dyNET38
Fork version of DyNet: DyNet38 shares wheels of DyNet for Python 3.8+
Decision gist · record as of 2026-08-14
Yes, if you are building NLP models with dynamic structures and prefer a C++-backed library with Python bindings. The permissive Apache 2.0 license and broad platform support (Linux, macOS, Windows) are favorable. However, the aging maintenance status (934 days since last release) and small repository star count (4) suggest limited active development and community; consider your long-term support needs.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires a C++ compiler and CMake to build from source; prebuilt wheels are available for Python 3.8–3.12 on Linux, macOS (x86_64 and ARM64), and Windows.
- Medium install friction due to compiled wheels for multiple platforms and Python versions (3.8–3.12).
- Package is aging—last release was 934 days ago—but the repository remains active with recent commits.
License · maintenance · safety
Apache 2.0 (permissive) — Licensed under Apache 2.0 (permissive), so you can use it freely in commercial and open-source projects without copyleft obligations.
last release 2024-01-23 (934 days) · last repo commit 2025-12-30 · 4 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 802,791 downloads/mo, #5,019 on PyPI
Alternatives
Verify before relying
pip install dynet38
from dynet38 import Model
model = Model()
W = model.add_parameters((10, 5))- Whether prebuilt wheels include GPU support (CUDA/cuDNN) or CPU-only computation.
- Current maintenance status and whether dynet38 is actively maintained separate from the original DyNet project.
- Specific performance characteristics or benchmarks for auto-batching on modern hardware.
- Whether the fork dynet38 maintains API compatibility with the original DyNet.
What it is and what it does
DyNet38 is a fork of DyNet that provides Python bindings to a C++ neural network library optimized for dynamic computation graphs—networks whose structure changes per training instance. This is particularly valuable in natural language processing, where sentence length, parse trees, and other structural properties vary. The library supports both CPU and GPU execution and includes an auto-batching feature that automatically groups computations for efficiency without requiring manual batching logic.
The package is built on top of cython and numpy, and ships as precompiled wheels for Python 3.8 through 3.12 across Linux, macOS, and Windows. It has been used to build state-of-the-art systems for syntactic parsing, machine translation, and morphological inflection. The aging maintenance status (last release 934 days ago) suggests it is stable but not under active development.
Use it for
- Build syntactic parsers or dependency parsers where tree structure varies per sentence.
- Implement sequence-to-sequence models for machine translation with dynamic graph construction.
- Train morphological inflection systems where input and output structure depend on linguistic properties.
- Prototype NLP models that benefit from auto-batching without manual minibatch management.
- Run neural network experiments on GPU or CPU with a single codebase.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, if you are building NLP models with dynamic structures and prefer a C++-backed library with Python bindings.
The permissive Apache 2.0 license and broad platform support (Linux, macOS, Windows) are favorable. However, the aging maintenance status (934 days since last release) and small repository star count (4) suggest limited active development and community; consider your long-term support needs.
Install
dynet38 on PyPI
Before you install
Medium install friction due to compiled wheels for multiple platforms and Python versions (3.8–3.12). Package is aging—last release was 934 days ago—but the repository remains active with recent commits. Depends on cython and numpy.
Requires a C++ compiler and CMake to build from source; prebuilt wheels are available for Python 3.8–3.12 on Linux, macOS (x86_64 and ARM64), and Windows.
License in practice
Licensed under Apache 2.0 (permissive), so you can use it freely in commercial and open-source projects without copyleft obligations.
Quickstart
pip install dynet38
from dynet38 import Model
model = Model()
W = model.add_parameters((10, 5))Verify before relying
- Whether prebuilt wheels include GPU support (CUDA/cuDNN) or CPU-only computation.
- Current maintenance status and whether dynet38 is actively maintained separate from the original DyNet project.
- Specific performance characteristics or benchmarks for auto-batching on modern hardware.
- Whether the fork dynet38 maintains API compatibility with the original DyNet.
Package facts
| License | Apache 2.0 permissive |
| Python support | Not specified |
| Install friction | Medium. Platform-specific wheel |
| Runtime dependencies | 2 packagescythonnumpy |
| Maintenance | Aging 934 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 802,791 / month, #5,019 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | Development Status :: 5 - Production/StableEnvironment :: ConsoleIntended Audience :: DevelopersIntended Audience :: Science/ResearchLicense :: OSI Approved :: Apache Software LicenseOperating System :: POSIXOperating System :: POSIX :: LinuxProgramming Language :: PythonProgramming Language :: Python :: 3.10Programming Language :: Python :: 3.11Programming Language :: Python :: 3.12Programming Language :: Python :: 3.8Programming Language :: Python :: 3.9Topic :: Scientific/Engineering :: Artificial Intelligence |
Evidence: dynet38-2.2-cp310-cp310-macosx_10_9_x86_64.whl; dyNET38-2.2-cp310-cp310-macosx_10_9_x86_64.whl; dyNET38-2.2-cp310-cp310-macosx_11_0_arm64.whl; dyNET38-2.2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl; dyNET38-2.2-cp310-cp310-manylinux_2_17_i686.manylinux2014_i686.whl; dyNET38-2.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl; dyNET38-2.2-cp310-cp310-musllinux_1_1_i686.whl; dyNET38-2.2-cp310-cp310-musllinux_1_1_x86_64.whl; dyNET38-2.2-cp310-cp310-musllinux_1_2_aarch64.whl; dynet38-2.2-cp310-cp310-win_amd64.whl; dynet38-2.2-cp311-cp311-macosx_10_9_x86_64.whl; dyNET38-2.2-cp311-cp311-macosx_10_9_x86_64.whl; dyNET38-2.2-cp311-cp311-macosx_11_0_arm64.whl; dyNET38-2.2-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl; dyNET38-2.2-cp311-cp311-manylinux_2_17_i686.manylinux2014_i686.whl; dyNET38-2.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl; dyNET38-2.2-cp311-cp311-musllinux_1_1_i686.whl; dyNET38-2.2-cp311-cp311-musllinux_1_1_x86_64.whl; dyNET38-2.2-cp311-cp311-musllinux_1_2_aarch64.whl; dynet38-2.2-cp311-cp311-win_amd64.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “dynamic neural networks”
- dyNET38DyNet38 is a Python binding for DyNet, a C++ neural network library…
- torch-geometricPyTorch Geometric is a library for building and training Graph Neural…
- jmpJMP provides mixed precision training support for JAX by managing…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also tensorrt-cu13 · nvidia-cudnn-cu13 · nvidia-cudnn-cu12 · tensorrt-cu13-bindings · tensorrt-cu12 · tensorrt-cu12-bindings · tensorrt-cu13-libs · dlib · torch · depthai