🗂️ Table of Contents
short.mp4
Welcome to Comic-ify, a revolutionary application fueled by cutting-edge LLM-based technology. Simply upload your mundane PDF, and prepare to be dazzled as it metamorphoses into a visually stunning comic-style masterpiece! Say goodbye to drowning in endless paragraphs and hello to a world where information leaps off the page, guiding you to your rightful destination in the most captivating way possible! Get ready to redefine your reading experience – Comic-ify is here to elevate it to unprecedented heights!
Comic-ify is not just another tool. It's a revolution in how we consume information. Imagine this: you have a stack of boring PDFs, full of dense text that drains your energy just thinking about it. We've all been there. But fear not! Comic-ify is here to transform your reading experience. Let's face it, deciphering pages of dull, overwhelming text is like trying to climb a mountain without a rope. It's tough, it's exhausting, and frankly, it's no fun at all. We've experienced this struggle since our early days, constantly grappling with the challenge of making tedious content engaging.
Lexicon, our fine-tuned model is employed to generate imaginative text and produce images inspired by the text. captivating readers as they explore the content
-
Input as prompt or PDF Displays modified text/images for prompts, generates a downloadable PDF for PDF input.
-
Theme-based generation Users can choose predefined themes. For example, in the 'Indian Mythology' theme, Lord Vishnu could provide insights into the mysteries of black holes in the universe.
-
Highlight User’s Prefered topics Example: If the emphasis is on Operating Systems, users can prompt for a detailed explanation of how system calls work, directing the explanation accordingly.
Comic-ify is revolutionizing how we engage with information. Powered by cutting-edge LLM-based models, our tool effortlessly converts dense PDFs into visually engaging comic-style content. Say goodbye to tedious paragraphs and hello to an immersive experience where information comes to life! Why Comic-ify? Because we believe knowledge should be accessible and enjoyable. With Comic-ify, understanding content becomes effortless, and the process becomes a pleasure. It's a win-win for all.
✨ Dataset: We have pushed our data and model to HuggingFace, however it is currently made private.
Explore our demo video on YouTube
YouTube Demo
Clone the Comic-ify repository to your local machine using the following command:
git clone https://github.com/S0L009/COMIC-IFY_OneAPI.gitRun the FastAPI server locally to enable all endpoints
Step1: Install the FastAPI directory from the repository and navigate into it in your terminal
Step2: Install all necessary dependencies
pip install -r requirements.txtStep3: Run the uvicorn ASGI server to get the endpoints functional
uvicorn main:app --reloadStep4: Test the APIs on Swagger-UI (your app should look like this)
| Sno | Intel Product Used | Description |
|---|---|---|
| 1 | Intel® Distribution for Python® | We are using this toolkit as it provides optimized Python libraries for numerical computing and data manipulation |
| 2 | Intel® OpenVINO™ | We are making the process of implementing Q-LoRA faster using OpenVINO |
| 3 | Intel® DevCloud | DevCloud offers a valuable environment for our project |
By leveraging the power of Intel® oneAPI libraries and frameworks, our models achieve remarkable performance enhancements and optimized memory utilization across various applications. The seamless integration of different components contributes to faster fine-tuning, efficient inference, and an improved overall user experience.
Different explaination styles
- Get your feet wet: This option, gives a short introductory summary of the PDF that the user has given.
- Deep Dive: This option, provides a deep and detailed description about the PDFs that the user has given as input
- Splash and Submerge: This option gives an short summary, but not as deep as the option "Deep Dive"
.png)
.png)
.png)
Comicify is currently under active development, but you can already experience its early capabilities by running the alpha version. Our dedicated team is diligently working on achieving the following key milestones:
-
Expanding Templates and Themes: We're dedicated to enriching the Comicify experience by adding more templates and themes for the output. These new designs will offer users a wider range of options to customize their comic-style content, ensuring greater flexibility and creativity in transforming text-heavy materials.
-
Scaling the Product: As Comicify gains momentum, we're focusing on scaling the product to accommodate growing demand and usage. This involves optimizing performance, enhancing scalability, and ensuring seamless user experiences even as our user base expands.
-
Innovative Feature Research: Our team is actively researching new and innovative features that align with our mission of making content consumption enjoyable. From interactive elements to gamification features, we're exploring ways to infuse Comicify with engaging functionalities that enhance user engagement and retention.
-
Continuous Improvement: Beyond adding new features, we're committed to continuously improving Comicify based on user feedback and emerging trends. This includes refining existing features, optimizing workflows, and addressing any issues or pain points identified by our users.
As we continue to evolve Comicify, our goal remains clear: to revolutionize the way people consume and interact with information by making content fun, engaging, and visually appealing.
Languages and Frameworks
 |
 |
 |
|---|
Libraries and Tools
 |
 |
 |
|---|
Functionalities
 |
 |
|---|
✅ Utilizing the Intel® AI Analytics Toolkit: We make the most of the Intel® AI Analytics Toolkit to supercharge our data science and analytics workflows on Intel® architecture. With familiar Python* tools and frameworks and optimized oneAPI libraries, we ensure peak performance across every task—from data preprocessing to deep learning and machine learning. Plus, our toolkit promotes seamless model development with interoperability.
✅ Seamless Adaptability: Thanks to our integration with the Intel® AI Analytics Toolkit, we smoothly handle machine learning and deep learning workloads with minimal adjustments.
✅ Fostered Collaboration: Our journey involved collaboration with experts in deep learning and data analysis, highlighting the importance of teamwork in achieving our shared goals.
Meet the talented minds behind Comic-ify, all hailing from Amrita Vishwa Vidyapeetham's AI branch:
Surya
Amrit
Navneet
Srikar
In our journey, we were fortunate to have the support of numerous individuals who lent their expertise and encouragement. Special thanks to Kavya Sakthivel, Premjith Sir, Dilip Parasu, Pratiti Bhadra mam and Vipin Das sir for their active support throughout the process. Your contributions were invaluable to our success.
If you're passionate about improving information consumption and want to contribute to Comic-ify, here's how you can get started:
- Fork the Comic-ify repository on GitHub.
- Make your desired changes or enhancements.
- Submit a pull request detailing your modifications.
- Collaborate with our team to review and integrate your contributions.
We welcome contributions from developers, designers, and enthusiasts alike!