# PaddlePaddle/PaddleSlim

PaddleSlim is an open-source library for deep model compression and architecture search.

Repository: https://github.com/PaddlePaddle/PaddleSlim
Canonical: https://ross.abutalabs.com/products/paddleslim
Homepage: https://paddleslim.readthedocs.io/zh_CN/latest/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: pruning, quantization, nas, bert, compression, detection, distillation, ernie, segmentation, sparsity, tensorrt, transformer, yolov6, yolov5, yolov7
Last push: 2026-01-04T09:30:21+00:00

## Health v2 (maintenance only)
Score: 50/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 60, release rhythm 8, longevity 100
- inputs: {"age_days": 2452, "days_push": 241, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1612, forks 353 (observed 2026-08-28T04:05:11.094325+00:00)

## What it is
PaddleSlim is an open-source library built on PaddlePaddle for deep learning model compression and architecture search. It provides low-bit quantization, knowledge distillation, sparsity/pruning, and neural architecture search (NAS) strategies, including automated compression pipelines for models like YOLO and BERT.

## Use cases
- compress a yolov5 model with int8 quantization for faster inference
- prune a detection model to reduce its size
- apply knowledge distillation to train a smaller student model
- run neural architecture search for a segmentation model
- quantize a bert transformer for deployment on tensorrt
- automatically compress an onnx model without changing training code

## When to choose
- you use PaddlePaddle and need to shrink models for edge or faster inference
- you want automated quantization (PTQ/QAT) for detection, segmentation, or NLP models
- you need pruning, distillation, or NAS in one toolkit

## When to avoid
- your models are in PyTorch or TensorFlow and you don't want to convert to Paddle
- you need general-purpose training rather than compression
- you need a framework-agnostic compression tool

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, llm-training
- domain: deep-learning, computer-vision, machine-learning
- platform: python, windows
- tags: model-compression, quantization, pruning, knowledge-distillation, neural-architecture-search, paddlepaddle, tensorrt, int8, yolo, natural-language-processing, linux, macos, gpu

## Member repositories
- PaddlePaddle/PaddleSlim (main) score 50

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:11.094325+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 2026-08-30T03:51:08.604443+00:00, confidence not recorded.
  - readme: https://github.com/PaddlePaddle/PaddleSlim (fetched 2026-08-28T04:05:11.094325+00:00, sha 1b51a0f89012)
- Data as of 2026-08-30T08:39:29.467469+00:00.