levmckinney

Levmckinney's group workspace

Group: pythia-2.8b-deduped

1-2

of 2

Timestamps visible

2023-05-05 00:54:09

Using 'GPTNeoXLayer' for transformer_auto_wrap_policy.

2023-05-05 00:54:21

Gradient accumulation steps: 64

2023-05-05 00:54:21

Using 262_144 tokens per training step.

2023-05-05 00:54:21

All processes have completed setup. Starting training.

2023-05-05 04:44:02

Training: 100%|██████████| 16000/16000 [3:49:41<00:00, 1.16it/s]

2023-05-05 04:44:02

Saving lens to /output/EleutherAI/pythia-2.8b-deduped-1683247839