I have lot of documents specific to a domain and I want to pretrain the model with the domain knowledge. How to achieve this? Thanks in advance

If I was Marvin the Paranoid Android from Douglas Adams 'The Hitch-hikers guide to the galaxy" My morning would have gone something like "Funny, how just when you think life can't possibly get any worse it suddenly does" i have spent the past hour trying to work out why my new crew isn't working against my ollama llama 3 when it was working earlier. Typically i jump straight in and look at all the possible issues, Libraries, bad code, numerous reads of woefully inadequate blog posts, and then i had an epiphany and checked the first thing i should have checked. .... YES, THE SERVICE WASN'T RUNNING!!! I had stopped ollama earlier to delete a model and didn't restart it. Moral of the story. sometimes its the little things that get you, not the code.

I'm very curious what everyone is using when they are writing code for CrewAI. Def leave a comment to give context if you want. If you can't pick multiple options leave the other ones you use in the comments.

Hey guys, do you have any tips on how to run projects locally ? I kind of struggle with this now... I am using a laptop with 16 GB (15,3) of RAM and 4 core Intel i5 (2,60 GHz) processor running on Ubuntu. No Nvidia graphics card unfortunately... I am trying to use only small models such as phi, phi3 or llama2 with Ollama, but most of the times it simply gets stuck while running the code or returns weird characters instead of the agents' work results. I don't have any errors in the code, the program runs well in the beginning, but it freezes usually after 4-5 minutes...

Capable of up to 1M token, but only with >100GB GPU memory. See https://ollama.com/library/llama3-gradient