English | 简体中文 A unified, high-performance framework for training LLMs, VLMs, diffusion, and embodied models. 🌐 Website · 📖 Docs · ✍️ Blog · ⚡ Quick Start · 📊 Performance · 🏛️ Supported Models · 💬 Contact LoongForge is a unified training framework for LLMs, VLMs, diffusion, and embodied models, covering pre-training, continued pre-training, and SFT.
[ Read the Docs ] 日本語 | 中文简体 | 中文繁體 --- Code and data for the following works: SWE-bench is a benchmark for evaluating large language models on real world software issues collected from GitHub. Given a codebase and an issue, a language model is tasked with generating a patch that resolves the described problem.
For embeddings, please see examples/embeddings/index.html WebGPU support is introduced via PR #215. Upon updating to V3.1, WebGPU will be enabled automatically.
[comment]: <> (Add badges for coverage when we have tests, update repo for other types of badges!) aequitas is an open-source bias auditing and Fair ML toolkit for data scientists, machine learning researchers, and policymakers. We provide an easy-to-use and transparent tool for auditing predictors of ML models, as well as experimenting with "correcting biased model" using Fair ML methods in binary classification settings.
Evidently An open-source framework to evaluate, test and monitor ML and LLM-powered systems. Documentation Evidently is an open-source Python library to evaluate, test, and monitor ML and LLM systems—from experiments to production.
⚠️ Warning: Not Actively Maintained This repository is no longer actively maintained. For a more up-to-date alternative, please visit the LitGPT project: https://github.com/Lightning-AI/litgpt , which serves as the successor to this repository.
A Rust, Python and gRPC server for text generation inference. Used in production at Hugging Face to power Hugging Chat, the Inference API and Inference Endpoints.
In a nutshell, we aim to generate polyphonic music of multiple tracks (instruments). The proposed models are able to generate music either from scratch, or by accompanying a track given a priori by the user.
This is a port of BlinkDL/RWKV-LM to ggerganov/ggml. Besides the usual FP32, it supports FP16, quantized INT4, INT5 and INT8 inference.
The LLM ecosystem has amazing infrastructure (LoRAX, PEFT, vLLM), but lacks standardized, high-quality capability adapters. Problem: Base models limited to 32K context, need 2M tokens for large repositories Solution: Progressive curriculum learning with vLLM + Unsloth hybrid approach Key Innovation: Hybrid optimization combining vLLM's inference speed with Unsloth's training efficiency - achieving 61x context extension with minimal compute!
LLM Finetuning toolkit is a config-based CLI tool for launching a series of LLM fine-tuning experiments on your data and gathering their results. From one single yaml config file, control all elements of a typical experimentation pipeline - prompts, open-source LLMs, optimization strategy and LLM testing.
This is the official repository for ICLR 2025 paper "Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing". Magpie generates high-quality alignment data by prompting aligned LLMs with their pre-query templates.
KoboldCpp is an easy-to-use AI text-generation software for GGML and GGUF models, inspired by the original KoboldAI. It's a single self-contained distributable that builds off llama.cpp and adds many additional powerful features.
We also introduce a diffusion-based value model that reduces variance and improves stability during optimization. Based on TraceRL, we derive a series of diffusion language models, TraDo, which achieve state-of-the-art performance on math and coding reasoning tasks.
RouteLLM is a framework for serving and evaluating LLM routers. [Blog ] [Paper ] Our core features include: From PyPI From source Let's walkthrough replacing an existing OpenAI client to route queries between LLMs instead of using only a single model.
MASFactory is a graph-centric framework for orchestrating Multi-Agent Systems with Vibe Graphing: Start from intent, generate a graph design, preview and refine it in a visual environment, compile it into an executable workflow, and trace node states, messages, and shared state at runtime. Turn natural-language intent into a structural design, then iteratively converge to an executable, reusable workflow.
Fine‑tune, evaluate, and run private, personalized LLMs xTuring makes it simple, fast, and cost‑efficient to fine‑tune open‑source LLMs (e.g., GPT‑OSS, LLaMA/LLaMA 2, Qwen3, MiniMax M2, GPT‑J, GPT‑2, DistilGPT‑2, Mamba) on your own data — locally or in your private cloud. Why xTuring: Run a small, CPU‑friendly example first: Want bigger models and reasoning controls?
I've started to work on reimplementation of the library here: FastTensors Please star it if you'd like to see GGML-compatible implementation in pure Go. Please check out my related project Booster We dream of a world where fellow ML hackers are grokking REALLY BIG GPT models in their homelabs without having GPU clusters consuming a shit tons of $$$.
This repo supports the paper "QLoRA: Efficient Finetuning of Quantized LLMs", an effort to democratize access to LLM research. QLoRA uses bitsandbytes for quantization and is integrated with Hugging Face's PEFT and transformers libraries.
Website • Docs • Community Slack NannyML is an open-source python library that allows you to estimate post-deployment model performance (without access to targets), detect data drift, and intelligently link data drift alerts back to changes in model performance. Built for data scientists, NannyML has an easy-to-use interface, interactive visualizations, is completely model-agnostic and currently supports all tabular use cases, classification and regression.
