Is there a character level Language model pre-trained on Wiki-Text?

I am aware that there is a pre-trained word level ULMIT language model available in fastai library.
Is there a corresponding character level language model available which is pre-trained on wiki text??

I am currently working on spell correction using neural machine translation and I am looking for a character level pretrained language model which I can use as an encoder for this.

Have you found any source for it? Is there any source i can use to create my own character-level langugage model. I have bee following this but it seems to keep running out of memory.