012 |
+
"learning_rate": 0.00019613305255035686,
|
| 2013 |
+
"loss": 2.9299,
|
| 2014 |
+
"step": 283
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 0.1842062591211286,
|
| 2018 |
+
"grad_norm": 1.4859459400177002,
|
| 2019 |
+
"learning_rate": 0.00019610485707394784,
|
| 2020 |
+
"loss": 2.8886,
|
| 2021 |
+
"step": 284
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 0.18485487270958326,
|
| 2025 |
+
"grad_norm": 1.5165742635726929,
|
| 2026 |
+
"learning_rate": 0.00019607656121982075,
|
| 2027 |
+
"loss": 2.7321,
|
| 2028 |
+
"step": 285
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 0.18550348629803795,
|
| 2032 |
+
"grad_norm": 1.711745023727417,
|
| 2033 |
+
"learning_rate": 0.00019604816501752947,
|
| 2034 |
+
"loss": 2.8954,
|
| 2035 |
+
"step": 286
|
| 2036 |
+
},
|
| 2037 |
+
{
|
| 2038 |
+
"epoch": 0.18615209988649262,
|
| 2039 |
+
"grad_norm": 1.5451345443725586,
|
| 2040 |
+
"learning_rate": 0.00019601966849673276,
|
| 2041 |
+
"loss": 2.9563,
|
| 2042 |
+
"step": 287
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 0.1868007134749473,
|
| 2046 |
+
"grad_norm": 1.3915531635284424,
|
| 2047 |
+
"learning_rate": 0.00019599107168719412,
|
| 2048 |
+
"loss": 3.0462,
|
| 2049 |
+
"step": 288
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 0.18744932706340198,
|
| 2053 |
+
"grad_norm": 1.2463390827178955,
|
| 2054 |
+
"learning_rate": 0.0001959623746187817,
|
| 2055 |
+
"loss": 2.7958,
|
| 2056 |
+
"step": 289
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 0.18809794065185664,
|
| 2060 |
+
"grad_norm": 1.683665156364441,
|
| 2061 |
+
"learning_rate": 0.0001959335773214685,
|
| 2062 |
+
"loss": 2.8798,
|
| 2063 |
+
"step": 290
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 0.18874655424031134,
|
| 2067 |
+
"grad_norm": 1.5754495859146118,
|
| 2068 |
+
"learning_rate": 0.00019590467982533215,
|
| 2069 |
+
"loss": 2.8401,
|
| 2070 |
+
"step": 291
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 0.189395167828766,
|
| 2074 |
+
"grad_norm": 1.6469128131866455,
|
| 2075 |
+
"learning_rate": 0.000195875682160555,
|
| 2076 |
+
"loss": 3.0458,
|
| 2077 |
+
"step": 292
|
| 2078 |
+
},
|
| 2079 |
+
{
|
| 2080 |
+
"epoch": 0.1900437814172207,
|
| 2081 |
+
"grad_norm": 1.4030909538269043,
|
| 2082 |
+
"learning_rate": 0.00019584658435742384,
|
| 2083 |
+
"loss": 2.8845,
|
| 2084 |
+
"step": 293
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 0.19069239500567536,
|
| 2088 |
+
"grad_norm": 1.4269638061523438,
|
| 2089 |
+
"learning_rate": 0.00019581738644633027,
|
| 2090 |
+
"loss": 3.1745,
|
| 2091 |
+
"step": 294
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 0.19134100859413006,
|
| 2095 |
+
"grad_norm": 1.8712362051010132,
|
| 2096 |
+
"learning_rate": 0.00019578808845777034,
|
| 2097 |
+
"loss": 2.9603,
|
| 2098 |
+
"step": 295
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 0.19198962218258472,
|
| 2102 |
+
"grad_norm": 1.3823246955871582,
|
| 2103 |
+
"learning_rate": 0.0001957586904223446,
|
| 2104 |
+
"loss": 2.8134,
|
| 2105 |
+
"step": 296
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 0.1926382357710394,
|
| 2109 |
+
"grad_norm": 1.6640461683273315,
|
| 2110 |
+
"learning_rate": 0.00019572919237075817,
|
| 2111 |
+
"loss": 3.0434,
|
| 2112 |
+
"step": 297
|
| 2113 |
+
},
|
| 2114 |
+
{
|
| 2115 |
+
"epoch": 0.19328684935949408,
|
| 2116 |
+
"grad_norm": 1.68658447265625,
|
| 2117 |
+
"learning_rate": 0.0001956995943338206,
|
| 2118 |
+
"loss": 3.2863,
|
| 2119 |
+
"step": 298
|
| 2120 |
+
},
|
| 2121 |
+
{
|
| 2122 |
+
"epoch": 0.19393546294794875,
|
| 2123 |
+
"grad_norm": 1.4889813661575317,
|
| 2124 |
+
"learning_rate": 0.00019566989634244584,
|
| 2125 |
+
"loss": 2.7607,
|
| 2126 |
+
"step": 299
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 0.19458407653640344,
|
| 2130 |
+
"grad_norm": 1.526188850402832,
|
| 2131 |
+
"learning_rate": 0.00019564009842765225,
|
| 2132 |
+
"loss": 2.9834,
|
| 2133 |
+
"step": 300
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.19458407653640344,
|
| 2137 |
+
"eval_loss": 3.1460793018341064,
|
| 2138 |
+
"eval_runtime": 35.0284,
|
| 2139 |
+
"eval_samples_per_second": 58.695,
|
| 2140 |
+
"eval_steps_per_second": 14.674,
|
| 2141 |
+
"step": 300
|
| 2142 |
}
|
| 2143 |
],
|
| 2144 |
"logging_steps": 1,
|
|
|
|
| 2167 |
"attributes": {}
|
| 2168 |
}
|
| 2169 |
},
|
| 2170 |
+
"total_flos": 2.1628629024768e+16,
|
| 2171 |
"train_batch_size": 4,
|
| 2172 |
"trial_name": null,
|
| 2173 |
"trial_params": null
|