Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Char-rnn aurkezpena
Search
Manex Agirrezabal
March 14, 2016
Research
0
78
Char-rnn aurkezpena
Manex Agirrezabal
March 14, 2016
Tweet
Share
More Decks by Manex Agirrezabal
See All by Manex Agirrezabal
The Flipped Classroom model for teaching Conditional Random Fields in an NLP course
manexagirrezabal
0
26
NLP for poetry generation and analysis
manexagirrezabal
0
63
Institut seminar 2020
manexagirrezabal
0
28
Automatic Scansion of Poetry (KU)
manexagirrezabal
0
480
RANLP talk
manexagirrezabal
0
70
Defense (Final version)
manexagirrezabal
0
68
Poesiaren eskantsio automatikoa: Bi hizkuntzen azterketa
manexagirrezabal
0
67
CodeFEST literature presentation
manexagirrezabal
0
57
Ongoing work (in mid 2016)
manexagirrezabal
0
20
Other Decks in Research
See All in Research
LiDARとカメラのセンサーフュージョンによる点群からのノイズ除去
kentaitakura
0
120
20241115都市交通決起集会 趣旨説明・熊本事例紹介
trafficbrain
0
180
Isotropy, Clusters, and Classifiers
hpprc
3
630
機械学習でヒトの行動を変える
hiromu1996
1
290
文化が形作る音楽推薦の消費と、その逆
kuri8ive
0
160
Weekly AI Agents News! 9月号 プロダクト/ニュースのアーカイブ
masatoto
2
130
メールからの名刺情報抽出におけるLLM活用 / Use of LLM in extracting business card information from e-mails
sansan_randd
2
130
論文紹介/Expectations over Unspoken Alternatives Predict Pragmatic Inferences
chemical_tree
1
260
Human-Informed Machine Learning Models and Interactions
hiromu1996
2
450
[ECCV2024読み会] 衛星画像からの地上画像生成
elith
1
630
授業評価アンケートのテキストマイニング
langstat
1
360
Generative Predictive Model for Autonomous Driving 第61回 コンピュータビジョン勉強会@関東 (後編)
kentosasaki
0
210
Featured
See All Featured
Facilitating Awesome Meetings
lara
50
6.1k
Rebuilding a faster, lazier Slack
samanthasiow
79
8.7k
Practical Tips for Bootstrapping Information Extraction Pipelines
honnibal
PRO
10
710
Practical Orchestrator
shlominoach
186
10k
Intergalactic Javascript Robots from Outer Space
tanoku
269
27k
How STYLIGHT went responsive
nonsquared
95
5.2k
ReactJS: Keep Simple. Everything can be a component!
pedronauck
665
120k
5 minutes of I Can Smell Your CMS
philhawksworth
202
19k
Mobile First: as difficult as doing things right
swwweet
222
8.9k
Unsuck your backbone
ammeep
668
57k
Testing 201, or: Great Expectations
jmmastey
38
7.1k
Fantastic passwords and where to find them - at NoRuKo
philnash
50
2.9k
Transcript
Poesiaren metrika DL bidez Manex Agirrezabal https://github.com/manexagirrezabal/char-rnn/
Proba ezberdinak TensorFlow: Sequence-to-sequence models https://www.tensorflow.org/versions/master/tutorials/seq2seq/index.html Torch: char-rnn (Andrew Karpathy)
https://github.com/karpathy/char-rnn/
Char-rnn http://karpathy.github.io/2015/05/21/rnn-effectiveness/ Karaktere mailako hizkuntz-ereduak sortzeko balio du. Sarrera gisa
testu hutsa.
Char-rnn Gure beharretarako moldatu behar: to swell the gourd and
plump the ha zel shells - ' - ' - ' - ' - ' wo man much missed how you call to me call to me ' - - ' - - ' - - ' - -
Char-rnn Dataset-a testu soil gisa: To_= swell_+ the_= gourd_+ and_=
plump_+ the_= ha_+ zel_= shells_+ To_= swell_+ the_= gourd_+ and_= plump_+ the_= hazel_+= shells_+ Wo_+ man_= much_= missed_+ how_= you_= call_+ to_= me_= call_+ to_= me_= Woman_+= much_= missed_+ how_= you_= call_+ to_= me_= call_+ to_= me_=
Char-rnn (training) $ th train.lua Parametroak: Model: [RNN, LSTM edo
GRU] rnn_size: LSTMaren (zelda) barruko tamaina num_layers: LSTMaren kapa kopurua seq_length: sekuentzian ikasteko karaktere kopurua
Char-rnn (prediction) $ th sample(mod).lua Parametroak: Model: eredu entrenatua Primetext:
sarrera testua (_ karakterearekin amaituta)
Char-rnn (prediction) Python programa bat (callSampleMod.py) aurreko programari deitzeko pausuz
pausu: $ th sampleMod.lua model M1 primetext “to_” = $ th sampleMod.lua model M1 primetext “to_= swell_” + $ th sampleMod.lua model M1 primetext “to_= swell_+ the_” = ...
Char-rnn (prediction) Arazoa: Hasieran, informazio gutxi duenez, batzuetan hanka sartzen
(+ propagatzen) du predikzioan. Adibidez, “to_” sarrerarekin Horrentzako soluzioa, predikzioa bi aldetara egitea.
Char-rnn (FW) Parametroak optimizatu nahi ditugu (seq_length, batch_size, rnn_size, ...)
Embedding-ak erabili nahi ditugu, baina gure hipotesia da ez dutela asko lagunduko.