| GPT Dataset | |
| Data prefix /data/arxiv/tokenized_text_document | |
| Dataset name valid | |
| Number of samples 51404800 | |
| Number of epochs 107 | |
| Sequence length 2048 | |
| Random seed 1234 | |
| Split 969, 30, 1 | |
| GPT Dataset | |
| Data prefix /data/arxiv/tokenized_text_document | |
| Dataset name valid | |
| Number of samples 51404800 | |
| Number of epochs 107 | |
| Sequence length 2048 | |
| Random seed 1234 | |
| Split 969, 30, 1 | |