There is another version at:

https://github.com/urchade/GLiNER

Looks like it’s still being maintained too?

GLiNER2 is an upgrade that allows for relationship extraction and classification, built on GLiNER, and added additional research / papers, then trained a new model.

Use Gliner2. Much better model.

Okay but there is a dependency on gliner1:

https://github.com/fastino-ai/GLiNER2/issues/69

As heavily discussed here 3 days ago (Persona is the same company LinkedIn uses for their ID verification process):

I verified my LinkedIn identity. Here's what I handed over

https://news.ycombinator.com/item?id=47098245

1.4K+ points, 490+ comments

> ERROR: Unknown input format pdf Pandoc can convert to PDF, but not from PDF

Darn

Thanks for sharing this! I had done a PWA that displays some revenue forecasting[1] based on Stripe Subscriptions but I found their APIs can be slow[2]. Having that data synced up in a db sounds like a good enhancement so will definitely look into this more.

[1] https://github.com/hbcondo/revenut-app

[2] https://github.com/stripe/stripe-dotnet/issues/2284#issuecom...

Yes, this is one of the bigger use-cases for this library. Lots of folks want custom analytics on their Stripe data and this provides a nice way to do just that.

Let me know how you find it! Happy to implement any fixes and PRs welcome!

Thanks for the strategy transparency; it's nice to see how the system and user prompts are written, especially the multi-step agents.

in no particular order:

  astronomy
  cryptocurrency
  climate change
  behavioral health
  prediction markets

Nice! Mermaid diagrams show up as their source code in Github Pages so perhaps this can be resolved now

https://github.com/orgs/community/discussions/65040

What was the level of (development) effort switching to subscription?

Not too bad technically – Apple handles most of it with StoreKit. Maybe a week of work for the initial implementation: setting up products in App Store Connect, handling purchase flow, receipt validation, restoring purchases.

The harder part was deciding on the model. Freemium limits? Trial length? What to lock? I changed this multiple times. The code changes for each iteration were small, but the decisions took longer.

Also submitted here 2 days ago (but no discussion yet): https://news.ycombinator.com/item?id=46475993