mega-base-wikitext / README.md
mnaylor's picture
Update README.md
3a4ca8d
|
raw
history blame
842 Bytes
metadata
license: apache-2.0
language:
  - en
library_name: transformers

Mega Masked LM on wikitext-103

This is the location on the Hugging Face hub for the Mega MLM checkpoint. I trained this model on the wikitext-103 dataset using standard BERT-style masked LM pretraining using the original Mega repository and uploaded the weights initially to hf.co/mnaylor/mega-wikitext-103. When the implementation of Mega into Hugging Face's transformers is finished, the weights here are designed to be used with MegaForMaskedLM and are compatible with the other (encoder-based) MegaFor* model classes.

This model uses the RoBERTa base tokenizer since the Mega paper does not implement a specific tokenizer aside from the character-level tokenizer used to illustrate long-sequence performance.