Skip to content

Instantly share code, notes, and snippets.

@halitbatur

halitbatur/AI.md

Created August 20, 2024 08:59
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save halitbatur/bfe89807b5b3d458ce5686aa87c16b6b to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save halitbatur/bfe89807b5b3d458ce5686aa87c16b6b to your computer and use it in GitHub Desktop.
Download ZIP
Ai related questions
Raw
AI.md

Discussion questions about AI

  1. What is Hugging Face, and what role does it play in the AI community?
  2. What is a Large Language Model (LLM)?
  3. What are some popular LLMs available on Hugging Face?
  4. How do you use a model from Hugging Face in a Python project?
  5. What is Gradio, and how does it simplify AI demos?
  6. What are some examples of AI applications built using Gradio?
  7. How can Gradio be integrated with Hugging Face models?
@thabiso-makolana
Copy link

thabiso-makolana commented Aug 20, 2024 •
edited
Loading

Lethukuthula
Samuel
Gerald
Thabiso

  1. Hugging Face is a company known for its work in natural language processing (NLP).They provide a platform where developers and researchers can share machine learning models. Hugging Face makes it easy to find, use, and share AI models, which has helped drive innovation and collaboration in AI.
  2. LLM is a type of AI model designed to understand and generate human language. These models are trained on a lot of text data, the more the better.
  3. GPT A model designed for generating human-like text.
    BERT A model good at understanding the context of words in a sentence.
    T5 A model that can convert one type of text into another, like translating languages or summarizing text.
    RoBERTa An improved version of BERT, trained with more data and better techniques.
    Llama - used for multilingual dialogue use case, Falcon - used for SEO optimisation, StableLM - processes complex prompts and generates clear text
  4. Hugging Face provides a user-friendly Python library called transformers that simplifies the process of loading and using pre-trained models.
  5. Gradio is a Python library that allows you to create user interfaces for your machine learning models with just a few lines of code. It's particularly useful for building interactive demos and sharing your work with others.
  6. Artbreeder: Creates and evolves images using neural networks.
    Stable Diffusion WebUI: Provides a user-friendly interface for interacting with Stable Diffusion models.  
    Pixray: Generates images from text descriptions using various AI models.  
    Dream Booth: Fine-tunes text-to-image models on a specific subject.  
    Reimagine Home: A tool for visualizing home renovations.
  7. Create a Gradio interface: You use Gradio to define how users will interact with the mode. Link the model to the interface, you connect the model’s function to the Gradio interface so that when users input something, the model processes it and displays the result.

@PhamelaMhlaba
Copy link

PhamelaMhlaba commented Aug 20, 2024

Team Members( Phamela, Nonhlahla and Ntandoyenkosi)

Answers.

  1. Hugging Face is an open-source platform for natural language processing (NLP) and computer vision projects1. It provides tools, resources, and a community for developers, researchers, and enthusiasts to work on NLP and machine learning234. Their goal is to make AI more accessible and democratize NLP research.

2.Large language models (LLM) are a type of machine learning model, which are trained on huge sets of data, that can recognize and generate text, among other tasks. An LLM is a program that has been fed enough examples (data gathered from the internet) to be able to recognize and interpret human language or other types of complex data.

  1. Some popular Large Language Models (LLMs) available on Hugging Face include:

GPT-4 by OpenAI: An advanced model known for its versatility and conversational abilities.
LLaMA (LLaMA 2) by Meta: A family of models designed for efficiency and scale.
BERT by Google: A transformer model optimized for understanding the context in text.
Falcon by TII: High-performance models focusing on open-source NLP tasks.
Bloom by BigScience: A multilingual model that supports 46 languages and 13 programming languages.

  1. To use a Hugging Face model in a Python project:

Install the Transformers library with pip install transformers.
Load a pre-trained model and tokenizer using AutoTokenizer and AutoModel classes.
Tokenize your input text and run it through the model for inference.

  1. Gradio is an open-source Python library designed for quickly creating demos or web applications for machine learning models, APIs, or any Python function. It includes built-in sharing capabilities for easy distribution, eliminating the need for expertise in JavaScript, CSS, or web hosting. Gradio serves as a link between your machine learning models and users, enabling the creation of user-friendly interfaces with minimal effort.

    • An interface can be built using Gradio for recommendation predictions using a trained recommendation model that suggests movies based on user input, such as their favorite genres or previously liked movies.
  • Image Classification Demos where users can upload an image (such as a picture of an animal), and the application will classify the image (e.g., "cat," "dog," "bird") using a pre-trained model.
  • Translation Services where a user inputs a sentence in English, and the application translates it into another language
  • Text Generation Demos where a user inputs a text prompt and an LLM generates a continuation of the text or generate a better spelling (auto-corrector). Gradio can be used to create an interactive interface for such an application.
  • Text-to-Speech Conversion where Gradio can be used to create an interface for user to type in text and listen to the audio output from text-to-speech LLMs
  1. Gradio can be seamlessly integrated with Hugging Face models to create intuitive interfaces for machine learning tasks. This involves loading a model from Hugging Face using the transformers library and then building a web interface with Gradio. Once installed, models like GPT-2 or BERT can be loaded and linked to a Gradio app, allowing users to input data and receive results, such as text generation or sentiment analysis. Gradio apps can be run locally or deployed publicly on Hugging Face Spaces, making them easily shareable with minimal coding effort.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment