Code for "Counterfactual Token Generation in Large Language Models", Arxiv 2024.
-
Updated
Nov 7, 2024 - Jupyter Notebook
Code for "Counterfactual Token Generation in Large Language Models", Arxiv 2024.
A GenAI RAG assistant designed to comprehend your PDF's context and provide accurate answers to your questions quickly-give it a try!
This project is a search-powered AI chatbot built with Streamlit and LangChain, designed to retrieve information from multiple sources like Wikipedia, Arxiv, and DuckDuckGo. It leverages LangChain's agent framework and Groq API for real-time, intelligent responses.
The Pictionary app uses LLaMA 3.1 to generate random drawing prompts and LLaMA 3.2 Vision to predict and judge user drawings based on these prompts. It provides an interactive and fun way to test your drawing skills within a set time limit.
Trained and evaluated traditional ML models, fine-tuned Dolphin 2.9.4 based on the Llama 3.1 (8B) model, and processed Bangla text to classify sentiments. (bnlp, nltk, bnlp_toolkit, banglanltk, huggingface_hub, transformers, torch)Β
A generalized framework for subspace tuning methods in parameter efficient fine-tuning.
Effortless Data Extraction, Powered by : Generative AI
Retrieval-Augmented Generation on YouTube transcripts and PDFs to deliver accurate and contextual answers.
This repository contains a real-time chatbot app using Groq's API and various LLMs. Built with Streamlit, it provides an interactive interface to select and chat with different models. Users can easily set up and run the chatbot locally.
Cloning Yourself using your whatsapp chat history and training a model on it.
Advanced AI functionalities, including tool usage, context aware similarity with Ollama models
Captionify is a versatile and user-friendly image captioning project that harnesses the power of two remarkable models: Llama-2-7b and Llama-3 for natural language understanding and Blip-Image-Captioning-Large for image caption generation.
βοΈ Fine-Tune π¦ Llama 3.1, Phi-3.. Models on custom DataSet using π΄οΈ unsloth & Saving to HuggingFace Hub
Try it now
RepoGenius aims to create a distributed system that, starting from a GitHub link or certain parameters, performs an analysis on the reference repository based on elements such as code, language or files
π Streamlit App : Weekly News Letter Crew AI Agents ποΈ
π Blog Writer Crew AI Agents - Streamlit App ποΈ
"Ask your PDF" ChatBot : Streamlit App, LangChain, llama3, Nomic embeddings
Add a description, image, and links to the llama3-8b topic page so that developers can more easily learn about it.
To associate your repository with the llama3-8b topic, visit your repo's landing page and select "manage topics."