SlideShare a Scribd company logo
1 of 19
Download to read offline
What’s New in
PyTorch Land
By Sam Witteveen
@sam_witteveen
About me
• Google Developer Expert for Machine Learning & Deep
Learning
• Deep Learning - Language & Dialogue
• Multiple Startups- B2B, B2C
• Red Dragon AI
• Deep Learning Consulting and Prototyping
• Education & Training
• Products:
• Conversational Computing
• Natural Voice Generation - multiple languages
• Knowledge Base Creation & Reasoning
PyTorch - What’s New
• PyTorch 0.4
• PyTorch Road Map
• ONNX
• Fast.ai
• Skorch
• Others
PyTorch 0.4
• “Tensors” and “Variables” have merged
• Device agnostic code - no need for .cuda()
• Support for scalar Tensors
• Windows support
PyTorch 0.4
Notebook
PyTorch Roadmap
• Path to 1.0 and to Production
• Imperative and declarative
• Easy for research, easy for production
• Folding Caffe2 into PyTorch
• ONNX/Caffe for mobile support
PyTorch Roadmap
• Fused kernels for declarative graphs
• Tools, pre-trained models & Datasets
• ELF (game playing), Glow (compiler), FAISS etc
• Currently in use in Facebook
• Azure & AWS Support
• Coming in the next few months.
ONNX
(Open Neural Network Exchange)
• Converting models to Caffe, CNTK, others
• 0.4 adds RNN support for PyTorch
• TensorFlow support is experimental currently
ONNX - CoreML
https://github.com/onnx/onnx-coreml
ONNX
(Open Neural Network Exchange)
Fast.ai Library
• Wrapper for PyTorch
• Fast.ai - Jeremy Howard & Rachel Thomas
• Adds features beyond PyTorch Vision and Text
• Abstracts a lot of the lower level PyTorch code
• Wraps Pandas, Spacy, TorchVision, TorchText to create
a set of often used data and ML functions
• Focuses on data handling and cutting edge techniques
Fast.ai Library
• date handling, text look ups and embeddings, tabular
data, categorical data etc
• Latest things like Stochastic Weight Averaging
• Very opinionated lib, often well explained in the
courses
• Function names are not always the easiest to decipher
• Documentation ….
Fast.ai Docs
• A scikit-learn compatible neural network library
that wraps pytorch
• Supports 0.4
• Allows you to use SKL pipelines, grid search etc
• https://github.com/dnouri/skorch
Other interesting projects
• Pyro 0.2 - https://github.com/uber/pyro
• Wisdom - https://github.com/facebookresearch/
visdom
• Tensor Board - https://github.com/lanpa/
tensorboard-pytorch
Links
• https://pytorch.org/2018/04/22/0_4_0-migration-
guide.html
• https://pytorch.org/2018/05/02/road-to-1.0.html
• https://github.com/onnx/onnx
• https://github.com/onnx/tensorflow-onnx
• https://github.com/fastai/fastai
• https://github.com/dnouri/skorch
How to Choose?
• Worth learning both
• If your are building for production go with
TensorFlow
• If you are doing research or trying to make
something new work go with PyTorch
The End
Contact
email: s@maclea.ai
twitter: sam_witteveen
https://github.com/samwit/

More Related Content

What's hot

5 NoSQL Options - Toronto - May 2018
5 NoSQL Options - Toronto - May 20185 NoSQL Options - Toronto - May 2018
5 NoSQL Options - Toronto - May 2018Matthew Groves
 
5 Popular Choices for NoSQL on a Microsoft Platform - Tulsa - July 2018
5 Popular Choices for NoSQL on a Microsoft Platform - Tulsa - July 20185 Popular Choices for NoSQL on a Microsoft Platform - Tulsa - July 2018
5 Popular Choices for NoSQL on a Microsoft Platform - Tulsa - July 2018Matthew Groves
 
If You Have The Content, Then Apache Has The Technology!
If You Have The Content, Then Apache Has The Technology!If You Have The Content, Then Apache Has The Technology!
If You Have The Content, Then Apache Has The Technology!gagravarr
 
Lessons from {distributed,remote,virtual} communities and companies
Lessons from {distributed,remote,virtual} communities and companiesLessons from {distributed,remote,virtual} communities and companies
Lessons from {distributed,remote,virtual} communities and companiesColin Charles
 
Automatic publication of library and museum data into the LOD cloud
Automatic publication of library and museum data into the LOD cloudAutomatic publication of library and museum data into the LOD cloud
Automatic publication of library and museum data into the LOD cloudhorvadam
 
