Star 总榜

Repositório da Speech2Learning, uma arquitetura baseada em reconhecimento de fala para promover a acessibilidade de Objetos de Aprendizagem.

ANDROID APP to AUTO GENERATE SUBTITLE FILE and TRANSLATED SUBTITLE FILE (using unofficial online Google Translate API) for any audio/video files using 2 ACTIVITIES

Speech to text bot for Discord using Mozilla's DeepSpeech

Don't write a bug report — record it. Local MCP: screen recording → transcript, frames, OCR, wall-clock evidence for coding agents.

Study Guide for the AI-102: Designing and Implementing a Microsoft Azure AI Solution Exam

Open-source iOS voice dictation keyboard — fully offline, private, no subscription required for core features.

Free, open-source Whisper desktop app — 100% offline speech-to-text for Windows, macOS & Linux. No cloud, no subscription.

DeepSeek-V3, R1 671B on 8xH100 Throughput Benchmarks

An open-source, model-agnostic agent harness for local LLMs. Define agents in YAML (tools, memory, deny-first permissions) and run them against any OpenAI-compatible endpoint: vLLM, Ollama, LM Studio, or llama.cpp.

A fine-tuning SDK — any open model, any harness, any method. 12 training methods behind one argument; pure-stdlib core.

This repository contains the source code and related resources for R-LoRA.

[EMNLP22] Improving Sharpness-Aware Minimization with Fisher Mask for Better Generalization on Language Models

T̶̘̊h̷̙͘į̸̀ș̷͌ ̴̳̀r̴̬̕e̷̬͐p̵͍̚o̵̧̎s̶̗͂i̷͚̿t̷̟͝õ̴͙ř̵̘y̷̛̪ ̴̮͌i̶͊͜s̴̠̊ ̴̼͗f̶͘͜i̵̮͊n̴̨̊e̶̖̍!̷̝͋ ̴̨͛T̷̐͜h̷̺̔e̶̩̍r̸̰͒é̶̥ ̸̻̇ȉ̶͍s̵̡̍ ̴̛̫n̶̼̓ọ̷̀t̸̊ͅh̵̙͑ĩ̶͚n̵͙̋g̴̫̃ ̸̼͊w̷̘̿r̶̩̓o̷̠͝n̷͉͌g̶̞͒ ̷̛̼ọ̸̓v̶͍̈́e̵̺͑r̸̻̄ ̴̲̀h̸̉ͅé̶͙r̷̻̾e̷̠͛.̸̨̌

PLMFit platform for TL on PLMs

A Streamlit-powered machine learning playground that automatically detects classification or regression tasks, builds pipelines with preprocessing, trains models interactively, and visualizes metrics using Plotly. Backward-compatible, fully responsive, and deployable on Streamlit Cloud or Docker.

Using langchain framework created some useful Ai applications.

基于YOLOV8的人脸检测模型，从huggingface上面下载，通过gradio进行部署，实现输入图片的URL或者上传图片，模型检测出所有的人脸，并返回位置信息

ContextQA - The open-source tool for data-driven conversations

Safetensors dump and load in Elixir

Fine-tune the newly released Llama-3.2 lightweight models.