levmckinney

Levmckinney's group workspace

Group: pythia-2.8b-deduped-v0

1-2

of 2

Timestamps visible

2023-05-05 20:01:49

Using 'GPTNeoXLayer' for transformer_auto_wrap_policy.

2023-05-05 20:01:59

Gradient accumulation steps: 64

2023-05-05 20:01:59

Using 262_144 tokens per training step.

2023-05-05 20:01:59

All processes have completed setup. Starting training.

2023-05-05 23:55:14

Training: 100%|██████████| 16000/16000 [3:53:13<00:00, 1.14it/s]

2023-05-05 23:55:13

Saving lens to /output/EleutherAI/pythia-2.8b-deduped-v0-1683316697