Project: llm-rk3588

Tech stack

Apache TVM, Machine Learning Compilation, CMake, Embedded system

Intro

Calling OpenAI’s API is the easiest way to access powerful Large Language Models. But I am not limited to this. I want bring a ChatGPT with me every where, without limition of the Internet.

That’s why I started this project, to make a portable Large Language Model, on ARM device, with GPU-acceleration.

Demo

check our blog for more information 🙂

Project: Hotel Reviewer

Tech stack

Python Flask, OpenAI API, Data Scraping

Intro

A good hotel stay can improve your overall experience when visiting a new city. Thus, reading the reviews from others could surely be helpful. However, searching and reading those reviews usually takes a long time, and I’d rather spend that time searching a places that worth to visit.

Therefore, I developed this API so that you only need to enter the name of the hotel you want to know about, and AI will automatically summarize the pros and cons of the place for you.

Demo

HTTP
curl http://{server_ip}:4878/reviews?hotel=Four%20Seasons%20Hotel%20San%20Francisco

Response

Plaintext
The Four Seasons Hotel in San Francisco offers excellence from start to finish. The staff provides a warm welcome and takes care of guests' needs, including fragile items in luggage. The check-in process is professional and welcoming, and the rooms are beautiful and comfortable. The hotel staff is friendly and helpful, providing great recommendations for local restaurants. The service is impeccable, with attention to detail and personalized touches. The property is well-maintained and offers a top-notch experience. The location is great, and the hotel provides access to a gym and pool club. However, some guests found the prices for amenities like cappuccino and valet parking to be steep. Overall, the hotel offers old-fashioned luxury and is well-located in the city center.

Project: Personal chat API

Tech stack

Python Flask, RESTful API, LLaMA

Intro

I am enthusiastic at new things and technology. As ChatGPT is a very hot topic recently; however, most of chatBot might using your chat data for training purpose, so I built a chatbot with a service running a LLaMA model remotely, it protects your privacy thoroughly your conversation.

Demo

HTTP
POST: http://{server_ip}:5050/c/vicuna

POST body:
line 2 – 4: History chat
line 5: New query

JSON
[
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "Hello there."},
    {"role": "assistant", "content": "Hi, how can I help you?"},
    {"role": "user", "content": "Who is Steve Jobs"}
]

Response:

JSON
{
    "choices": [
        {
            "message": {
                "content": " Steve Jobs was the co-founder and former CEO of Apple Inc., a technology company that designs and manufactures consumer electronics and computer software. He was widely regarded as a pioneer of the personal computer revolution and for his influential career, which spanned both the computer engineering and the entertainment industries. Jobs died on October 5, 2011 due to complications from pancreatic cancer.",
                "role": "assistant"
            }
        }
    ],
    "model": "ggml-nous-gpt4-vicuna-13b",
    "usage": {
        "completion_tokens": 411,
        "prompt_tokens": 323,
        "total_tokens": 734
    }
}

Project: Carbon Energy

also see this for Android client

Skills

System Design, Spring Boot, NoSQL (MongoDB), RESTful API, Android, Gradle, Git

Introduction

The earth’s climate directly affects our living environment. In the past 10 years, the average temperature of the earth has continued to increase at a rate of 0.32 F per year, and this will cause serious disasters, such as wildfires, hurricanes, floods, famines, etc. Increases in greenhouse gases (carbon emissions) are the main cause of these climate problems.

I want to use my knowledge to protect the beautiful environment we are living in. So I built this platform to encourage people to walk more and reduce the use of cars and other modes of transportation that increase greenhouse gases.

In addition, users can plant a real tree using the points they earned to reduce the impact of greenhouse gases and protect our environment.