Projects

Overview: This repository contains the source code and results for the paper "Unified-modal Salient Object Detection via Adaptive Prompt Learning" presented at TCSVT 2025. Key features: - Unified-modal Salient Object Detection framework - Baseline SOD framework comparison - Performance metrics for RGB, RGB-D, and RGB-T modalities Tech stack: Python Why it matters: This project offers a comprehensive solution for salient object detection across different modalities, demonstrating the effectiveness of adaptive prompt learning in unified modal detection. Repository: Angknpng/UniSOD GitHub stars: 13 GitHub description: Unified-modal Salient Object Detection via Adaptive Prompt Learning

### Overview: This repository contains the code for the paper "Generating Diverse and Meaningful Captions: Unsupervised Specificity Optimization for Image Captioning" by Annika Lindh et al., presented at ICANN 2018. The project focuses on improving image captioning models to generate more diverse and meaningful captions without human supervision. ### Key Features: - **Unsupervised Specificity-Guided Optimization**: The code optimizes image captioning models to encourage diversity in the generated captions. - **Image Captioning Model**: The repository includes a specific image captioning model that is trained unsupervised to generate captions that are both diverse and meaningful. - **Unsupervised Learning**: The project leverages unsupervised learning techniques to enhance the diversity of captions. ### Tech Stack: - Python 2.7 - PyTorch 0.3.1 - torchvision - h5py 2.7.1 - sklearn 0.19.1 - scipy 1.0.1 - scikit-image (skimage) 0.13.1 - ijson ### Why It Matters: The project aims to address the challenge of generating captions that are both diverse and meaningful, which is crucial for applications such as image retrieval and image captioning. By optimizing the model to focus on diversity, the project seeks to improve the quality of generated captions, making them more useful and relevant for various applications. ### Repository: The repository is hosted on GitHub and is described as "Generating Diverse and Meaningful Captions

### Overview: This guide provides a comprehensive setup and configuration for a 2-node NVIDIA DGX Spark cluster, optimized for distributed training and inference tasks. The cluster is equipped with 200 Gb/s direct fabric networking and NVIDIA's latest NVMe-TCP shared storage, ensuring high performance and efficiency. ### Key features: - Distributed training with DDP scaling efficiency of 93.5% - CUDA inference with EXO model serving - NCCL tuning for optimal performance - 200 Gb/s direct fabric networking - Grace Blackwell GB10 DGX Spark systems connected point-to-point - NCCL configuration and GID index setup - End-to-end distributed training pipeline - GPU Direct RDMA explanation and solution ### Tech stack: Python ### Why it matters: This guide is essential for developers and researchers looking to deploy and optimize distributed training and inference tasks on the DGX Spark cluster, ensuring high performance and efficient resource utilization.

Overview: This project utilizes the Stable Diffusion Pipeline to generate images from text prompts, offering a simple GUI application for users to input prompts and view generated images. Key features: - Generates images based on user prompts - Uses the Stable Diffusion Pipeline from the diffusers module - Provides a Python GUI application - Supports text-based image generation Tech stack: Python, Jupyter Notebook Why it matters: This project simplifies the process of generating images from text prompts, making it accessible for a wide range of users, including artists, designers, and anyone interested in exploring the intersection of text and image generation.

Overview: Engram is a decentralized vector database on Bittensor, designed to store and retrieve text, images, and documents securely and permanently. It leverages content-addressed semantic memory to ensure each piece of knowledge has a unique identifier, eliminating the need for a central authority. Key features: - Content-addressed: Each piece of knowledge is uniquely identified by its CID, ensuring consistency across miners. - Decentralized: Replicated across competing miners on Bittensor subnet 450, providing redundancy and fault tolerance. - Permanent: Stores binary files (images, PDFs) on Arweave, ensuring permanence and durability. - Private: X25519 hybrid encryption with differential privacy noise, and namespace isolation at every endpoint. - Incentivized: Miners earn TAO for storing and serving vectors, incentivizing the network. - Verifiable: Merkle commitment over the full memory corpus with HMAC challenge-response proofs for transparency. - Mobile: Mine from your phone via managed Akash Network nodes, with pay-as-you-go usage on Base with USDC. Why it matters: Engram replaces a single-vendor vector database with a decentralized, incentivized network, ensuring the security and reliability of AI memory without relying on a single point of failure. Tech stack: Python Why it matters: Engram is a Bittensor subnet that turns text, images, and documents into permanently stored, content-addressed memories, providing a decentralized and private solution for AI memory. GitHub stars: 13 Why it matters: The repository has a moderate number of

