| train: openwebtext | |
| valid: wikitext103 | |
| tokenizer_name_or_path: ibm-research/materials.selfies-ted | |
| cache_dir: /share/kuleshov/ssahoo/textdiffusion/data | |
| wrap: True | |
| streaming: False | |
| train: openwebtext | |
| valid: wikitext103 | |
| tokenizer_name_or_path: ibm-research/materials.selfies-ted | |
| cache_dir: /share/kuleshov/ssahoo/textdiffusion/data | |
| wrap: True | |
| streaming: False | |