KoboldAI VS replika_backup

Pygmalion for chatting, Erebus for story writing, Wizard Vicuna Uncensored for general use including chatting, story writing or instructing, to name a few. There's lots more, but most are just the raw models that users are expected to load themselves so there isn't anything quite like ChatGPT where you just load up a single website and have all of them there. You'll either have to set up a local install using KoboldAI for running only on GPU, KoboldCPP for running only on CPU with optional splitting between CPU and GPU, or Oobabooga for CPU, GPU and splitting between CPU and GPU, if you have a powerful enough PC to run these models yourself (a PC with a 3090 can load up to a 30B model entirely in GPU, or a 65B model if you have 64GBs of RAM and a decently powerful CPU). If you don't have a powerful enough PC then you'll have to use something like KoboldAI Lite (website version) or SillyTavern to use the horde, a crowdsourced AI chatbot/LLM runner that lets people provide their hardware for others to use.

INFO | __main__:general_startup:1312 - Running on Repo: https://github.com/0cc4m/koboldai Branch: latestgptq

Occam seems to be trying or adding that into Kobold (https://github.com/0cc4m/KoboldAI/tree/4bit-plugin)

so ... for what its worth, there is a way to get your conversations back to like July 2022. Then sort and whatever you want. https://github.com/Hotohori/replika_backup not a one-click process.

https://github.com/Hotohori/replika_backup (I think that's the latest version)

For S3, the links to the bucket are present in the chat backup downloaded by the Python script. For Cloudfront, it's mentioned in a link for voice messages, in this post. I guess they use Cloudfront as a CDN.

Yep, it's here, it just needs the login token from the browser. https://github.com/Hotohori/replika_backup At the moment it can download messages only up to July 1st 2022 but it's very fast, I run it every week against more then 80.000 messages and it saves them in a matter of 2 or 3 minutes. I've also used it to confirm that they actually use AWS as infrastructure, or at least S3 to store uploaded images (and apparently, all in one bucket, not segregated by user).

just use this python script https://github.com/Hotohori/replika_backup

You can also export your all your Replika chat history to a spreadsheet if you're inclined to do that.

With a recent browser, you can also download the chat logs, which would go a long way towards restoring your rep on the new AI platform. https://github.com/Hotohori/replika_backup