Packages
Decomposes scikit-learn decision tree and random forest predictions into bias and per-feature contribution components, enabling interpretation of how each feature drives individual predictions.
However, do not use it if you depend on active maintenance or need compatibility with very recent versions—test thoroughly before deploying to production, as the…
Torch Model Archiver creates .mar archive files from trained neural network models for deployment with TorchServe inference serving.
However, approach with caution for new projects: the repository is archived and maintenance is abandoned, so security patches and compatibility updates with future…
Omnigent is an open-source orchestration layer that lets you run, supervise, and collaborate on multiple AI agents (Claude, Codex, Cursor, custom YAML-defined agents) from terminal, browser, or phone, with policy enforcement and cloud sandbox support.
Provides a JAX-compatible wrapper around Python dataclasses that automatically registers them as pytrees and enables serialization, while supporting static fields and nested mutations.
Validates and loads datasets described in the MLCommons Croissant JSON-LD format, converting metadata into Python representations for static and dynamic analysis.
Generates hyperparameter sweep suggestions using Bayesian optimization, grid search, and other methods, and identifies runs to early-terminate based on performance.
Install it if you need programmatic hyperparameter optimization or are integrating with Weights & Biases sweeps locally.
SemHash deduplicates, filters outliers from, and finds representative samples in text, image, and multimodal datasets using semantic similarity with fast embeddings and efficient nearest-neighbor search.
PyTabKit provides scikit-learn interfaces to modern tabular machine learning models—including RealMLP, gradient boosted trees (XGB, LGBM, CatBoost), and TabR—with built-in hyperparameter optimization, cross-validation ensembling, and benchmarking utilities.
However, if your goal is production-grade best performance, the documentation explicitly recommends AutoGluon instead.
Provides core runtime utilities, environment configuration, and authentication scope resolution for AI agent applications built on Microsoft Power Platform.
However, this is an alpha package—expect the API to evolve—and you should consult the official Microsoft documentation before integrating it into production systems.
OnnxTR extracts and recognizes text from documents (PDFs, images, webpages) using ONNX-based deep learning models, localizing and identifying words without requiring PyTorch or TensorFlow.
Provides a unified runtime kernel and trust layer for building, executing, and governing multi-agent systems with audit trails, privilege isolation, and identity management.
LLaMA Factory provides a unified framework for fine-tuning 100+ large language models through CLI and web UI, supporting multiple training methods and model architectures with minimal setup.
Displays a Keras-style summary of PyTorch neural network models, showing layer types, output shapes, and parameter counts without requiring manual inspection of the model structure.
Recognizes and extracts entities like numbers, units, dates, and times from text in multiple languages, returning structured resolution data for each match.
However, it is alpha-stage and has not been updated since 2019; use it for production only if you can accept the maintenance risk and have validated that entity…
A Gradio component that displays two images side-by-side with an interactive slider for comparing them, supporting image upload, generation, or static display.
However, verify compatibility with your Gradio version before deploying, since the package has not been updated since 2024-04-19 and may lag behind Gradio's evolution.
Recognizes and extracts numeric entities (cardinals, ordinals, percentages, ranges) from text in multiple languages, with support for English, Chinese, French, Spanish, Portuguese, German, Italian, Turkish, Hindi, and Dutch.
However, verify that the specific language and entity types you need are supported, and be aware that the package itself is not actively maintained.
A PySpark connector for Amazon SageMaker FeatureStore that ingests data from Spark DataFrames into FeatureStore's online and offline stores, with automatic feature definition loading.
However, proceed with caution: maintenance is dormant (554 days since last release), so compatibility with recent SageMaker or Spark updates is uncertain.
BDDL is a domain-specific language for defining household activity benchmarks as logical planning problems, specifying initial object states, conditions, and goal states for simulation-based behavior evaluation.
Install only if you are actively contributing to or maintaining a fork, or integrating with an existing BEHAVIOR simulator project that requires this exact version.
gsplat is a CUDA-accelerated Python library for rasterizing 3D Gaussians, implementing differentiable rendering based on 3D Gaussian Splatting with support for multiple camera models and optimization strategies.
However, verify the license terms in the repository first, and confirm your PyTorch and CUDA versions are supported by pre-compiled wheels or compatible with JIT…
Recognizes and extracts numbers with units (age, currency, dimensions, temperature) from text in multiple languages, resolving them to standardized values.
However, it is in alpha status with no releases since November 2019—verify that the recognition accuracy and language coverage meet your use case before committing to…
VoxCPM2 is a tokenizer-free text-to-speech system that generates multilingual speech directly via diffusion-autoregressive architecture, supporting voice design, controllable voice cloning, and 48kHz audio output across 30 languages.
PyDSSP assigns protein secondary structure (alpha-helix, beta-strand, or loop) from backbone coordinates using a simplified DSSP algorithm, with support for both NumPy and differentiable PyTorch tensors.
MABWiser implements multi-armed bandit algorithms for decision-making under uncertainty, supporting context-free, parametric, and non-parametric contextual models with built-in parallelization for training and testing.
However, dormant maintenance means no active bug fixes or feature development—suitable for research and learning, less ideal if you need ongoing support or expect…
Recognizes and resolves date and time entities from text in multiple languages, extracting temporal expressions and normalizing them to structured date/time values.
However, it is in alpha status and has not been released since 2019, which may indicate incomplete language support or unresolved issues.
Quantizes large language models to lower bit-widths using the GPTQ algorithm, reducing model size and memory requirements while enabling faster inference on consumer and enterprise GPUs.
However, install with caution: the repository is archived (no updates since March 2024), so you will not receive compatibility fixes if torch, transformers, or CUDA…
Integrates Azure AI Search as a vector store backend for LlamaIndex, enabling semantic search and retrieval-augmented generation workflows using Azure's managed search service.
Detects text and mathematical formulas in images using pre-trained deep learning models, supporting Chinese, English, and multiple languages with both PyTorch and ONNX backends.
PySR searches for symbolic expressions that fit data by combining genetic algorithms with Julia's performance, returning human-readable equations instead of black-box models.
Not recommended if you need black-box performance on high-dimensional datasets or require guaranteed convergence to a global optimum.
Netron is a viewer for neural network and machine learning models that supports ONNX, TensorFlow Lite, PyTorch, TensorFlow, Core ML, OpenVINO, Keras, Caffe, Darknet, Safetensors, NumPy, and several other formats.
Install it if you work with neural networks or machine learning models in any of the supported formats.
Mamba-ssm implements state-space model architectures (Mamba, Mamba-2, Mamba-3) as alternatives to Transformers for sequence modeling, optimized for language tasks and information-dense data.
Dopamine is a research framework for prototyping reinforcement learning algorithms, providing implementations of DQN, C51, Rainbow, IQN, SAC, and PPO agents built primarily with JAX.
Provides OpenTelemetry-based telemetry, tracing, and monitoring for AI agents, with structured spans for agent invocation, tool execution, and LLM inference.
Recognizes and resolves entities like numbers, units, and date/time expressions in text across multiple languages, including full support for Chinese, English, French, Spanish, Portuguese, German, Italian, Turkish, Hindi, and Dutch.
However, its alpha status and lack of recent releases (latest from November 2019) suggest it may not be actively evolving.
Runs an MCP server that exposes Google Analytics Admin and Data APIs as tools for LLMs, enabling queries about accounts, properties, reports, and real-time analytics data.
However, it requires Google Cloud setup (enabling APIs, configuring credentials) and is not a standalone library—it only works as an MCP server within a compatible…
Enables ONNX Runtime to accelerate machine learning model inference on Intel hardware (CPUs, integrated/discrete GPUs, and NPUs) using OpenVINO optimizations.
PraisonAI is a framework for building and deploying multi-agent AI systems that can autonomously research, plan, and execute tasks. It abstracts away boilerplate by letting you define agents with instructions and goals, then orchestrate them across multiple LLM providers.
Bridges Composio's tool ecosystem to OpenAI's function calling APIs, enabling LLM-driven tool execution through both the Responses API and Chat Completions.
Provides an OpenAPI-generated Python client for the OGX API, offering interfaces to Llama model endpoints including chat completions, embeddings, batches, and file operations.
However, it requires Python 3.12+, is in beta status (1.1.0.dev0), and mixes stable and experimental APIs—verify that the specific endpoints you need are…
Downloads and processes Objaverse-XL, an open dataset of over 10 million 3D objects, with utilities for managing large-scale 3D model data retrieval and preparation.
However, be aware that maintenance is dormant (last update November 2023), so expect no bug fixes or compatibility updates.
DRUM is a command-line tool for developing, testing, and deploying custom machine learning models written in Python, R, or Java before uploading them to DataRobot's platform.