We are the Cynaptics Club of IIT Indore, the AI/ML enthusiasts' hub, and we're excited to present you with our induction task! This induction is open to both first-year and second-year students, with tailored requirements for each. For first years, Task 1 is mandatory, while Task 2 is optional as a bonus. For second years, both Task 1 and Task 2 are compulsory. Submission guidelines and detailed explanations for each task can be found in their respective task folders. Best of luck, and we look forward to your innovative solutions!
Sub-Task 1: AI vs. Real Image Classification
Your objective is to classify AI-generated images from real ones. We've provided a dataset and a baseline code to help you get started. Train your model and aim for higher accuracies by experimenting with different techniques.
Sub-Task 2: Custom GAN Training
In this task, you are required to pick a dataset of your choice from the internet and train your own GAN (Generative Adversarial Network) model to generate images. Showcase your creativity and technical skills!
Sub-Task 3: Design Your Own X-mas GAN
Get into the holiday spirit by designing your very own Christmas-themed GAN. Let your model generate festive and unique outputs
Detailed descriptions for each sub-task are provided in their respective folders. Good luck, and have fun exploring these challenges!
Note: You will be evaluated only on sub-task 1 and 2. Sub-task 3 is for you to explore GANs more deeply.
In this task, your challenge is to train a machine learning model capable of playing the classic Hangman game. We've provided you with baseline code for both playing the game and building the model. Your goal is to enhance the model's performance by improving its accuracy and efficiency. Experiment with different strategies and techniques to make your model a Hangman champion!
Detailed instructions and guidelines can be found in the task folder.
Note: This is a bonus task for both the first years and second years.
In this task, you are supposed to fine-tune a LLM to a custom persona-based chat dataset provided to you. The main goal of the task is to fine-tune an LLM of your choice in limited computation power to align with the persona of a human and give a better experience to user to chat with the LLM.
Dataset and details can be found in the respective task folder.