Iteration Plan for release in Mar

[scarlett2018]

Extending this short iteration to a regular iteration for Feb-Mar.
Release Manager: @liuzhe-lz
Code freeze date: 3/2 3/26 (Wed) 3/27
End Game Issue: #2132 2132
Target release date: 3/6 3/31 (Tue)
Release Note: https://github.com/microsoft/nni/releases/tag/v1.5

---

Planned Items

Code completed

DLTS/DLWorkspace integration

• Integration with DLWS PR DLTS integration #1945 @QuanluZhang, @Gerhut;

  • support pause status @chicm-ms PR Add PAUSED status for trial jobs (#2089) #2090 (review @liuzhe-lz @QuanluZhang )
  • support pause status on webui @lvybriage (pending)
  • fix IT fail @Gerhut
  • current log path @Gerhut

• model compression examples refactor @chicm-ms / reviewer: @Cjkkkk @QuanluZhang PR Refactor pruner examples #2099

  • prepare GPU vm for pipeline (1.14), migrate remote docker to the GPU vm @SparkSnail

• speedup example refactor @QuanluZhang

• refactor the logic of applying mask file @QuanluZhang

• compression framework refactor @Cjkkkk

• validation of configuration passed to compression algorithms @chicm-ms

In progress

NAS

• NAS visualization @ultmaster / Reviewer: @lvybriage @QuanluZhang @SparkSnail PR NAS visualization #2085
• NAS installation @SparkSnail (3/23)

New Algo

• support HPO algo - PBT Add Population Based Training Tuner #1470 @RayMeng8 (draft pr add PBT tuner #2139, debugging, a pr ready in next week)

AutoNLP

• AutoNLP @QuanluZhang
  • search algo @ultmaster
  • retrain / Add evaluation script for TextNAS Add evaluation script for TextNAS #2204 @pkuyym pr Add evalutaion scripts for TextNAS #2220

IT/UT

• add IT/UT @chicm-ms PR Refactor integration tests #2190 need review
  • refactor framework (eta: wed)
  • resume experiment
  • trial uses dict metric
  • experiment view, experiment foreground
  • IT for NAS
  • IT for model speedup
  • replace pai IT with new ones @SparkSnail PR Add pai pipeline #2092 (debugging)
  • IT failure (smac) @chicm-ms

---

Bug Fix

• web ui bug: issue WebUI crash after refresh #2087 @lvybriage PR # 2162 waiting review
• launcher exception: issue nnictl stop cannot stop a failed experiment sometime #2080 PR Handle exception and kill restserver in nnictl #2086 @SparkSnail review: @chicm-ms @QuanluZhang
• webui supports NaN/Infinite @lvybriage PR support show NaN in detail table #2183 review: @chicm-ms @liuzhe-lz

---

##Other candidates

• remove deprecated nas examples @QuanluZhang @ultmaster
• customized homepage for sphinx doc build (before code freeze) @leckie-chn @lvybriage
• PT test @ultmaster (择吉日)

---

Deferred Items

TF 2.0 @liuzhe-lz #2127

• structure codes + ENAS example as first pr, eta 3/18 3/20 NAS for TensorFlow #2115

• Examples - Rewrite TFv1 example trials with TFv2

• NAS SDK and 1-2 examples

• bug - clone fail “Unknown error" for clone trial with MNIST example #1985

• clean issues assigned to each member - 121 issues in total by 2/21

• lottery ticket pruner, add exporting winning ticket @QuanluZhang

[scarlett2018]

left work is tracking in endgame ticket.