Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Lightning Talk @ Chainer Meetup

Sponsored · Your Podcast. Everywhere. Effortlessly. Share. Educate. Inspire. Entertain. You do you. We'll handle the rest. →
Avatar for Sinhrks Sinhrks
October 14, 2015
Programming
53k
0

Lightning Talk @ Chainer Meetup

Avatar for Sinhrks

Sinhrks

October 14, 2015

More Decks by Sinhrks

See All by Sinhrks
daskperiment: Reproducibility for Humans
Avatar for Sinhrks sinhrks
1
450
PythonとApache Arrow
Avatar for Sinhrks sinhrks
6
2k
大規模データの機械学習におけるDaskの活用
Avatar for Sinhrks sinhrks
10
3.3k
機械学習と解釈可能性
Avatar for Sinhrks sinhrks
7
5.8k
LIME
Avatar for Sinhrks sinhrks
2
1.5k
データ分析言語R 1年の振り返り
Avatar for Sinhrks sinhrks
5
2.6k
pandasでのOSS活動事例と最初の一歩
Avatar for Sinhrks sinhrks
2
20k
Dask Distributedによる分散機械学習
Avatar for Sinhrks sinhrks
4
1.6k
Data processing using pandas and Dask
Avatar for Sinhrks sinhrks
1
300

Other Decks in Programming

See All in Programming
為什麼你並不需要ViewModel / No, you don't need a ViewModel
Avatar for Elvis Shi lovee
1
450
複数の Claude Code が"放置"されてしまう問題をCLI ダッシュボードを自作して解決した話
Avatar for sumihiro3 sumihiro3
0
570
SREの積み重ねがAI駆動開発のガードレールになった ― 7つの実践/SRE Guardrails The 7
Avatar for TomoyaKitaura tomoyakitaura
9
5.8k
霧の中の代数的エフェクト
Avatar for funnycat funnyycat
1
450
AI時代のPHPer生存戦略 ～「言語、もうなんでもよくない？」に本気で向き合う～
Avatar for viviON vivion
0
210
型も通る、synthも通る、それでも危ない 〜AIのCDKの権限とコストを機械で検証する〜 / It Passes Type Checks, It Passes Synth Checks, but It’s Still Risky — Automatically Verifying Permissions and Costs in AI’s CDK —
Avatar for shiro seike seike460
PRO
1
500
なぜ関数型プログラミングで「型」と「証明」が語られるのか #fp_matsuri
Avatar for Takuma Kajikawa kajitack
3
1.1k
Go言語とトイモデルで学ぶTransformerの気持ち / fukuokago23-transformer
Avatar for monochromegane monochromegane
0
150
言語を使う側から、作る側へ。 自作 Lisp で得た新たな気づき。
Avatar for ANDPAD inc andpad
0
140
【やさしく解説 設計編・中級 #4】ルールの寿命と、システムの年輪
Avatar for HideyukiKitao panda728
PRO
2
180
ルールを書いて終わらせないハーネスエンジニアリング
Avatar for Yuji Yamaguchi yug1224
4
1.8k
地域 SRE コミュニティ最前線 - ホンマでっかSRE勉強会
Avatar for tk3fftk tk3fftk
0
290

Featured

See All Featured
The AI Revolution Will Not Be Monopolized: How open-source beats economies of scale, even for LLMs
Avatar for Ines Montani inesmontani
PRO
3
3.7k
Testing 201, or: Great Expectations
Avatar for Joseph Mastey jmmastey
46
8.2k
Beyond borders and beyond the search box: How to win the global "messy middle" with AI-driven SEO
Avatar for David Carrasco davidcarrasco
3
190
Between Models and Reality
Avatar for mayunak mayunak
4
380
Building Flexible Design Systems
Avatar for Yesenia Perez-Cruz yeseniaperezcruz
330
40k
Lightning Talk: Beautiful Slides for Beginners
Avatar for Ines Montani inesmontani
PRO
2
620
Digital Ethics as a Driver of Design Innovation
Avatar for Per Axbom axbom
PRO
1
350
Stewardship and Sustainability of Urban and Community Forests
Avatar for Eric Wiseman pwiseman
0
400
Visualizing Your Data: Incorporating Mongo into Loggly Infrastructure
Avatar for mongodb mongodb
49
10k
Hiding What from Whom? A Critical Review of the History of Programming languages for Music
Avatar for Tomoya Matsuura tomoyanonymous
3
1.1k
Design and Strategy: How to Deal with People Who Don’t "Get" Design
Avatar for Morgane Peng morganepeng
133
19k
The Director’s Chair: Orchestrating AI for Truly Effective Learning
Avatar for Mike Taylor tmiket
1
230

Transcript

  1. Lightning Talk@ Chainer Meetup by sinhrks

  2. ࣗݾ঺հ • ϝʔΧʔʹͯσʔλ෼ੳ • OSS ׆ಈ: • PyData Development Team

    (pandasίϛολ) • Blaze Development Team (Daskίϛολ) • NumPyαϒηοτͷฒྻ෼ࢄndarrayͷ࣮૷͋Γ • GitHub: https://github.com/sinhrks

  3. Chainer ͷར༻৔໘ • ֶࣗࣗश • ۀ຿Ͱͷ Deep Learning ׆༻͸ͳ͠

  4. Chainer ͷྑ͍ͱ͜Ζ • ڧ͍ํ͕࡞͍ͬͯΔ • ϨΠϠఆ͕ٛ௚઀తͰ໌ྎ • ಺෦ͷ࣮૷͕ಡΈ΍͍͢

  5. Chainer ͷվળ఺ • ֶशશମͷॲཧΛ౎౓هड़͢Δඞཁ͕͋Δ • ҎԼͷΑ͏ͳఆ൪ॲཧͷͨΊͷ؆қه๏(ؔ਺/Ϋ ϥε)͕ཉ͍͠ • ॳճͷΈάϥϑग़ྗ •

    epoch ͝ͱͷଛࣦ • early stopping • ՄࢹԽ

  6. Chainer ͷ ՄࢹԽπʔϧ࡞Γ͍ͨ

  7. ֓ཁ • ֶशͷϞχλϦϯά • Computational Graph • ॏΈߦྻ (ະணख) •

    ͳͲΛɺϦϞʔτ (Jupyter / Webϒϥ΢β) ͔Β ֬ೝ͍ͨ͠ • →Bokeh

  8. ֶशͷϞχλϦϯά • ͪ͜ΒͷϒϩάΤϯτϦͷ಺༰ import chainer_vis.visualizer as visualizer iv = visualizer.JupyterVisualizer()

    monitor = iv.get_classification_monitor() for … monitor.update(test_loss=loss, test_accuracy=acc)