Building a knowledge graph of the Belgian War Press
Building a knowledge graph of the Belgian War PressBuilding a knowledge graph of the Belgian War Press
Building a knowledge graph of the Belgian War PressOpen Knowledge Belgium
 
State of the Art Logging. Kibana4Solr is Here!
State of the Art Logging. Kibana4Solr is Here! State of the Art Logging. Kibana4Solr is Here!
State of the Art Logging. Kibana4Solr is Here! lucenerevolution
 
Extending Pandas using Apache Arrow and Numba
Extending Pandas using Apache Arrow and NumbaExtending Pandas using Apache Arrow and Numba
Extending Pandas using Apache Arrow and NumbaUwe Korn
 
What is Hydra?
What is Hydra?What is Hydra?
What is Hydra?Findwise
 
Free and Open Source Software Movement in Libraries of Pakistan
Free and Open Source Software Movement in Libraries of PakistanFree and Open Source Software Movement in Libraries of Pakistan
Free and Open Source Software Movement in Libraries of PakistanAta Rehman
 
NOSQL Databases for the .NET Developer
NOSQL Databases for the .NET DeveloperNOSQL Databases for the .NET Developer
NOSQL Databases for the .NET DeveloperJesus Rodriguez
 
Apache Arrow: Cross-language Development Platform for In-memory Data
Apache Arrow: Cross-language Development Platform for In-memory DataApache Arrow: Cross-language Development Platform for In-memory Data
Apache Arrow: Cross-language Development Platform for In-memory DataWes McKinney
 
PyCon Singapore 2013 Keynote
PyCon Singapore 2013 KeynotePyCon Singapore 2013 Keynote
PyCon Singapore 2013 KeynoteWes McKinney
 
Rust is for Robots!
Rust is for Robots!Rust is for Robots!
Rust is for Robots!Andy Grove
 

What's hot (19)

Find your data
Find your dataFind your data
Find your data
 
Wot’s in a name
Wot’s in a nameWot’s in a name
Wot’s in a name
 
5 NoSQL Options - Toronto - May 2018
5 NoSQL Options - Toronto - May 20185 NoSQL Options - Toronto - May 2018
5 NoSQL Options - Toronto - May 2018
 
Fedora4
Fedora4Fedora4
Fedora4
 
5 Popular Choices for NoSQL on a Microsoft Platform - Tulsa - July 2018
5 Popular Choices for NoSQL on a Microsoft Platform - Tulsa - July 20185 Popular Choices for NoSQL on a Microsoft Platform - Tulsa - July 2018
5 Popular Choices for NoSQL on a Microsoft Platform - Tulsa - July 2018
 
If You Have The Content, Then Apache Has The Technology!
If You Have The Content, Then Apache Has The Technology!If You Have The Content, Then Apache Has The Technology!
If You Have The Content, Then Apache Has The Technology!
 
Lessons from {distributed,remote,virtual} communities and companies
Lessons from {distributed,remote,virtual} communities and companiesLessons from {distributed,remote,virtual} communities and companies
Lessons from {distributed,remote,virtual} communities and companies
 
Automatic publication of library and museum data into the LOD cloud
Automatic publication of library and museum data into the LOD cloudAutomatic publication of library and museum data into the LOD cloud
Automatic publication of library and museum data into the LOD cloud
 
Knowledge graph
Knowledge graphKnowledge graph
Knowledge graph
 
Building a knowledge graph of the Belgian War Press
Building a knowledge graph of the Belgian War PressBuilding a knowledge graph of the Belgian War Press
Building a knowledge graph of the Belgian War Press
 
State of the Art Logging. Kibana4Solr is Here!
State of the Art Logging. Kibana4Solr is Here! State of the Art Logging. Kibana4Solr is Here!
State of the Art Logging. Kibana4Solr is Here!
 
Extending Pandas using Apache Arrow and Numba
Extending Pandas using Apache Arrow and NumbaExtending Pandas using Apache Arrow and Numba
Extending Pandas using Apache Arrow and Numba
 
What is Hydra?
What is Hydra?What is Hydra?
What is Hydra?
 
