FYI: Databricks has released an open source ChatGPT-like clone called Dolly.

  • Thread starter Deleted member 13126
  • Start date
Status
Not open for further replies.
D

Deleted member 13126

Guest
This is the link to the open source code in case anyone wants to tinker with it. I'm going to try and see if there's any way I can get it plugged in to my Logos library (although folks who have plain text works from the Puritans should have comparatively little trouble). The company says that with one server, $30 and a few hours, you can basically have something very similar to ChatGPT running for your own limited use case.
 
I'll be insterested to see what you discover. My guess is that Logos files are encrypted in such a way that you won't be able to ingetst them into something that Dolly will be able to apply the algorithm to.

It would be something worth starting a thread with Logos to suggest they add such a feature.
 
I'll be insterested to see what you discover. My guess is that Logos files are encrypted in such a way that you won't be able to ingetst them into something that Dolly will be able to apply the algorithm to.

It would be something worth starting a thread with Logos to suggest they add such a feature.

Thanks for the encouragement.

Chat-GPT4, unlike Chat-GPT3, will do image recognition. I'm not sure that Dolly will. If Dolly could do OCR, though, I'm comfortable using Google's Tesseract-OCR and screen scraping the content that way. Unfortunately, it's been an uphill battle for more than a decade with little to no progress made with FaithLife in the way of giving users access to the raw text outside of the confines of their platform.

There may be an API I can use, though. Or there may be another API I can leverage with Archive.org or some other website that has a bunch of Puritan stuff on it. I'll keep this thread updated if I find something neat, or make a new one if it makes sense.
 
Thanks for the encouragement.

Chat-GPT4, unlike Chat-GPT3, will do image recognition. I'm not sure that Dolly will. If Dolly could do OCR, though, I'm comfortable using Google's Tesseract-OCR and screen scraping the content that way. Unfortunately, it's been an uphill battle for more than a decade with little to no progress made with FaithLife in the way of giving users access to the raw text outside of the confines of their platform.

There may be an API I can use, though. Or there may be another API I can leverage with Archive.org or some other website that has a bunch of Puritan stuff on it. I'll keep this thread updated if I find something neat, or make a new one if it makes sense.
Are you a data scientist? I'm just asking because I know enough to talk with them about how the dicipline works.

What would be interesting is to tag all the various works so that we knew that works by Owen were congregational, Puritan, paedoaptist, Reformed, etc, all date tagged and other works were likewise tagged. It would be really interesting for an algorithm to be able to gain insights across authors, movements, etc. Of course, it would be intersting if some of the other data science and engieering tools were used so that it could "show its work" rather than just giving a balck box answer.
 
Are you a data scientist? I'm just asking because I know enough to talk with them about how the dicipline works.

What would be interesting is to tag all the various works so that we knew that works by Owen were congregational, Puritan, paedoaptist, Reformed, etc, all date tagged and other works were likewise tagged. It would be really interesting for an algorithm to be able to gain insights across authors, movements, etc. Of course, it would be intersting if some of the other data science and engieering tools were used so that it could "show its work" rather than just giving a balck box answer.
Regrettably, no. I work in a niche within Information Security called IAM, interestingly enough. But I do like to tinker, and may be able to get something going.
 
Status
Not open for further replies.
Back
Top