Packages
Provides a catalog of pre-defined Pydantic schemas for extracting structured data from images, videos, and documents using Vision Language Models.
Python client library for the Needle API, enabling document indexing, collection management, and semantic search to power retrieval-augmented generation (RAG) pipelines.
However, the aging maintenance status (no updates in months) and modest adoption (45 repository stars) suggest you should verify that Needle's API remains actively…
Connects LangChain applications to SambaNova's language models and embeddings through a unified interface supporting both SambaCloud and SambaStack deployments.
Install it if SambaNova is your chosen model provider; skip it if you use a different LLM backend.
Provides tools to load, parse, and analyze the nuScenes autonomous driving dataset, including 3D object detection, tracking, prediction, and semantic segmentation of lidar and camera data.
However, verify that the non-commercial license restriction aligns with your use case before committing to production work.
fhlmi provides a unified async Python interface to multiple large language models, handling authentication, rate limiting, cost tracking, and tool calling across different LLM providers.
However, verify that all runtime dependencies fit your environment before committing; the dependency footprint is substantial.
Provides ATIF schema models for the NVIDIA NeMo Agent Toolkit, enabling structured representation and validation of agent evaluation and interaction data.
Zenpyre provides utility functions and helpers for working with LangChain, a framework for building applications with language models and agents.
However, it is in alpha (0.0.1a64), so the API is unstable and may change.
Provides a Python client to query text generation models hosted on Hugging Face Inference Endpoints or the Hugging Face Hub, with support for both synchronous and asynchronous generation and token streaming.
anima-python provides a command-line interface to the anima consciousness engine for measurement, evaluation, serialization, and corpus generation, runnable with only numpy on systems without the hexa toolchain.
However, verify the license status before production use, and confirm the package is available on PyPI (documentation indicates it may still be awaiting publication).
Basic Pitch converts audio files to MIDI transcriptions with pitch bend detection using a lightweight neural network, supporting polyphonic and instrument-agnostic note detection.
However, the package is dormant (last release 728 days ago), so verify that the license is compatible with your use case (it is currently unclear in the metadata) and…
whylogs generates statistical profiles of datasets to track data distributions, detect drift, and validate data quality across machine learning pipelines.
However, the 619-day gap since the last release means no active support for new Python versions, dependency updates, or bug fixes—verify compatibility with your…
Generates deterministic Python code that enforces business policy constraints on AI agent tool invocations before they execute, preventing policy violations at the tool level.
Provides evaluation components and CLI commands for the NVIDIA NeMo Agent Toolkit, enabling assessment of agent performance.
Install only if you need evaluation capabilities; it is a specialized subpackage.
Integrates Ollama's local embedding models with LlamaIndex for generating text embeddings without external API calls.
Install it if you're already running Ollama and want seamless vector embedding support; skip it if you rely on cloud embedding APIs.
Provides pre-trained transformer protein language models (ESM-2, ESMFold, ESM-1v, MSA Transformer, ESM-IF1) for protein structure prediction, embedding generation, variant effect prediction, and inverse folding directly from sequence.
However, the repository is archived and unmaintained since February 2024, so there will be no bug fixes or compatibility updates.
Quantizes LiteRT models to reduce size and improve inference performance on edge devices, supporting dynamic, weight-only, and static quantization strategies with configurable recipes.
Converts text to phonemes for Vietnamese, Thai, and Indonesian with English code-switching support, using a memory-mapped binary dictionary and Rust-based engine for fast batch processing.
Python SDK for the VLM Run API platform, providing access to vision-language models for image and document processing, chat completions, and CLI-based agent interactions.
CUGA is an enterprise-grade generalist agent framework that orchestrates complex tasks across web interfaces and APIs using configurable reasoning modes, integrated tools (OpenAPI, MCP, LangChain), and policy-driven behavior controls.
Converts PyTorch models to .tflite format for on-device deployment on Android, iOS, and IoT devices via LiteRT, with CPU coverage and initial GPU and NPU support.