### Overview: Qwen3.8-Flash-Next is a high-performance AI model optimized for running on a single RTX 3090 with 64 GB of RAM. This setup allows for up to 2,100 tokens per second prefill and 47 tokens per second decode, making it ideal for tasks that require quick response times. ### Key Features: - **GPU:** One NVIDIA RTX 3090 with 24 GB of memory. - **RAM:** 64 GB of system RAM. - **Runtime:** vLLM with hot MoE experts on the GPU and cold experts computed on the CPU. - **Int8 KV Cache:** Utilizes INT8 KV cache to optimize memory usage. - **Docker:** Runs in a Docker container for easy deployment. - **API Compatibility:** Compatible with OpenAI's API. ### Tech Stack: - Python-based runtime. ### Why It Matters: This setup provides a powerful and efficient way to run Qwen3.8-Flash-Next, leveraging the full potential of the RTX 3090 GPU and 64 GB of RAM to achieve high performance and fast response times. The use of INT8 KV cache and hot MoE experts ensures optimal performance and resource efficiency. ### Repository: Repository: DominikBucko/qwen38-flash-next-3090 GitHub Stars: 13 GitHub Description: Run Qwen3.8-Flash-Next on ONE RTX 3090 (24 GB) +

### Overview: This repository is a comprehensive guide to Large Language Models (LLMs), covering both theoretical and practical aspects of these models. It provides insights into the development, evaluation, and applications of LLMs, focusing on their capabilities and limitations. ### Key Features: - **Theoretical and Practical Aspects**: Covers both the theoretical foundations and practical implementation of LLMs. - **Statistical Language Modeling**: Explains N-gram models and their applications. - **Neural Modeling**: Discusses neural language modeling techniques (NLM) and their importance. - **Evaluation of LMs**: Details methods for evaluating the performance of language models. - **Transformer-Based Language Models**: Focuses on the advancements in transformer-based models. - **Generative Configuration Parameters**: Provides guidance on configuring parameters for inference. - **Computational Challenges and Quantization**: Addresses the computational challenges and techniques for efficient model deployment. - **Prompt Engineering**: Covers techniques for in-context learning and auto prompt engineering. - **Fine-Tuning**: Discusses the process of fine-tuning models for specific tasks. - **RAG (Retrieval-Augmented Generation)**: Explains how to incorporate retrieval in language models. - **LangChain**: Demonstrates how to build applications using LangChain. - **LLM Model Evaluation**: Provides evaluation metrics and techniques for LLMs. - **AI Agents**: Introduces the concept of AI agents and their integration with LLMs. - **References**: Offers additional resources for further reading. ### Tech Stack: The primary language used in this repository is Jupyter

### Overview: This repository is a comprehensive guide to system design for the AI era, focusing on modern patterns and infrastructure for large language models (LLMs), vector databases, and agent architectures. ### Key Features: - **Modern Patterns**: Covers AI infrastructure patterns that have been deployed in production after 2022. - **Original Synthesis**: Provides original explanations and diagrams instead of aggregating links. - **Reference Manual**: A free, open-source reference manual designed for software engineers preparing for system design interviews or building production systems using LLMs, agents, and vector databases. ### Tech Stack: The repository uses Markdown for documentation and Mermaid for flowcharts. ### Why It Matters: - **Beyond the Primer**: Adds AI infrastructure patterns that have reached production after 2022. - **Original Synthesis**: Provides original explanations and diagrams instead of aggregating links. - **Reference Manual**: A free, open-source reference manual designed to be maintained indefinitely. ### Who This Is For: - **Software Engineers**: With 2+ years of experience, preparing for system design interviews at L4–Staff level or building production systems using LLMs, agents, and vector databases. - **Not For**: Total beginners, ML researchers, or those seeking code tutorials. ### Structure Overview: ```mermaid flowchart LR F[Foundations] --> A[AI Era] A --> M[Modern Patterns] M --> C[Case Studies] C --> I[Interview Guide] ``` ### Table of Contents: - **Foundations** - [

Overview: This project implements four different multi-modal neural machine translation models based on the PyTorch port of OpenNMT, an open-source neural machine translation system. Key features: - Four different multi-modal neural machine translation models - Uses PyTorch and torchtext for implementation - Based on the research papers by Iacer Calixto - Supports PyTorch version 1.3.1 and torchtext version 0.2.3 - Includes pre-processing scripts for text and image data Tech stack: PyTorch, torchtext Why it matters: This project provides a practical demonstration of multi-modal neural machine translation models, showcasing their effectiveness in handling multimodal data for translation tasks.

