Python 排行榜

Thinking-native compiler for LEET Language Understanding item sets — killer-item design grammar, verification separation, deterministic gate chain, 3-model benchmark

Hikari's curated English measured notes on local LLM systems (serving, specdec, training, steering).

Turn any NVIDIA GPU into a local AI platform. Inference + fine-tuning in your browser. One command to start, automatic clustering.

TokenPowerBench: Benchmarking the power consumption of LLM inference across engines and deployment configurations

Automated Triton w8a8 block FP8 kernel tuning tool for vLLM. Auto-detects model architecture, supports Qwen3-Coder-30B-A3B-Instruct-FP8/DeepSeek-V3/custom models, multi-GPU parallel tuning, and generates optimized kernel configs for quantization.

A curated collection of Claude Code agent skills that accelerate the entire vLLM development lifecycle.

⚡️ The "1-Minute RAG Audit" — Generate QA datasets & evaluate RAG systems in Colab, Jupyter, or CLI. Privacy-first, async, visual reports.

A simplified fine tune and deploy code based on bert for text matching.

Fine-tune a local LLM on your own app's data in 15 minutes. Runs entirely on-device, zero cloud after training. Apple Silicon + CUDA.

🏠 OpenStreetMap, AI import tool for buildings in Poland

Measuring generalization properties of graph neural networks

LLM finetuning for Sudoku solving

JGLUE: Japanese General Language Understanding Evaluation for huggingface datasets

Customer Conversational Intelligence Platform Powered by an LLM Agent

Chat with your documents using Generative AI & Retrieval-Augmented Generation (RAG)

Know your VRAM before you run. Instant GPU memory estimates for any HuggingFace model.

Local-first RAG platform — Ollama + Qdrant + Redis + MinIO. No cloud, no API keys, runs entirely on your machine.

A repository dedicated to exploring, experimenting with, and applying Generative AI concepts and tools in practical, everyday scenarios.

An optimized FastAPI server for OpenAI's Whisper whisper-large-v3-turbo model using MLX optimization

We introduce a large semi-automatically generated dataset of ~400,000 descriptive sentences about commonsense knowledge that can be true or false in which negation is present in about 2/3 of the corpus in different forms that we use to evaluate LLMs