Provides a collection of alternative optimization algorithms for PyTorch that are compatible with the standard optim module interface.
Lightly provides self-supervised learning models and loss functions for computer vision, enabling you to train neural networks on unlabeled image data using methods like MoCo, SimCLR, BYOL, and DINO.
Install it if you need to train on unlabeled image data or experiment with SSL methods.
MLServer is an open-source inference server that exposes machine learning models through REST and gRPC APIs compliant with the KFServing V2 Dataplane specification, supporting multi-model serving, adaptive batching, and parallel inference.
However, the aging maintenance status warrants checking whether security patches and bug fixes align with your production requirements.
Treelite-runtime provides a Python runtime for loading and executing decision tree forest models that have been serialized in the Treelite format, enabling fast inference on tree-based models.
A toolkit for building and deploying AI agents on DataRobot, providing unified LLM routing, agentic tools, multi-framework support (LangGraph, LlamaIndex, CrewAI, NIM), and orchestration through a low-code workflow interface.
KServe Python SDK provides server and client libraries for deploying and managing machine learning model inference services, with built-in support for model loading from multiple storage backends and standardized prediction/preprocessing handlers.
torchax is a PyTorch backend that runs PyTorch code on Google Cloud TPUs via JAX, with bidirectional interoperability allowing you to call JAX functions from PyTorch and vice versa.
However, it is in alpha (Development Status :: 3), so expect incomplete operation coverage and potential API changes.
Kumo SDK provides a Python interface to programmatically interact with the Kumo machine learning platform, enabling users to build and manage ML workflows on modern data infrastructure.
Cleanlab TLM scores the trustworthiness of LLM responses in real-time, detecting hallucinations and incorrect outputs by evaluating every response against the original prompt.
However, the aging maintenance signal (only 24 stars, no recent activity beyond the last commit date) and reliance on an external API key and service mean you should…
Provides a collection of modern optimizers, learning rate schedulers, and loss functions for PyTorch training with a consistent API.
Install it if you want to experiment with modern training methods or need access to research optimizer variants without implementing them yourself.
Provides the database backend for FiftyOne, a computer vision framework for managing and analyzing image and video datasets.
BoxMOT provides pluggable multi-object tracking modules that work with bounding box detections from any model, supporting both axis-aligned and oriented bounding boxes through a unified Python API and CLI.
sgl-kernel provides optimized CUDA compute primitives for LLM inference, enabling efficient execution of custom kernel operations for large language models and vision-language models.
Loads and runs pretrained EfficientNet image classification models in PyTorch, supporting both standard and adversarially-trained variants for inference and feature extraction.
NeMo Evaluator runs standardized benchmarks against language models through a configurable framework with built-in scoring, proxy interception, and multi-format result reporting.
Phidata is a framework for building multi-modal agents that can use tools, maintain memory, and reason to accomplish tasks—either individually or as coordinated teams.
However, the dormant maintenance status (564 days since last release) means you should verify compatibility with your Python version and dependencies before…
Cua Agent provides a framework for building agentic workflows that interact with computers through AI models, supporting macOS, Windows, and Linux sandboxes with liteLLM integration.
However, verify the license terms before use—the unclear license treatment is a blocker for some contexts.
Serializes and deserializes PyTorch modules and tensors to/from HTTP, HTTPS, S3, Redis, and local filesystem endpoints with streaming support for fast model loading.
Automates IBM Quantum backend connection, authentication, and selection by detecting your subscription plan and returning the least-busy available QPU resource for immediate use in Qiskit circuits.
However, note the aging maintenance status (219 days since last release) and verify that the connector's error handling and backend-selection logic align with your…
Ranger is a PyTorch optimizer that combines RAdam (Rectified Adam) and LookAhead techniques to improve training stability and convergence in deep learning models.
No—install only if you are reproducing historical research or maintaining legacy code.