ToolAI.io · Канал GitHub
Открывайте недавно добавленные на GitHub проекты AI с открытым исходным кодом: сведения о репозитории, заметки по настройке, лицензии и связанные ресурсы.
Данные публичного репозитория
microsoft/agent-framework
An open-source framework for building, orchestrating, and evaluating AI agents across Python and .NET.
huggingface/smolagents
A lightweight Python library for building agents that use tools and execute code to solve tasks.
CrewAIInc/crewAI
A Python framework for coordinating role-based AI agents and structured task workflows.
dottxt-ai/outlines
A library for generating structured, constrained outputs from language models.
stanfordnlp/dspy
A framework for programming and optimizing language-model pipelines instead of hand-tuning prompts.
huggingface/trl
A library for post-training transformer models with supervised fine-tuning and preference-optimization methods.
typesafe-ai/typesafe-sdk-python
Connect Python services to the TypeSafe API using typed questions and results. The repository provides the client library rather than model weights.
NVIDIA/warp
Compiles typed Python functions into compute kernels with geometry, physics and differentiable programming capabilities for simulation and machine learning.
zai-org/GLM-OCR
An OCR project for tables, formulas and complex layouts, with hosted and self-hosted pipelines, a CLI and a Python SDK.
PaddlePaddle/PaddleSpeech
A PaddlePaddle-based speech toolkit covering speech recognition, text-to-speech, punctuation restoration, and other audio tasks, with command-line, Python interface, and service examples.
google-deepmind/mujoco
A multijoint contact dynamics engine maintained by Google DeepMind, providing model description, physics stepping, an interactive viewer, and a Python interface. It is suitable for developing robotics, control, and reinforcement learning environments.
hydra-ecosystem/hydra
A configuration framework for Python applications that manages experiments through configuration composition, command-line overrides, and multirun execution. It is suitable for organizing data, model, and training parameters into reusable configuration structures.
Страница 1 / 7 · 76 проектов