### Overview: Kedro Serving is a plugin for the Kedro framework that allows developers to serve Kedro Pipelines as an API. This plugin simplifies the deployment and accessibility of complex data pipelines, making it easier to share and integrate with other applications. ### Key Features: - **Serves Kedro Pipelines as APIs**: Easily expose Kedro pipelines as RESTful APIs, enabling seamless integration with other systems. - **Integrated with Kedro**: Seamlessly integrates with the Kedro ecosystem, ensuring compatibility and ease of use. - **Flexible Configuration**: Configurable settings for API endpoints, authentication, and more. ### Tech Stack: Python ### Why It Matters: Kedro Serving simplifies the deployment and accessibility of Kedro Pipelines, making it easier to share and integrate with other applications. This plugin is particularly useful for developers working on complex data pipelines who need to expose their workflows as APIs for external use.

### Overview: This repository contains the evaluation code for a paper published in *npj Digital Medicine* titled "*An In-Depth Evaluation of Federated Learning on Biomedical Natural Language Processing for Information Extraction*". The datasets used in the paper include the NCBI-Disease, 2018 n2c2, and GAD datasets for named entity recognition (NER) and relation extraction (RE). ### Key features: - Federated Learning Implementation - Comparison of LLMs with Blue BERT and GPT 2 trained with federated learning - Results show improvements in F1 scores for NER and RE tasks compared to Blue BERT and GPT 2 ### Tech stack: Jupyter Notebook ### Why it matters: This repository provides a comprehensive evaluation of LLMs in the context of biomedical NLP, demonstrating the effectiveness of federated learning in improving information extraction tasks.

Overview: This repository contains an implementation of a recommendation model using Milvus, a vector database designed to store and search high-dimensional vectors efficiently, deployed with Docker. The recommendation model leverages the capabilities of Milvus to perform fast similarity searches, enabling efficient retrieval of similar items based on user preferences or item attributes. Key features: - Utilizes Milvus for efficient storage and retrieval of high-dimensional vectors. - Implements recommendation algorithms for personalized recommendations. - Supports both user-based and item-based recommendation strategies. - Provides easy-to-use APIs for integrating the recommendation model into existing applications. - Supports scalability for handling large datasets and user bases. - Dockerized deployment for easy setup and management. Tech stack: Jupyter Notebook Why it matters: This project showcases the integration of Milvus, a powerful vector database, with a recommendation model, demonstrating how it can be used to efficiently retrieve similar items based on user preferences or item attributes, making it valuable for applications needing personalized recommendations. Repository: HGSChandeepa/Recommendation-Model-using-Milvus-vector-db GitHub stars: 13 GitHub description: Recommendation Model using Milvus vector db

Overview: Nebula is a Kubernetes control plane designed to manage GPU workloads across various cloud environments, including NeoClouds, hyperscalers, and on-premises infrastructure. Key features: - Kubernetes-native experience for existing Deployments or Jobs. - Multi-provider support with modal, RunPod, and AWS providers. - Policy-driven placement for flexible NodePools. - Cost tracking for per-instance cost. - Observability with rich metrics and logs. Tech stack: Go Why it matters: Nebula simplifies GPU workload management across different cloud providers, ensuring seamless and cost-effective GPU usage without the need for complex per-cloud glue.

Overview: BAGM is a backdoor attack designed to manipulate text-to-image generative models, targeting three stages of the generative process. This attack increases bias in generated outputs by up to 515% without compromising model robustness. Key features: - Targets three unique text-to-image architectures across three stages of the generative process. - Modifies the behavior of the embedded tokenizer and pre-trained language and visual neural networks. - Includes surface, shallow, and deep attack forms. - Quantitative metrics for assessing backdoor attacks on generative AI models. - Utilizes a Marketable Foods dataset for the target domain. Tech stack: PyTorch, NVIDIA GeForce RTX 4090 GPU, cloud computing services. Why it matters: BAGM highlights the security challenges in generative AI, providing a framework for preemptive efforts to address these challenges.

Overview: A simple tool for taking a picture of text on a website and extracting the text. Key features: - Takes a picture of text on a website. - Extracts the text from the image. - Can be used on any website. Tech stack: Vue.js Why it matters: This tool simplifies the process of extracting text from websites, making it easier for developers to integrate text extraction into their applications.

