Android: Cannot load models, stopCompletions not working.
Vali-98 opened this issue · comments
As it says on the tin. Loading small 3b models ala Tiny Llama or StableLM models do not work. Tested models:
- https://huggingface.co/TheBloke/stablelm-zephyr-3b-GGUF
- https://huggingface.co/TheBloke/TinyLlama-1.1B-Chat-v0.3-GGUF
- https://huggingface.co/TheBloke/phi-2-GGUF/tree/main
Attempting to call initLlama results in
Error: Failed to initialize context
Which I can only assume is here:
I do not know enough about native functions to investigate further.
In addition, stopCompletions() does not stop a completion on Android.
Thanks for your work, the project is fantastic otherwise.
Can you share the following info? Thanks!
- llama.rn version (latest: 0.3.0-rc.9)
- Used model quantized type or file size
- Android device version & arch & total memory
Also you can use adb logcat -s "RNLLAMA_LOG_ANDROID","RNLLAMA_ANDROID_JNI"
to get logs.
Sure thing:
-
llama.rn version: 0.3.0-rc.9
-
Models used & quantizations:
-
Not working:
-- phi-2.Q3_K_M.gguf
-- stablelm-zephyr-3b.Q3_K_L.gguf -
Working:
-- tinyllama-1.1b-chat-v0.3.Q3_K_L.gguf -
Android Devices Tested on:
-- Samsung A71 - Android 13 - 8GB
-- Emulated Pixel 3a - Android 14 - 8GB allocated -
logcat log:
RNLLAMA_ANDROID_JNI: [RNLlama] is_model_loaded false
RNLLAMA_ANDROID_JNI: [RNLlama] is_model_loaded false
RNLLAMA_ANDROID_JNI: [RNLlama] is_model_loaded false
After further testing, this seems to be caused by an older version of the package being installed instead. Purging cache and downloading latest fixed the issue.
A bit of an embarassing mistake by me, but I will leave this up for future reference. Thank you for your time.