woheller69
woheller69
has anyone found a reasonable solution for this? Or am I the only one not willing to wait until the model finishes without killing the job and losing context?
gpt4all python bindings offer a similar way which allows stopping with the next token
I don't plan to add it to bookmarks. But I might add it to the settings per TAB in future
see #6 for icons I would have to allow all URLs, including tracking...
For those who do not care about tracking... Here is an unblocked version of gptAssist: [gptAssist_unblocked.apk.zip](https://github.com/woheller69/gptAssist/files/12258010/gptAssist_unblocked.apk.zip) -Uninstall -Remove .zip and install
this would be a different app. I also have one for huggingChat, where you can select different open source models. https://f-droid.org/en/packages/org.woheller69.hugassist/
that does not work for me. In my tkinter app mainloop() must at the end of the with statement. So I would have to restart the app which - with...
Excellent, thanks! That works. https://github.com/woheller69/gpt4all-TK-CHAT/blob/main/appGUI.py
This does not happen with smaller models, such as Llama 3 8B Instruct Q8 which is 8.5GB in size. Dolphin 2.7 Mixtral 8x7b Q4_K_M is 26 GB. I have 36...
With v2.6.0 Dolphin 2.7 is held in cache and reloads quickly:  I notice the same with llama-cpp-python. Has there been a degradation in llama.cpp ?