Tools to Fine-Tune and Train AI Models Locally

LiteRT Torch is a Python library for converting PyTorch models into .tflite files for the LiteRT runtime. Developers can prepare models on Linux and deploy them into Android, iOS and IoT applications that run inference on the device. Code is licensed under Apache 2.0.

DeepTrack2 is an open-source Python library for researchers building image data pipelines for machine learning and digital microscopy. It brings image generation, manipulation and analysis into a modular framework that runs in your own Python environment. Its focus is scientific imaging, with applications in microscopy, optics and biophysics.

sd-scripts is a collection of Python scripts for people who want to train image models on their own hardware. It suits users comfortable working with scripts who need to adapt a model to their own images or subjects.

Reef is self-hosted infrastructure for developers who want AI agents to learn from their interactions. It connects live requests and feedback to changes in model weights or the prompts, rules, and skills that guide an agent.

Cleanlab is a Python library for data scientists and ML engineers checking training data for errors. It uses model predictions or feature embeddings to identify problems in text, image, audio and tabular datasets.

TensorBoard helps ML developers and researchers inspect TensorFlow experiments through a browser. It shows how training progresses and lets you compare runs when evaluating changes to a model or its hyperparameters.

Second Me is an open-source prototype for people who want to train and host an AI representation of themselves. It uses personal memories to shape a model around their identity and context.

Opacus is a Python library for training PyTorch models with differential privacy. It's for machine learning practitioners adapting existing training code and researchers experimenting with private training methods.

SecretFlow is an open-source Python framework for teams that need to analyze data or train machine learning models across separately held datasets. It brings several privacy-preserving computing methods into one framework, so applications can combine them according to their security assumptions and performance needs.

CellProfiler Analyst helps biologists explore measurements from image-based experiments and train machine learning classifiers to recognize cell phenotypes. It runs on your own machine and works with data imported from CellProfiler, with a focus on large, high-dimensional datasets from high-throughput experiments.

openWakeWord is a wake word detection library for developers building voice apps that listen for a chosen word or phrase. It lets an application detect voice triggers on its own hardware.

LatentSync is an open-source AI lip-sync tool for video creators and researchers who want a face's mouth movements to match supplied audio. It processes video on your own GPU, with demos covering filmed faces and anime characters.

Applio converts recordings and microphone input into another voice for musicians making AI covers, streamers and speech creators. You can choose community voice models or train a model from your own recordings.

Doccano is a text annotation tool for people preparing training or evaluation datasets for machine learning. Its browser interface lets teams label text together on their own machine or server.

Musubi Tuner provides training scripts for people who want to adapt image and video generation models with their own datasets. It's aimed at users comfortable with a script-based workflow.

PaddleNLP is a Python toolkit for developers training and deploying language models on their own infrastructure. Built on PaddlePaddle, it covers pretraining, fine-tuning, model compression and inference across several accelerator families.

MeloTTS is a Python text-to-speech library for developers adding spoken output to their applications. It runs locally and can generate speech in real time on a CPU, so a dedicated GPU isn't required for inference.

Hugging Face Datasets is a Python library for developers and researchers preparing data for AI training and evaluation. It handles local files and datasets from the Hugging Face Hub through the same data-processing tools.

Synthetic Data Vault is a Python library for teams that need synthetic versions of their tabular datasets. It learns patterns from real records and generates data that emulates them. Training and generation run on your own hardware with standard CPUs.

SpeechBrain is a toolkit for researchers and developers building speech assistants, transcription systems and conversational AI. It provides pretrained models and training recipes for adapting speech, audio and text models to your own data.

RamaLama runs AI models for local chat or a self-hosted REST API. It's for developers who want a container workflow for trying models on their own hardware and serving them in production.

Stable Diffusion Core ML is a developer toolkit for adding local AI image generation to Mac, iPhone, and iPad apps. It pairs model conversion tools with a Swift package for embedding generation in native software.

InterpretML is an open-source Python library for data scientists who need to explain machine learning predictions or build models whose decisions they can inspect directly. It runs locally on Windows, macOS and Linux under the MIT license. Its tools support model debugging and checks for potential bias, including work in healthcare and finance.

FlagEmbedding is a Python toolkit for developers building semantic search and retrieval-augmented generation (RAG). It runs embedding models to find relevant text and rerankers to refine the order of retrieved results.

UNO is a local AI image generation framework for researchers and creators who want to preserve subjects from reference images in generated scenes. One model handles both single-subject images and compositions with several subjects.

MLflow helps developers and teams inspect AI applications, compare changes and track model experiments. It brings agent debugging and evaluation into a platform you can run locally or on your own servers.

NVIDIA cuML is a Python machine learning library for data scientists whose scikit-learn, UMAP or HDBSCAN workloads take too long on a CPU. It runs on NVIDIA GPUs in Linux or WSL 2, on a workstation or across a cluster. The project is open source under Apache 2.0.

ControlNet adds visual guidance to Stable Diffusion for artists and developers who need generated images to follow a reference's layout or shape. It lets a sketch, pose or scene map guide the result alongside a text prompt.

DeepLabCut is an open-source toolbox for researchers who want to track animal movement and behavior on their own computers without attaching physical markers. It estimates the positions of visible features you choose, with support for animals, humans and objects rather than a fixed set of species or body parts.

MediaPipe is a toolkit for developers adding on-device AI to applications. It provides pretrained models and task APIs for applications that need to process images, video, text or audio on the user's device.

TensorRT Model Optimizer (ModelOpt) is a Python library for developers who need smaller models and faster inference. It accepts Hugging Face, PyTorch and ONNX models and lets developers combine optimization techniques before deployment.

ONNX Runtime is an AI inference and training engine for developers adding models to apps and services. It supports deployment on a user's device or a server, including applications that need the same model format across platforms.

Augmentoolkit turns documents into datasets for training an LLM on a particular subject. It's for people building custom models around material such as research papers or fictional lore.

AReaL is open-source reinforcement learning infrastructure for researchers and developers training LLMs and AI agents on their own compute. Its asynchronous training approach targets large reasoning models and agent applications, with support for single-node runs and distributed training. It uses the Apache 2.0 license.