KO
|
EN
gitlite — search
Search
#javascript
#python
#hacktoberfest
#react
#ai
#typescript
#llm
#go
#golang
#android
#machine-learning
#rust
#deep-learning
#linux
LocalGPT-Android
★ 74
Open GitHub ↗
Run GPT4ALL locally on your device
Download README (.md)
Explore Similar Repositories
Local-AI
:
A curated guide to running Large Language Models (LLMs) on your own machine. Covers tools like Ollama, LM Studio, LocalAI, GPT4All, and more!
llm-agent
:
LLM using long-term memory through vector database
Eunomia
:
Analyze your code locally using a GPT LLM.
E2E-AI-Chatbot
:
Pipeline LLMs offline combine with own context for generating answer
SimpleMamba
:
Implementation of a modular, high-performance, and simplistic mamba for high-speed applications
// repository documentation
Was this content helpful?
★ 0
(0 ratings)
Select Rating:
★
★
★
★
★
Submit Feedback
Recent Feedback
×
Download README
Do you want to download the
README.md
file for
LocalGPT-Android
?
Download (.md)
# LocalGPT-Android - The Local GPT Android is a mobile application that runs the GPT (Generative Pre-trained Transformer) model directly on your Android device. This app does not require an active internet connection, as it executes the GPT model locally. - (It still uses Internet to download the model, you can manually place the model in data directory and disable internet). ## Requirements - Android 11+ - At least 3-4 GB of Free RAM (Again "Free" RAM) - I tested it on iQoo 11 with 16GB RAM and Snapdragon 8 Gen 2 ## Features - It works :) - Only Text Completions as of now - If you want to to chat, send messages like `### Human:\n prompt\n ### Assistant:` - The model used does not care about ethics afaik ## How to? - Download the latest release - Wait for the download to complete - The model is ggml-gpt4all-j-v1.3-groovy.bin (Downloaded from gpt4all.io) - The model will get loaded - You can start chatting ## Benchmarks | Device Name | SoC | RAM | Model Load Time | Average Response Initiation Time | |-----------------|------------|-------|-----------------|----------------------------------| | iQoo 11 | SD 8 Gen 2 | 16 GB | 4 seconds | 2 seconds | | Galaxy S21 Plus | SD 888 | 8 GB | 7 seconds | 6 seconds | | LG G8X | SD 855 | 6 GB | Did not run | NaN | | Xiaomi Poco F5 | SD 7+Gen2 | 8 GB | 7 seconds | 5 seconds | | Pixel 6 Pro | Tensor | 12 GB | 5 seconds | 5 seconds | ## Demo - [Demo Video](demo.mp4) or https://youtube.com/shorts/J7DJ-40Uy1k?feature=share ## TODO - [x] Add very basic chat support - [ ] Reduce the size of the model - [ ] Fix UI issues ## Notes - It is very much buggy. I haven't tested it much. - You'll need a good device to run this. - Response initiation time and RAM usage for Chat Completion increases with the number of messages. - This is because Chat Completion is using Text Completion, and with every message the prompt size increases. - This could be fixed by training the model with Chat model in mind. I'm looking into this. ## Credits - nomic-AI
