Skip to content
View Suyashkb's full-sized avatar

Block or report Suyashkb

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Suyashkb/README.md

Hi there, I'm Suyash 👋

Typing SVG


B.Tech ECE (DTU, 2026)

I work on document image analysis and vision–language models, multimodal learning. I have personal interests in biomedical signal and image processing, financial analysis using machine learning .

Previously at HyperVerge, MIND Lab (IIT Delhi), IIT Jodhpur, and IIT Guwahati.


📄 Publications

VedOCR: A Dataset and Baseline for Sanskrit Manuscript OCR DALL Workshop @ ICDAR 2026, Vienna

MorphBG: Colour-Aware Binarization for Chromatically Stained Manuscripts Work-in-Progress track, IEEE ICIP 2026, Tampere

Multimodal Assessment of Emotional Regulation: EEG–fNIRS Based Quality Analysis of Self-Compassion Narrative Processing PREMI 2025 (poster) — with E. Chopra, N. Dhiman, P. Sahni


🛠️ Tools

Python · PyTorch · TensorFlow · Hugging Face Transformers · OpenCV · scikit-learn · NumPy/Pandas · LaTeX · MATLAB


LinkedIn Kaggle

Pinned Loading

  1. VedOCR VedOCR Public

    Expert-annotated OCR dataset and fine-tuned TrOCR baseline for Vedic Sanskrit manuscripts. DALL Workshop @ ICDAR 2026.

    Jupyter Notebook 1

  2. VLM-distillation VLM-distillation Public

    Distilling Florence-2 into a compact ~60M-param VQA student (feature-matching + attention transfer) — ~70% teacher-accuracy retention at 138x inference throughput with bf16 + torch.compile on H100.

    Python 2

  3. Hierarchical-Signal-Classification-using-Deep-Learning Hierarchical-Signal-Classification-using-Deep-Learning Public

    A hierarchical deep learning pipeline for automatic modulation classification of I/Q radio signals using CNN and CNN-LSTM architectures. Trained on 462,000 signals across 11 modulation types, achie…

    Jupyter Notebook

  4. Customer-Support-Chatbot Customer-Support-Chatbot Public

    This repo contains a fine-tuned version of DialoGPT (Conversational model of GPT-2) explicitly for customer support chatbots.

    Python 2

  5. Vmamba-explainer Vmamba-explainer Public

    Python

  6. AttentionMask-VLM AttentionMask-VLM Public

    Tests and Experiments for a attention based masking method to improve the robustness of VLMs

    Jupyter Notebook