Skip to main content
haok
Projects
flame-moe
Run groups
dense-191m
Logs
Group
Workspace
Runs
Logs
Files
Artifacts

Annavettoruzzo's group workspace

Timestamps visible
2025-05-14 22:43:25
 [2025-05-14 18:43:25] iteration     2500/    2503 | consumed samples:      2560000 | elapsed time per iteration (ms): 553.4 | throughput per GPU (TFLOP/s/GPU): 242.7 | learning rate: 3.225516E-05 | global batch size:  1024 | lm loss: 3.493463E+00 | loss scale: 1.0 | grad norm: 0.150 | num zeros: 0 | number of skipped iterations:   0 | number of nan iterations:   0 |
2025-05-14 22:43:56
(min, max) time across ranks (ms):
2025-05-14 22:43:56
    evaluate .......................................: (30962.97, 30964.47)
2025-05-14 22:43:56
------------------------------------------------------------------------------------------------
2025-05-14 22:43:56
 validation loss at iteration 2500 | lm loss value: 3.498568E+00 | lm loss PPL: 3.306805E+01 |
2025-05-14 22:43:56
------------------------------------------------------------------------------------------------
2025-05-14 22:43:58
(min, max) time across ranks (ms):
2025-05-14 22:43:58
    save-checkpoint ................................: (1423.93, 1426.21)
2025-05-14 22:44:31
(min, max) time across ranks (ms):
2025-05-14 22:44:31
    evaluate .......................................: (29897.12, 29898.28)
2025-05-14 22:44:31
------------------------------------------------------------------------------------------------------------------
2025-05-14 22:44:31
 validation loss at iteration 2503 on validation set | lm loss value: 3.499209E+00 | lm loss PPL: 3.308927E+01 |
2025-05-14 22:44:31
------------------------------------------------------------------------------------------------------------------
2025-05-14 22:45:18
(min, max) time across ranks (ms):
2025-05-14 22:45:18
    evaluate .......................................: (47441.41, 47441.99)
2025-05-14 22:45:18
------------------------------------------------------------------------------------------------------------
2025-05-14 22:45:18
 validation loss at iteration 2503 on test set | lm loss value: 3.496324E+00 | lm loss PPL: 3.299395E+01 |
2025-05-14 22:45:18
------------------------------------------------------------------------------------------------------------