microsoft/nni

Does nni contain some new pruning base methods？

Open

#4,567 建立於 2022年2月21日

在 GitHub 查看
 (6 留言) (0 反應) (1 負責人)Python (13,504 star) (1,830 fork)batch import
good first issuemodel compressionnew feature

描述

Describe the issue: Sorry to bother you, in nni, the current basic pruning strategies are Norm, FPGM, ActivationPruner, TaylorFOWeightPruner, etc. Are there other basic strategies, such as HRank, if I want to add it myself, where do I need to rewrite it?

Environment:

  • NNI version: V2.6
  • Training service (local|remote|pai|aml|etc): local
  • Client OS: linux
  • Server OS (for remote mode only):
  • Python version: 3.8
  • PyTorch/TensorFlow version: pytorch1.8
  • Is conda/virtualenv/venv used?: conda
  • Is running in Docker?: no

How to reproduce it?:

貢獻者指南

技術棧
pythonpytorch
領域
machine learningtooling
議題類型
research
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
1-3 hours
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
stale
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
mostly clear
前置要求
Familiarity with NNI pruning moduleBasic PyTorch knowledge
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
30
研究方向
The issue asks about adding a new pruning method (e.g., HRank) to NNI. The user wants to know which base class to extend and where to implement it. Refer to the abstract pruner in nni/compression/pytorch/pruning/basic pruner.py and the existing pruners like NormPruner. The research should explore how to create a new pruner class that inherits from the base and registers it. Also check the latest NNI documentation and examples for custom pruners.