### Overview: JimmyMax/ninfer-v100 is a C++/CUDA inference engine optimized for a single Tesla V100-32GB GPU (compute capability 7.0, `sm_70`). It supports the Qwen3.8-27B family, Qwen3.8-Flash-Next, and Gemma 4, with native NVFP4, FP8, and GGUF-derived artifacts, achieving 1.00k tokens per second on inference and 78+ tokens per second on decoding. ### Key features: - **Operator Routing and Weight Prepacking**: Optimizations applied at load time or on first use, ensuring no route changes. - **Native NVFP4 and FP8 Profiles**: Supports quantized models from HF safetensors, compressed-tensors, and block-FP8 sources. - **GGUF-Converted Artifacts**: Converts models from `.gguf` format, providing a 2-4 bit block format with mixed-precision artifacts, achieving 12% below native profile. - **27K-Tok Prefill and 1.00k Tokens/s Inference**: Prefills 27K tokens and achieves 1.00k tokens per second on inference. - **27K-Tok Prefill and 78+ Tokens/s Decode**: Prefills 27K tokens and decodes 78+ tokens per second on decoding. ### Tech stack: C++/CUDA ### Why it matters: JimmyMax/ninfer-v

Overview: This project includes ten secure ComfyUI nodes designed for WaveSpeed image generation and editing, featuring an in-canvas key flow, explicit spend control, and native `IMAGE` sockets. The nodes are production-tested and safe by default. Key features: - Secure in-canvas key flow - Explicit spend control - Native ComfyUI `IMAGE` sockets - Safe-by-default workflows Tech stack: Python Why it matters: These nodes offer enhanced security and ease of use for ComfyUI users, ensuring a safer and more controlled workflow environment. Repository: JsonMatrixLab/matrix-power-nodes-wave1 GitHub stars: 13 GitHub description: Ten secure WaveSpeed image generation and editing nodes for ComfyUI.

Overview: This project details the successful implementation of Gemma-4-31B with a TurboQuant KV cache on an AMD RDNA4 GPU (gfx1201), achieving full 256K context loading and prefilling at 735 tok/s. The repository includes a measured study showing that this setup uses only 22.9 GB VRAM and 9 GB free, with a crash-free decode performance of ~22 tok/s at low context and ~9.4 tok/s at 128K context. Key features: - TurboQuant KV cache - HIP-graph-safe Flash-Attention for llama.cpp - Fully measured on real hardware Tech stack: Python Why it matters: This project showcases a novel approach to running large language models on GPUs, specifically targeting the AMD RDNA4 architecture, which significantly reduces the memory footprint and improves performance compared to previous setups.

Overview: This comprehensive toolkit provides a wide range of tools and resources for building AI-powered applications, covering model selection, deployment, and everything in between. Key features: - OpenAI API - Grok (xAI) - Google Gemini - Hugging Face - Anthropic Claude - DeepSeek - Together AI - OpenRouter - Mistral AI - LLaMA (Meta) - AWS Bedrock - NVIDIA NeMo - FastAPI - Flask Tech stack: Python Why it matters: This toolkit offers a diverse set of tools and resources to support the entire AI development process, from model selection to deployment, ensuring that developers can create high-quality AI applications efficiently.

### Overview: The MCP Vision Bridge is a text-only LLM coding agent that provides visual insights into images through multimodal models. It integrates seamlessly with various Claude Code, Codex, opencode, Kimi, PI, and other compatible agents, enhancing their capabilities to analyze and debug code more effectively. ### Key Features: - **Image Analysis**: Routes images through a chosen multimodal model to provide detailed textual descriptions. - **Versatile Models**: Works with Claude Code, Codex, opencode, Kimi, PI, and any MCP client. - **Auto-Loop Hook**: Automatically updates the MCP server and skill on startup, ensuring the latest version is available. - **No Image Input**: The agent only needs to provide the image file path, which is automatically inserted into the agent's context. ### Tech Stack: - Node.js ≥ 18 - TypeScript ### Why It Matters: - **Enhanced Debugging**: Provides a detailed textual description of images, enabling agents to debug and fix issues more effectively. - **Versatility**: Works with multiple compatible agents and models, enhancing the overall functionality of the agent. - **Efficiency**: No need to manually transcribe or paste images, making the process more efficient and error-free. ### Repository: - **Repository**: KuaaMU/mcp-vision-bridge - **GitHub Stars**: 13 - **GitHub Description**: MCP server that gives text-only LLM coding agents vision — analyze images via any multimodal model (mimo, Claude, Gemini, OpenAI-compatible). Works with Claude Code,