Free and Open Source Software Movement in Libraries of Pakistan
Free and Open Source Software Movement in Libraries of PakistanFree and Open Source Software Movement in Libraries of Pakistan
Free and Open Source Software Movement in Libraries of Pakistan
 
BPAA PD Day: ONIX 101
BPAA PD Day: ONIX 101BPAA PD Day: ONIX 101
BPAA PD Day: ONIX 101
 
NOSQL Databases for the .NET Developer
NOSQL Databases for the .NET DeveloperNOSQL Databases for the .NET Developer
NOSQL Databases for the .NET Developer
 
Apache Arrow: Cross-language Development Platform for In-memory Data
Apache Arrow: Cross-language Development Platform for In-memory DataApache Arrow: Cross-language Development Platform for In-memory Data
Apache Arrow: Cross-language Development Platform for In-memory Data
 
PyCon Singapore 2013 Keynote
PyCon Singapore 2013 KeynotePyCon Singapore 2013 Keynote
PyCon Singapore 2013 Keynote
 
Rust is for Robots!
Rust is for Robots!Rust is for Robots!
Rust is for Robots!
 

Similar to PyTorch 04 What's New in PyTorch Land

Data Science at Scale: Using Apache Spark for Data Science at Bitly
Data Science at Scale: Using Apache Spark for Data Science at BitlyData Science at Scale: Using Apache Spark for Data Science at Bitly
Data Science at Scale: Using Apache Spark for Data Science at BitlySarah Guido
 
Apache Content Technologies
Apache Content TechnologiesApache Content Technologies
Apache Content Technologiesgagravarr
 
A Jupyter kernel for Scala and Apache Spark.pdf
A Jupyter kernel for Scala and Apache Spark.pdfA Jupyter kernel for Scala and Apache Spark.pdf
A Jupyter kernel for Scala and Apache Spark.pdfLuciano Resende
 
Markup languages and warp-speed documentation
Markup languages and warp-speed documentationMarkup languages and warp-speed documentation
Markup languages and warp-speed documentationLois Patterson
 
Lois Patterson: Markup Languages and Warp-Speed Documentation
Lois Patterson:  Markup Languages and Warp-Speed DocumentationLois Patterson:  Markup Languages and Warp-Speed Documentation
Lois Patterson: Markup Languages and Warp-Speed DocumentationJack Molisani
 
APACHE TOREE: A JUPYTER KERNEL FOR SPARK by Marius van Niekerk
APACHE TOREE: A JUPYTER KERNEL FOR SPARK by Marius van NiekerkAPACHE TOREE: A JUPYTER KERNEL FOR SPARK by Marius van Niekerk
APACHE TOREE: A JUPYTER KERNEL FOR SPARK by Marius van NiekerkSpark Summit
 
Better Tools, Better Mindset
Better Tools, Better MindsetBetter Tools, Better Mindset
Better Tools, Better MindsetYusup
 
An Incomplete Data Tools Landscape for Hackers in 2015
An Incomplete Data Tools Landscape for Hackers in 2015An Incomplete Data Tools Landscape for Hackers in 2015
An Incomplete Data Tools Landscape for Hackers in 2015Wes McKinney
 
3 python packages
3 python packages3 python packages
3 python packagesFEG
 
PyData: The Next Generation | Data Day Texas 2015
PyData: The Next Generation | Data Day Texas 2015PyData: The Next Generation | Data Day Texas 2015
PyData: The Next Generation | Data Day Texas 2015Cloudera, Inc.
 
Introduction to Jupyter notebook and MS Azure Machine Learning Studio
Introduction to Jupyter notebook and MS Azure Machine Learning StudioIntroduction to Jupyter notebook and MS Azure Machine Learning Studio
Introduction to Jupyter notebook and MS Azure Machine Learning StudioMuralidharan Deenathayalan
 
Introduction to Jupyter notebook and MS Azure Machine Learning Studio
Introduction to Jupyter notebook and MS Azure Machine Learning StudioIntroduction to Jupyter notebook and MS Azure Machine Learning Studio
Introduction to Jupyter notebook and MS Azure Machine Learning StudioMuralidharan Deenathayalan
 
Data Science at Scale by Sarah Guido
Data Science at Scale by Sarah GuidoData Science at Scale by Sarah Guido
Data Science at Scale by Sarah GuidoSpark Summit
 
Lessons Learned in the Development of a Web-scale Search Engine: Nutch2 and b...
Lessons Learned in the Development of a Web-scale Search Engine: Nutch2 and b...Lessons Learned in the Development of a Web-scale Search Engine: Nutch2 and b...
Lessons Learned in the Development of a Web-scale Search Engine: Nutch2 and b...Chris Mattmann
 
360|Flex Recap - San Jose 2010
360|Flex Recap - San Jose 2010360|Flex Recap - San Jose 2010
360|Flex Recap - San Jose 2010David Ortinau
 
DevDay: Node Analytics with Python, Chainhaus
DevDay: Node Analytics with Python, ChainhausDevDay: Node Analytics with Python, Chainhaus
DevDay: Node Analytics with Python, ChainhausR3
 
