Hacker News new | past | comments | ask | show | jobs | submit
loading story #42341865
loading story #42342860
loading story #42342047
loading story #42344515
Please help me understand something.

I've been out of the loop with HuggingFace models.

What can you do with these models?

1. Can you download them and run them on your Laptop via JupyterLab?

2. What benefits does that get you?

3. Can you update them regularly (with new data on the internet, e.g.)?

4. Can you finetune them for a specific use case (e.g. GeoSpatial data)?

5. How difficult and time-consuming (person-hours) is it to finetune a model?

(If HuggingFace has answers to these questions, please point me to the URL. HuggingFace, to me, seems like the early days of GitHub. A small number were heavy users, but the rest were left scratching their heads and wondering how to use it.)

Granted it's a newbie question, but answers will be beneficial to a lot of us out there.

loading story #42342569
loading story #42342540
loading story #42343625
loading story #42343181
loading story #42345145
loading story #42343540
loading story #42342775
loading story #42343492
loading story #42345895
loading story #42342498
loading story #42343895
loading story #42342010
loading story #42343698
loading story #42344058
loading story #42345616
loading story #42342580
loading story #42342253
loading story #42347631
loading story #42346995
loading story #42346694
loading story #42345670
loading story #42342301
loading story #42346331
loading story #42341995