# PetrochukM/PyTorch-NLP

Basic Utilities for PyTorch Natural Language Processing (NLP)

Repository: https://github.com/PetrochukM/PyTorch-NLP
Canonical: https://ross.abutalabs.com/products/pytorch-nlp
Homepage: https://pytorchnlp.readthedocs.io
Language: Python
License: BSD-3-Clause
License Family: permissive
Topics: pytorch, nlp, natural-language-processing, pytorch-nlp, torchnlp, data-loader, embeddings, word-vectors, python, deep-learning, dataset, metrics, neural-network, sru, machine-learning
Archived: true
Last push: 2023-07-04T21:11:26+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 3111, "days_push": 1156, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2222, forks 253 (observed 2026-08-28T04:06:27.672376+00:00)

## What it is
PyTorch-NLP (torchnlp) is a Python library of basic utilities for natural language processing built on PyTorch, providing text data loading, datasets, encoders, word vector embeddings, and metrics. The repository has been archived, with the author recommending Hugging Face and PyTorch toolchain alternatives.

## Use cases
- load NLP datasets like IMDB into PyTorch
- load pretrained word vectors and embeddings in PyTorch
- compute NLP metrics for model evaluation
- build text data processing pipelines for PyTorch models
- tokenize and encode text for neural networks

## When to choose
- you are maintaining a legacy codebase that already depends on torchnlp
- you need a lightweight, self-contained PyTorch NLP utility library and accept it is unmaintained

## When to avoid
- you are starting a new NLP project - use Hugging Face Datasets, Tokenizers, or torchtext instead
- you need actively maintained code, bug fixes, or support for recent PyTorch versions

## Facets
- artifact type: library
- maturity: abandoned
- function: machine-learning, nlp, data-science, serialization
- domain: machine-learning, deep-learning
- platform: python
- tags: pytorch, torchnlp, text-processing, word-vectors, embeddings, datasets, archived, natural-language-processing

## Member repositories
- PetrochukM/PyTorch-NLP (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:27.672376+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:45:45.045191+00:00, confidence not recorded.
  - readme: https://github.com/PetrochukM/PyTorch-NLP (fetched 2026-08-28T04:06:27.672376+00:00, sha 8a0edb3e53cf)
- Data as of 2026-08-30T08:39:29.467469+00:00.