LexiestLeszek

scrapeGPT

ScrapeGPT is a RAG-based Telegram bot designed to scrape and analyze websites, then answer questions based on the scraped content. The bot utilizes Retrieval Augmented Generation and webscraping to return natural language answers to the user's queries.

namegen

Self-contained, minimalistic implementation of a language model that generates coherent and normal sounding names. It uses an input dataset of names and probability distribution to generate new names based on the sequences of four characters.

sova_ollama

Open source implementation of Sova - RAG-based Web search engine using power of LLMs. Using Langchain, Ollama, HuggingFace Embeddings and scraping google search results.

InvestorGPT

InvestorGPT is an AI agent that analyzes the stock that dropped in its price recently and calculates chance for the company to recover the stock price back.

perplexity-clone-local

Perplexity.ai clone that works completely locally using open source tech.

qwen_rag_ollama

Simple implementation of RAG with new Qwen 1.5 model using Ollama, Langchain and HuggingFaceEmbeddings.

web-search-ollama-qwen-local

Local LLM Web search using qwen model and Ollama

Autonomous-CLI-Agent

Autonomous CLI Agent that uses Local LLM with Ollama to execute tasks in CLI.

hf_daily_papers_tg_bot

Bot that downloads Hugging Face daily papers and summarizes them using LLMs

ollama1-reasoning

o1-like reasoning with Ollama.

langmodel

Simple one-file implementation of a word-based language model that learns to predict the next word based on the previous one by training on a provided text dataset (Karpathy's Tiny Shakespeare in this case).

bookr

auto-email-sender

Simple tool to run email marketing campaigns.

compression

Simple compression algorithm

document_ranking

Document ranking based on heat maps.

flask-chat

Simple flask chat app using flask-socketio

hallucination_checker

A python library to check your LLM outputs for hallucinations.

incomment.js

JS Library for in-product commenting, enabling users to create threaded comments with unique identifiers, appearing on right-click at specified positions within a web application.

llm-ollama-inference

Playing around with Ollama to create function calling agent.

llmware

Providing enterprise-grade LLM-based development framework, tools, and fine-tuned models.

montestocks

Monte Carlo simulation for stocks

poker_engine

rag-ollama

Gardio app with RAG implementation using Ollama.

rl-holdem-agent

Reinforcement Learning agent that learns to play Texas Hold'em Poker.

spcheck

spell checking using Levenshtein's distance

talkr

Local AI assistant to talk about stuff (using Ollama and Whisper)

text2sql_CSV

Simple script using local LLM to ask CSV/XLS files with natural language.

tictactoe_ai

Tic Tac Toe agent implementation using Q-Learning.

twittorGPT

TwittorGPT is a tool that uses GPP4ALL model to generate tweets and then uses JS frontend Twitter API to post them.

web-user-behaviour

JS Library for user behaviour tracking from the browser, using mouse movements, clicks, scroll, and time on page.