English | 简体中文 A unified, high-performance framework for training LLMs, VLMs, diffusion, and embodied models. 🌐 Website · 📖 Docs · ✍️ Blog · ⚡ Quick Start · 📊 Performance · 🏛️ Supported Models · 💬 Contact LoongForge is a unified training framework for LLMs, VLMs, diffusion, and embodied models, covering pre-training, continued pre-training, and SFT.
[ Read the Docs ] 日本語 | 中文简体 | 中文繁體 --- Code and data for the following works: SWE-bench is a benchmark for evaluating large language models on real world software issues collected from GitHub. Given a codebase and an issue, a language model is tasked with generating a patch that resolves the described problem.
For embeddings, please see examples/embeddings/index.html WebGPU support is introduced via PR #215. Upon updating to V3.1, WebGPU will be enabled automatically.
[comment]: <> (Add badges for coverage when we have tests, update repo for other types of badges!) aequitas is an open-source bias auditing and Fair ML toolkit for data scientists, machine learning researchers, and policymakers. We provide an easy-to-use and transparent tool for auditing predictors of ML models, as well as experimenting with "correcting biased model" using Fair ML methods in binary classification settings.
Evidently An open-source framework to evaluate, test and monitor ML and LLM-powered systems. Documentation Evidently is an open-source Python library to evaluate, test, and monitor ML and LLM systems—from experiments to production.
⚠️ Warning: Not Actively Maintained This repository is no longer actively maintained. For a more up-to-date alternative, please visit the LitGPT project: https://github.com/Lightning-AI/litgpt , which serves as the successor to this repository.
A Rust, Python and gRPC server for text generation inference. Used in production at Hugging Face to power Hugging Chat, the Inference API and Inference Endpoints.
In a nutshell, we aim to generate polyphonic music of multiple tracks (instruments). The proposed models are able to generate music either from scratch, or by accompanying a track given a priori by the user.
This is a port of BlinkDL/RWKV-LM to ggerganov/ggml. Besides the usual FP32, it supports FP16, quantized INT4, INT5 and INT8 inference.
The LLM ecosystem has amazing infrastructure (LoRAX, PEFT, vLLM), but lacks standardized, high-quality capability adapters. Problem: Base models limited to 32K context, need 2M tokens for large repositories Solution: Progressive curriculum learning with vLLM + Unsloth hybrid approach Key Innovation: Hybrid optimization combining vLLM's inference speed with Unsloth's training efficiency - achieving 61x context extension with minimal compute!
LLM Finetuning toolkit is a config-based CLI tool for launching a series of LLM fine-tuning experiments on your data and gathering their results. From one single yaml config file, control all elements of a typical experimentation pipeline - prompts, open-source LLMs, optimization strategy and LLM testing.
This is the official repository for ICLR 2025 paper "Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing". Magpie generates high-quality alignment data by prompting aligned LLMs with their pre-query templates.
KoboldCpp is an easy-to-use AI text-generation software for GGML and GGUF models, inspired by the original KoboldAI. It's a single self-contained distributable that builds off llama.cpp and adds many additional powerful features.
We also introduce a diffusion-based value model that reduces variance and improves stability during optimization. Based on TraceRL, we derive a series of diffusion language models, TraDo, which achieve state-of-the-art performance on math and coding reasoning tasks.
RouteLLM is a framework for serving and evaluating LLM routers. [Blog ] [Paper ] Our core features include: From PyPI From source Let's walkthrough replacing an existing OpenAI client to route queries between LLMs instead of using only a single model.
MASFactory is a graph-centric framework for orchestrating Multi-Agent Systems with Vibe Graphing: Start from intent, generate a graph design, preview and refine it in a visual environment, compile it into an executable workflow, and trace node states, messages, and shared state at runtime. Turn natural-language intent into a structural design, then iteratively converge to an executable, reusable workflow.
Fine‑tune, evaluate, and run private, personalized LLMs xTuring makes it simple, fast, and cost‑efficient to fine‑tune open‑source LLMs (e.g., GPT‑OSS, LLaMA/LLaMA 2, Qwen3, MiniMax M2, GPT‑J, GPT‑2, DistilGPT‑2, Mamba) on your own data — locally or in your private cloud. Why xTuring: Run a small, CPU‑friendly example first: Want bigger models and reasoning controls?
I've started to work on reimplementation of the library here: FastTensors Please star it if you'd like to see GGML-compatible implementation in pure Go. Please check out my related project Booster We dream of a world where fellow ML hackers are grokking REALLY BIG GPT models in their homelabs without having GPU clusters consuming a shit tons of $$$.
This repo supports the paper "QLoRA: Efficient Finetuning of Quantized LLMs", an effort to democratize access to LLM research. QLoRA uses bitsandbytes for quantization and is integrated with Hugging Face's PEFT and transformers libraries.
Website • Docs • Community Slack NannyML is an open-source python library that allows you to estimate post-deployment model performance (without access to targets), detect data drift, and intelligently link data drift alerts back to changes in model performance. Built for data scientists, NannyML has an easy-to-use interface, interactive visualizations, is completely model-agnostic and currently supports all tabular use cases, classification and regression.