deep-learning

A minimal implementation of a denoising diffusion model in PyTorch.

Finetune LLaMA-7B with Chinese instruction datasets

Implementation of STAM (Space Time Attention Model), a pure and simple attention model that reaches SOTA for video classification

Official implementation of the WACV 2025 paper "Pix2Poly: A Sequence Prediction Method for End-to-end Polygonal Building Footprint Extraction from Remote Sensing Imagery".

[AAAI 2024] The official PyTorch implementation of "Diverse and Aligned Audio-to-Video Generation via Text-to-Video Model Adaptation"

A comprehensive AI learning roadmap covering Python fundamentals, mathematics, machine learning, deep learning, LLMs, and agentic systems — focused on hands-on projects, practical tools, and real-world deployment.

Simple Implementation of Pix2Seq model for object detection in PyTorch

Efficient World Models with Context-Aware Tokenization. ICML 2024

A curated list of the top 10 computer vision papers in 2021 with video demos, articles, code and paper reference.

StableDiffusionTelegram is a telegram bot that allows to generate images using the stable diffusion AI from a telegram bot, in a much more comfortable and simple way.

Application of Connectionist Temporal Classification (CTC) for Speech Recognition (Tensorflow 1.0 but compatible with 2.0).

BanglaSpeech2Text: An open-source offline speech-to-text package for Bangla language. Fine-tuned on the latest whisper speech to text model for optimal performance.

AI-powered digital picture frame. Generate captivating and unique art from spoken conversations.

[SIGGRAPH 2024] "EASI-Tex: Edge-Aware Mesh Texturing from Single Image", ACM Transactions on Graphics.

A local and uncensored AI entity.

Implementation of RQ Transformer, proposed in the paper "Autoregressive Image Generation using Residual Quantization"

Visual knowledge bank for understanding large language models, with 180 concept cards from tokenization to deployment.

A complete collection of RAG interview questions, answers (548 questions & 41 RAG types), system design scenarios, architecture patterns, and production-ready concepts.

Implementation of the Large Behavioral Model architecture for dexterous manipulation from Toyota Research Institute

Segmentation deep learning ALgorithm based on MONai toolbox: single and multi-label segmentation software developed by QIMP team-Vienna.