One of our EMNLP papers got 6 reviews. That's insane!! Who has the energy to debate with 6 reviewers?!! 😂

So happy to see that our model, ChartGemma, is helping people analyze data in charts 📊

Happy to start a new chapter of my life! I am beginning my PhD in Computer Science this Fall at York University under the supervision of Prof. Enamul Hoque Prince. Building on my master's work, my research will focus on multimodal chart understanding and multimodal LLMs.

If you are planning to join MILA for graduate studies next year and are interested in working on Multimodal AI, especially document understanding, I highly recommend considering Sai as a co-supervisor.

Excited to dive into large language models and their applications to visualizations? Don’t miss my #LLM4Vis tutorial at IEEE VIS on Monday, October 14, 2024, from 8:30 AM to 11:30 AM EST! More details at nlp4vis.github.io. #ieeevis

Happy to announce our paper "ChartGemma" has been accepted to the COLING 2025 Industry Track! ChartGemma is a state-of-the-art chart understanding model📊developed over PaliGemma Check it out! 🚀 Demo: huggingface.co/spaces/ahmed-m… 📄 Paper: huggingface.co/papers/2407.04…

Thank you for highlighting our work! The first version of ColFlor did face some performance drops due to its smaller and faster design. We're now working on version 2, where we've made significant performance improvements without compromising efficiency. Stay tuned!

If you are looking for a cheap open-source alternative, check out our ColFlor model. It's super efficient (only 174M parameters) with strong performance on English Documents. 📄 Blog post: huggingface.co/blog/ahmed-mas… 💻 Demo: huggingface.co/spaces/ahmed-m… 🧠 Model:

Hosting Spaces with HuggingFace's ZeroGPU (A100s) via a Pro account is underrated! For just $9/month, I’ve hosted demos for models from my research like ChartGemma and ColFlor, with hundreds of weekly visitors. Demos make it easy for others to explore and adopt your work.