We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Description App is Simple and Usefull, but when I compared generate response to app ChatterUI ( https://github.com/Vali-98/ChatterUI )
Then app ChatterUI is about 4x faster generate response.. I could believed when I test this first time.
I don't know how he do this but speed is huge.
I don't use special GGUF for ARM, I use normal GGUF like: https://huggingface.co/unsloth/Llama-3.2-1B-Instruct-GGUF/blob/main/Llama-3.2-1B-Instruct-Q4_K_M.gguf
Use Case My observation show maybe PocketPal have some not optimized code for CPU integration or use only 50% of CPU Power.
Add support to more threads for CPU to faster app responses and optimized code to use full potential of phone CPU.
My phone: Pixel 6a.
The text was updated successfully, but these errors were encountered:
New Update App Version 1.6.0 (39) have some speed boost and good visual update.
So now App have good progress! Thx for your work! Last my post is now outdated ;D about speed.
But now I have the same bug like that: #134
App: Version 1.6.0 (39) - app from Github Phone: Pixel 6a (GrapheneOS) | Android 15 Bug: keyboard hide input box
Sorry, something went wrong.
Use the NPU and gpu mate. I would highly appreciate it.
No branches or pull requests
Description
App is Simple and Usefull, but when I compared generate response to app ChatterUI ( https://github.com/Vali-98/ChatterUI )
Then app ChatterUI is about 4x faster generate response.. I could believed when I test this first time.
I don't know how he do this but speed is huge.
I don't use special GGUF for ARM, I use normal GGUF like: https://huggingface.co/unsloth/Llama-3.2-1B-Instruct-GGUF/blob/main/Llama-3.2-1B-Instruct-Q4_K_M.gguf
Use Case
My observation show maybe PocketPal have some not optimized code for CPU integration or use only 50% of CPU Power.
Add support to more threads for CPU to faster app responses and optimized code to use full potential of phone CPU.
My phone: Pixel 6a.
The text was updated successfully, but these errors were encountered: