Packages
MMEngine is a foundational PyTorch training library that provides a configurable runner, metrics system, and integration with distributed training frameworks to streamline deep learning model development.
Install it if you want a structured, configurable training abstraction; skip it if you prefer writing training loops directly.
Optimizes GPU memory usage for running large generative models on consumer GPUs with limited VRAM by managing model loading, quantization, and offloading.
However, verify the unclear license status before use in proprietary work, and test your specific model and hardware combination first.
mmh3 is a Python binding to MurmurHash3, a fast non-cryptographic hash function that returns 32-bit or 128-bit hash values for bytes and strings, with optional seeding.
mmhash3 is a Python wrapper for MurmurHash3, providing fast non-cryptographic hash functions for strings and byte data with configurable output sizes (32-bit, 64-bit, or 128-bit) and seed values.
However, the package is abandoned—no updates since 3.0.1—so consider it stable and feature-complete rather than actively maintained.
Decodes and encodes the macromolecular transmission format (MMTF), a binary encoding for biological structures, and provides a Python API to fetch and inspect PDB structure data.
However, be aware that it will not receive updates for Python version incompatibilities or format changes.
MNE-Python provides tools for exploring, visualizing, and analyzing human neurophysiological data including MEG, EEG, sEEG, and ECoG recordings, with modules for preprocessing, source estimation, time-frequency analysis, connectivity, and machine learning.
Install it if you work with MEG, EEG, or related brain recordings.
MNE-BIDS reads and writes neuroimaging datasets in BIDS format, enabling standardized organization and analysis of MEG, EEG, and iEEG data with MNE-Python.
Generates mnemonic word sequences and converts them to cryptographic seeds following the Bitcoin standard for deterministic wallet creation.
However, dormant maintenance (no updates in 952 days) means no active support for bugs or security issues—acceptable for a narrow, mature specification but worth…
MNN is a lightweight deep learning inference and training framework that runs neural network models efficiently on mobile devices, embedded systems, and servers with support for multiple hardware backends including CPU, GPU, and NPU.
Provides null-safe dot-notation access to nested dictionaries and JSON-like data structures, replacing standard dicts with a `Data` object that supports both attribute and path-based access without raising errors on missing keys.
Provides a flat-namespace compatibility layer for Python 2/3 code migration, consolidating commonly needed imports into a single `mo_future` module to avoid discovering them across multiple sub-modules.
However, the aging maintenance status (425 days since last release) and small user base (2 GitHub stars) suggest this is a niche tool—if you are starting a new…
Provides patterns to resolve circular module dependencies in Python through delayed imports and expect/export mechanisms, keeping imports at the top of files while breaking cycles.
Provides an `@override` decorator that lets you pass function parameters as a dictionary via a `kwargs` argument, ignoring extra keys and allowing parameter overrides without raising errors.
Provides structured JSON logging with integrated exception handling that tracks calling context automatically, allowing you to log messages with named parameters and chain exceptions without creating custom exception types.
mo-parsing is a PEG parser generator that lets you define grammars using Python operators and predefined patterns, then parse strings into structured parse trees accessible as lists or dictionaries.
Parses SQL queries into JSON-serializable parse trees, and can generate SQL back from JSON, supporting multiple database flavors including MySQL, SQLServer, and BigQuery.
Mobly is a Python test framework designed to orchestrate test cases involving multiple devices, complex environments, or custom hardware setups, with built-in support for Android device control.
mobsfscan is a static analysis tool that detects insecure code patterns in Android and iOS source code, supporting Java, Kotlin, Swift, Objective-C, and Android XML files.
Install it if you develop mobile apps and want automated pattern-based vulnerability detection integrated into your workflow.
mock provides a backport of Python's standard library unittest.mock for Python 3.6 and later, allowing you to replace parts of your system under test with mock objects and assert how they were used.
Provides mock helpers for SQLAlchemy sessions in unit tests, enabling assertion on query calls and filter operations without accessing a real database.
However, be aware that maintenance is dormant—last release was 2023-03-26—so compatibility with very recent SQLAlchemy versions is uncertain.
Provides an in-memory mock of the Google Cloud Firestore Python client for use in unit tests, supporting collections, documents, queries, transactions, and field transforms without requiring a live database.
Provides a drop-in replacement for mock.mock_open that supports multiple files, persistent file state across open calls, and standard file operations like read, write, seek, and tell.
Provides a Python context manager that runs an in-process SSH server for testing, built on paramiko, without requiring OpenSSH binaries.
Mocket is a socket mocking framework that intercepts and replaces socket and SSL module calls, enabling you to test socket-based clients without real network traffic.
Mockito is a spying and stubbing framework for Python that lets you mock out dependencies, verify interactions, and stub side effects in unit tests.
Provides an in-process SFTP server for testing SFTP client code, with fixtures for easy integration into test suites.
However, the abandoned status (no updates since 2018-03-21) means you should test compatibility first and be prepared to maintain a fork if critical issues arise.
Modal is a Python client library that lets you run Python functions on serverless cloud compute infrastructure directly from your local scripts, handling deployment, scaling, and execution on Modal's platform.
This package is a deprecated compatibility shim that redirects users to install the `modal` package instead; it is no longer functional.
Mode is a minimal asyncio library that lets you build programs as composable services with lifecycle callbacks (start, stop, restart) and background tasks, organized into a dependency graph.
Model Archiver creates `.mar` archive files from trained neural network models for use with MXNet-Model-Server inference.
Model Bakery creates test fixtures for Django models automatically, generating realistic data for object instances with a single line of code.
Compresses and optimizes neural networks through quantization, pruning, and hardware-aware techniques, reducing model size and computational cost for edge deployment.
Install it if you need to compress trained models for deployment; the main constraint is the Python >=3.10 requirement and the need to provide a pre-trained model as…
Provides standardized FastAPI-based handlers and decorators for deploying ML models (vLLM, TensorRT-LLM) to Amazon SageMaker with unified `/ping` and `/invocations` endpoints.
Maintains a structured index of machine learning model metadata (parameters, results, training data, paper links) stored as YAML, JSON, or markdown annotations, with optional integration to Papers with Code.
Model Mommy generates Django model instances with random data for testing, automating the creation of test fixtures without manual field specification.
Model2Vec converts sentence transformers into small, fast static embedding models that generate vector representations of text for tasks like retrieval, classification, and clustering.
Reads and writes mmCIF and BinaryCIF files conforming to the ModelCIF extension, with optional BinaryCIF support via msgpack.
Install it if you work with mmCIF or BinaryCIF files in structural biology; skip it if you have no need for these formats.
Provides a typed Python interface to models.dev, a community-maintained catalog of 2000+ LLM models from 75+ providers, with pricing, context limits, and capability metadata—no HTTP calls needed.
ModelScope provides unified Python interfaces for inference, fine-tuning, and evaluation across machine learning models spanning NLP, computer vision, speech, multi-modal, and scientific computing domains.
Download, upload, and manage AI models, datasets, and other assets on ModelScope Hub through a unified Python SDK and CLI interface.