Wojtekb30's Blog (EN)

Hello readers! Me and my friend Patryk (who is also my classmate, we're both studying computer science at the same university) are taking part in the 2026 edition of IBM AI Racing Competition. The goal of it is to develop an AI agent capable of driving a car in Torcs simulator on "Corkscrew" track and perform a full lap. For our solution, we trained an TD3 RL neural network agent to drive the car. Reinforcement Learning networks can control things and learn thru points - in a nutshell, they recieve positive points (reward) for doing something correctly, and negative points (punishment) for doing something incorrectly. They learn how to behave to maximize the reward. Of course, they lack understanding of purpose, so we have to program constrains and how the point system works ourselves. It is challenging, as they like to find workarounds to get free points or stop losing them. If the grading system is designed wrong, the neural network may end up learning to quickly fail o...

Hello everyone! In this post I will tell you about how I trained a GPT-2 "Small" AI model from scratch on a huge dataset of around 32 GB of raw text files with use of Torch and HuggingFace's Transformers libraries. I will explain steps of such a project, tell you about how it went for me and how I recommend doing these things now, and of course I will show you the code that pre-trains the AI model from a dataset too. Let's start from the beginning: 1. The idea, purpose and raw dataset First and foremost, a project like this needs an idea and should have an purpose. I pre-trained (trained from scratch) the AI model just to learn more about the process, but pre-training LLM AI models of that scale is done rarely and is usually pointless, as there are lot of base models already trained on specific languages that can then just be fine-tuned into a specific response format for example. But let's assume that we want to pre-train such an model anyway, for example as an a...

Hello. In this follow-up, we will continue development of the solution to let ChatGPT describe videos. In my approach, the video is split into fragments and then these fragments are turned into "comics". Then, every fragment gets described and all the descriptions are sent to the master AI agent, which interacts with the user.  Since the last post, I updated the code to make it more modular and move AI processiong out of initalizer: from moviepy.video.io.VideoFileClip import VideoFileClip from PIL import Image import os import easyapiopenai from WojSafeAPI import * #My own library for loading API keys safely, you cannot use it. class Video_ChatGPT: def __init__(self,openai_api_key:str, video_path: str, time_between_frames_seconds: float, frames_per_scene: int, process_audio=True, frame_height = 480, watcher_model='gpt-4o-mini',master_model='gpt-4o', watcher_token_limit = 500, master_token_limit=15*10...

Hello everyone! Today, we will create a way to let ChatGPT watch videos. But first, I will describe my approach to this problem. So models like GPT-4o and GPT-4o-mini are able to analyze static images. My approach to let the AI analyze movies will look like this: - The video will be split into individual frames on specific interval. - Certain amounts of the frames will become scenes (for example 8 frames when the frames are took every 4 seconds will make a 32 second long scene in total). The scenes will then be assembled into images containing all the frames, creating sort of a "comic". - If set to do so, audio fragment for each scene will be processed thru speech recognition to get transcript of dialogues. As I am writing this, OpenAI released a model gpt-4o-audio-preview, which they claim can analyze audio; but since I haven't used it yet, let's first extract speech with a classic speech recognition library. In this part I am not going to add audio support yet tho. ...

ISKRA experiment 02_10_2024-1 - "Defeat the Minecraft Creeper" Hello everyone! In this post, I will describe my recent and first public ISKRA Project experiment. I have conducted some experiments before, but I believe they are not worth publishing. Let's consider this one the first official ISKRA Project experiment. Let's begin with the description of it. I've built an installation in Minecraft Bedrock Edition: This was the starting point and viewport of our AI: The goal of the AI was to defeat the Creeper by using a set of pre-defined control commands: W - move forward 1 block S - move backwards 1 block A - strafe left 1 block D - strafe right 1 block P - place a brown dirt block in front of you R - destroy block in front of you F - punch to attack/fight The commands P and R were never actually implemented since they were unnecessary for this experiment. The Creeper was trapped in a structure to prevent...

Hello everyone reading this! Before I start with the topic of this post, I would like to inform that due to being busy next  ISKRA Project  articles will sadly be delayed. Sadly, I won't be able to publish them in regular intervals. Anyway, let's get into the topic I would like to write about here. I have fine-tuned different AI models in the past, but I never fine-tuned a latest OpenAI's GPT model. Some time ago, I decided to finally try that. But for those of you who don't know, I will first explain what fine-tuning is.  In a nutshell, it is a process of training artificial intelligence, usually to shape its output (responses) in a specific manner, teach it dealing with new tasks or to provide it with new data. It is usually done by providing lot of examples of input (for example questions) and desired output for each of the inputs (for example correct answers to all the questions).  A example could be: if you fine-tune a text generator AI on theatrical play script...