HPCC Systems Engineering Summit: Community Use Case: Because Who Has Time for...
HPCC Systems Engineering Summit: Community Use Case: Because Who Has Time for...HPCC Systems Engineering Summit: Community Use Case: Because Who Has Time for...
HPCC Systems Engineering Summit: Community Use Case: Because Who Has Time for...HPCC Systems
 

Similar to PyTorch 04 What's New in PyTorch Land (20)

Data Science at Scale: Using Apache Spark for Data Science at Bitly
Data Science at Scale: Using Apache Spark for Data Science at BitlyData Science at Scale: Using Apache Spark for Data Science at Bitly
Data Science at Scale: Using Apache Spark for Data Science at Bitly
 
Apache Content Technologies
Apache Content TechnologiesApache Content Technologies
Apache Content Technologies
 
A Jupyter kernel for Scala and Apache Spark.pdf
A Jupyter kernel for Scala and Apache Spark.pdfA Jupyter kernel for Scala and Apache Spark.pdf
A Jupyter kernel for Scala and Apache Spark.pdf
 
Markup languages and warp-speed documentation
Markup languages and warp-speed documentationMarkup languages and warp-speed documentation
Markup languages and warp-speed documentation
 
Lois Patterson: Markup Languages and Warp-Speed Documentation
Lois Patterson:  Markup Languages and Warp-Speed DocumentationLois Patterson:  Markup Languages and Warp-Speed Documentation
Lois Patterson: Markup Languages and Warp-Speed Documentation
 
APACHE TOREE: A JUPYTER KERNEL FOR SPARK by Marius van Niekerk
APACHE TOREE: A JUPYTER KERNEL FOR SPARK by Marius van NiekerkAPACHE TOREE: A JUPYTER KERNEL FOR SPARK by Marius van Niekerk
APACHE TOREE: A JUPYTER KERNEL FOR SPARK by Marius van Niekerk
 
Better Tools, Better Mindset
Better Tools, Better MindsetBetter Tools, Better Mindset
Better Tools, Better Mindset
 
An Incomplete Data Tools Landscape for Hackers in 2015
An Incomplete Data Tools Landscape for Hackers in 2015An Incomplete Data Tools Landscape for Hackers in 2015
An Incomplete Data Tools Landscape for Hackers in 2015
 
3 python packages
3 python packages3 python packages
3 python packages
 
PyData: The Next Generation | Data Day Texas 2015
PyData: The Next Generation | Data Day Texas 2015PyData: The Next Generation | Data Day Texas 2015
PyData: The Next Generation | Data Day Texas 2015
 
Introduction to Jupyter notebook and MS Azure Machine Learning Studio
Introduction to Jupyter notebook and MS Azure Machine Learning StudioIntroduction to Jupyter notebook and MS Azure Machine Learning Studio
Introduction to Jupyter notebook and MS Azure Machine Learning Studio
 
Introduction to Jupyter notebook and MS Azure Machine Learning Studio
Introduction to Jupyter notebook and MS Azure Machine Learning StudioIntroduction to Jupyter notebook and MS Azure Machine Learning Studio
Introduction to Jupyter notebook and MS Azure Machine Learning Studio
 
Decode2018 report
Decode2018 reportDecode2018 report
Decode2018 report
 
Data Science at Scale by Sarah Guido
Data Science at Scale by Sarah GuidoData Science at Scale by Sarah Guido
Data Science at Scale by Sarah Guido
 
Lessons Learned in the Development of a Web-scale Search Engine: Nutch2 and b...
Lessons Learned in the Development of a Web-scale Search Engine: Nutch2 and b...Lessons Learned in the Development of a Web-scale Search Engine: Nutch2 and b...
Lessons Learned in the Development of a Web-scale Search Engine: Nutch2 and b...
 
360|Flex Recap - San Jose 2010
360|Flex Recap - San Jose 2010360|Flex Recap - San Jose 2010
360|Flex Recap - San Jose 2010
 
London level39
London level39London level39
London level39
 
DevDay: Node Analytics with Python, Chainhaus
DevDay: Node Analytics with Python, ChainhausDevDay: Node Analytics with Python, Chainhaus
DevDay: Node Analytics with Python, Chainhaus
 
HPCC Systems Engineering Summit: Community Use Case: Because Who Has Time for...
HPCC Systems Engineering Summit: Community Use Case: Because Who Has Time for...HPCC Systems Engineering Summit: Community Use Case: Because Who Has Time for...
HPCC Systems Engineering Summit: Community Use Case: Because Who Has Time for...
 
