Hardware specs for offline version? | Socrates

Socrates

Q: Hardware specs for offline version?

Apologies, if this question sounds naive, as I'm not with my laptop this week. How do you know if it will run on a i5 windows PC and what are the minimum specs; for example does it need to have a GPU?

WhatNuwithyouPLUSNov 10, 2024

Founder Team

Jon_Socrates

Nov 10, 2024

A: I'd recommend minimum 16 GB of ram to run local models. No, doesn't need GPU!

Helpful?

Related questions

We're excited about Socrates and about to pull the trigger on the top tier, but have a few questions: 1. Will any of the AppSumo tiers include Flow AI when it's released? 2. Is API access planned? We'd love to be able to add sources and chat via an API. 3. Are there any functionalities we lose if we use only the Web UI instead of the app? We don't care about running local models. 4. Are the hot-linked citations only available in normal chat and not deep dive? We tested on a free account and could not get deep-dive to return the hot-linked [#] citation numbers--they only displayed when deep-dive was turned off. 5. Do you have a Discord or community we can join/check out? Thanks. We're excited about this product and look forward to your reply.
What does 250 prompt loop segments mean? What does 5 cloud chats mean? I will only be able to analyze 5 documents in total, if I select cloud mode? What about the limit for local mode?
So the deal is for one user, with the only community components being the ability to share 1 chat (Tier 3) or 5 chats (Tier 4)? What is the mechanism of sharing a chat? Is it via embed? Is it by giving access to specific people who log in somewhere? Could you point me to an example of a shared chat that I can use? Thanks!
1. No stop or edit button? - In testing the desktop version the LLM got carried away and you can't stop it and you can't edit the query. Odd omissions. 2. When deep dive comes to the local version of Socrates is there parity between the local and online limits? (How do you know which tier is best if you mostly use the Local LLM?). 3. If working locally, can you access the live chat files and make a backup copy (.db-wal file)? One of the nice things about using LM studio is you can delete and restore the chat files (.json). Literally cut and paste (including the uploaded files) from the file manager. Exporting a chat each and every time would get tiresome as a independent backup process. 4. Linking to a folder locally or a cloud drive with a refresh option would be great for obsidian users. Saves reuploading edited md files.
Hi. I purchased tier 1. I uploaded a document (with pictures) and I activated the Deep Dive feature, but it got stuck at 5% for the last 10 hours. Any recommendations on how to unstick the process? See screenshot: https://share.zight.com/nOu48lgD

View product details

Socrates

Share Socrates

Related questions