playground ↗
documentation ↗
NV
Infrastructure
Documentation
Support
EL
Training
Epoch 14 / 20
ETA: 4h 23m
24h
7d
30d
Export
Model 120B Fine-tune
Training Loss
0.0234
-42%
Per epoch
Validation Loss
0.0312
+2%
Per epoch
Learning Rate
2.4e-5
GPU Memory
76.2 GB
/ 80 GB
Throughput
1,842
tokens/s
Checkpoint Evaluations
Checkpoint
Epoch
Train Loss
Val Loss Trend
Status
ckpt-014
14
0.0234
0.031
Current
ckpt-010
10
0.0298
0.029
Best
ckpt-005
5
0.0512
0.048
Saved