Add the directories for each OS Signed-off-by: AT <manyoso@users.noreply.github.com>
2.1 KiB
QA Checklist
-
Ensure you have a fresh install by backing up and then deleting the following directories:
Windows
- Settings directory:
C:\Users\{username}\AppData\Roaming\nomic.ai
- Models directory:
C:\Users\{username}\AppData\Local\nomic.ai\GPT4All
Mac
- Settings directory:
/Users/{username}/.config/gpt4all.io
- Models directory:
/Users/{username}/Library/Application Support/nomic.ai/GPT4All
Linux
- Settings directory:
/home/{username}/.config/nomic.ai
- Models directory:
/home/{username}/.local/share/nomic.ai/GPT4All
^ Note: If you've changed your models directory manually via the settings you need to backup and delete that one
- Settings directory:
-
Go through every view and ensure that things display correctly and familiarize yourself with the application flow
-
Navigate to the models view and download Llama 3 Instruct
-
Navigate to the models view and search for "TheBloke mistral 7b" and download "TheBloke/Mistral-7B-Instruct-v0.1-GGUF"
-
Navigate to the chat view and open new chats and load these models
-
Chat with the models and exercise them. Rename the chats. Delete chats. Open new chats. Switch models when in chats.
-
Create a new localdocs collection from a directory of .txt or .pdf files on your hard drive
-
Enable the new collection in chats (especially with Llama 3 Instruct) and exercise the localdocs feature
-
Go to the settings view and explore each setting
-
Remove collections in localdocs and re-add them. Rebuild collections
-
Now shut down the app, go back and restore any previous settings directory or model directory you had from a previous install and re-test #1 through #11 :)
-
Try to break the app
EXTRA CREDIT
-
If you have a openai api key install GPT-4 model and chat with it
-
If you have a nomic api key install the remote nomic embedding model for localdocs (see if you can discover how to do this)
-
If you have a python script that targets openai API then enable server mode and try this
-
Really try and break the app
All feedback is welcome