Spaces:
Running
Small model for parsing city directories?
Hello!
Love the work you're doing here.
I've recently been working on a personal project I call Directory Pipeline and for the next phase of work I want to do on it, I'm eager to make it more specifically useful for parsing the thousands of city directories published in the US in nearly every city in the country in the 19th and 20th centuries. You can see one example product of what the pipeline produce for city directories here: https://hadro.github.io/tulsa-city-directories/1921#about
Based on things both of you have written in recent months, I feel fairly confident I'm up for the task of training a fine-tuned small model specific to city directory data, and I think it could potentially be useful to a wide variety of GLAM institutions.
But before I dive headfirst into the work with no prior experience, I figured I would ask you here if you have any specific pointers or words of advice you'd give
to someone embarking on this for the first time?
Likewise, if in your small models for GLAM travels you've encountered anything near enough to this kind of thing that I should put my energy elsewehre, please let me know, I certainly don't want to reinvent any wheels!