Language Models

A demo repo for running extreme-quantized language models locally without needing a research cluster.

It breaks complex tasks across a team of specialized LLM agents that refine their own skills as they work.

Xime is a deliberately minimal, Rime-based Android input method that serves as its author's personal testbed for on-device AI experiments in predictive text and speech recognition.

It turns your local machine into an OpenAI-compatible inference endpoint so agents and IDEs can run on offline models without reconfiguration.

Uses multimodal LLMs to transcribe PDFs into Markdown, preserving complex layouts that traditional extractors mangle.

Curated technical deep-dives covering everything from NVLink signal integrity to Kubernetes GPU scheduling and Huawei NPU porting.

It split off from `verl` to give diffusion, video, and omni-modality models an RL post-training framework that doesn't treat them like chatbots.

Bytez wraps 175,000+ AI models behind a single endpoint so you don't have to host them yourself.

HRM-Text claims to cut pretraining costs by 130–600× compute and 150–900× data, shipping a full 1B-parameter framework with FSDP2, FlashAttention 3, and a hierarchical recurrent architecture.

OpenMythos is an independent attempt to reconstruct Anthropic’s rumored Claude Mythos architecture as a trainable Recurrent-Depth Transformer with switchable attention and sparse MoE layers.

A modular speech-to-speech pipeline that exposes an OpenAI Realtime-compatible WebSocket API so you can run voice agents on local or open-source models instead of proprietary cloud services.

Apfel surfaces Apple’s built-in FoundationModels as a pipe-friendly UNIX tool and OpenAI-compatible local server, no API keys required.

This Go CLI turns a single sentence into a full novel by making Architect, Writer, and Editor LLM agents plan, draft, and review inside a long-loop state machine—no human hand-holding required.

A thin Python wrapper that lets you swap GPT-4o for Claude or Gemini without touching your code.

It corrals the latest subquadratic sequence-model research into hardware-efficient, training-ready PyTorch layers verified across NVIDIA, AMD, and Intel GPUs.

DataFlex stops LLM training loops from wasting compute on static data mixes by dynamically selecting, mixing, and reweighting samples inside LLaMA-Factory.

NexaSDK is a local inference engine that squeezes frontier LLMs and vision models onto Qualcomm silicon through NPU, GPU, and CPU backends.

Petals lets you run and fine-tune models like Llama 3.1 405B from a desktop by distributing layers across a public swarm of consumer GPUs.

A PyTorch implementation of "Attention Is All You Need" that scales from 13M to multi-billion parameter models.

MiniMind-O packs listen-see-speak intelligence into a 0.1B-parameter model you can retrain from the first line of code on a single desktop GPU.