Research notes and extra resources for all the work at explodinggradients.com

🚀 MiksaGPT, part of the 'Miksa' project, is a groundbreaking voice assistant utilizing Claude 3 and APIs from 'anthropic' and 'elevenlabs'. It enables real-time Opus two-way voice chat with seamless interruptibility, built with Flutter and available for free on GitHub.

The Strongest Open-Sourced AI Reasoning Model Architecture.

A fun collection of notebooks for finetuning AI models... Share ready-to-run notebooks… tips & tricks for finetuning... Hugging Face Transformers, Unsloth, vLLM, PyTorch / CUDA magic

基于Gin构建的土拨鼠社区基础服务

Anima Portable Standalone Trainer—a completely self-contained, browser-based UI for fine-tuning the cutting-edge Anima diffusion architecture (DiT + Qwen3 Text Encoder). Say goodbye to dependency hell and complex command lines; simply run the installer, launch the sleek web interface, and start training your LoRAs / Model Finetune

Enter a topic and the AI will search the web, scan YouTube, and synthesize a comprehensive research report — powered by Gemini, Google Search, and YouTube APIs.

An AI-powered clinical assistant using Retrieval-Augmented Generation (RAG) on the MIMIC-IV DiReCT dataset. It retrieves relevant patient cases and generates diagnostic reasoning using LLMs. Built with Streamlit, Transformers, FAISS, and SentenceTransformers.

This repository provides a merged LoRA fine-tuned version of Mistral-7B, specialized for health insurance Q&A and domain-specific chat tasks. The LoRA weights are already integrated into the base model, making it usable as a standalone model for instruction-following and conversational use cases.

An open-source conversational AI assistant for rural & semi-urban India. Supports voice-based queries in Hindi, Punjabi, Tamil, Bengali and more. Converts speech to text (ASR), retrieves answers via RAG, translates between Indic ↔ English, and responds back with Indic TTS. Built with AI4Bharat models, Whisper, Flan-T5, IndicTrans2, and IndicLID.

ConversaVault_AI is a multi-user conversational AI backend built with FastAPI, Ollama, and PostgreSQL, designed for persistent session memory, secure user-level conversation isolation, and scalable local LLM orchestration.