20100306 Datasalon 4 : code4lib
20100306 Datasalon 4 : code4lib20100306 Datasalon 4 : code4lib
20100306 Datasalon 4 : code4lib
 

More from Sam Witteveen

Tensor flow 06 tips and tricks
Tensor flow 06 tips and tricksTensor flow 06 tips and tricks
Tensor flow 06 tips and tricksSam Witteveen
 
Tensor flow05 neural-machine-translation-seq2seq
Tensor flow05 neural-machine-translation-seq2seqTensor flow05 neural-machine-translation-seq2seq
Tensor flow05 neural-machine-translation-seq2seqSam Witteveen
 
Tensor flow intro and summit info feb 2017
Tensor flow intro and summit info feb 2017Tensor flow intro and summit info feb 2017
Tensor flow intro and summit info feb 2017Sam Witteveen
 
Tensor flow03 generativemodels part2 8x super resolution
Tensor flow03 generativemodels part2 8x super resolutionTensor flow03 generativemodels part2 8x super resolution
Tensor flow03 generativemodels part2 8x super resolutionSam Witteveen
 
Conversational agents
Conversational agentsConversational agents
Conversational agentsSam Witteveen
 
Hr Summit Biz Genius Presentation
Hr Summit Biz Genius PresentationHr Summit Biz Genius Presentation
Hr Summit Biz Genius PresentationSam Witteveen
 

More from Sam Witteveen (8)

Tensor flow 06 tips and tricks
Tensor flow 06 tips and tricksTensor flow 06 tips and tricks
Tensor flow 06 tips and tricks
 
Tensor flow05 neural-machine-translation-seq2seq
Tensor flow05 neural-machine-translation-seq2seqTensor flow05 neural-machine-translation-seq2seq
Tensor flow05 neural-machine-translation-seq2seq
 
Tensor flow intro and summit info feb 2017
Tensor flow intro and summit info feb 2017Tensor flow intro and summit info feb 2017
Tensor flow intro and summit info feb 2017
 
Tensor flow03 generativemodels part2 8x super resolution
Tensor flow03 generativemodels part2 8x super resolutionTensor flow03 generativemodels part2 8x super resolution
Tensor flow03 generativemodels part2 8x super resolution
 
Conversational agents
Conversational agentsConversational agents
Conversational agents
 
Traction
TractionTraction
Traction
 
Product development
Product developmentProduct development
Product development
 
Hr Summit Biz Genius Presentation
Hr Summit Biz Genius PresentationHr Summit Biz Genius Presentation
Hr Summit Biz Genius Presentation
 

Recently uploaded

Microscopic Analysis of Ceramic Materials.pptx
Microscopic Analysis of Ceramic Materials.pptxMicroscopic Analysis of Ceramic Materials.pptx
Microscopic Analysis of Ceramic Materials.pptxpurnimasatapathy1234
 
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...Soham Mondal
 
What are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptxWhat are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptxwendy cai
 
Concrete Mix Design - IS 10262-2019 - .pptx
Concrete Mix Design - IS 10262-2019 - .pptxConcrete Mix Design - IS 10262-2019 - .pptx
Concrete Mix Design - IS 10262-2019 - .pptxKartikeyaDwivedi3
 
Sachpazis Costas: Geotechnical Engineering: A student's Perspective Introduction
Sachpazis Costas: Geotechnical Engineering: A student's Perspective IntroductionSachpazis Costas: Geotechnical Engineering: A student's Perspective Introduction
Sachpazis Costas: Geotechnical Engineering: A student's Perspective IntroductionDr.Costas Sachpazis
 
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptxDecoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptxJoão Esperancinha
 
Oxy acetylene welding presentation note.
Oxy acetylene welding presentation note.Oxy acetylene welding presentation note.
Oxy acetylene welding presentation note.eptoze12
 
Introduction to Microprocesso programming and interfacing.pptx
Introduction to Microprocesso programming and interfacing.pptxIntroduction to Microprocesso programming and interfacing.pptx
Introduction to Microprocesso programming and interfacing.pptxvipinkmenon1
 
Architect Hassan Khalil Portfolio for 2024
Architect Hassan Khalil Portfolio for 2024Architect Hassan Khalil Portfolio for 2024
Architect Hassan Khalil Portfolio for 2024hassan khalil
 
Artificial-Intelligence-in-Electronics (K).pptx
Artificial-Intelligence-in-Electronics (K).pptxArtificial-Intelligence-in-Electronics (K).pptx
Artificial-Intelligence-in-Electronics (K).pptxbritheesh05
 
Past, Present and Future of Generative AI
Past, Present and Future of Generative AIPast, Present and Future of Generative AI
Past, Present and Future of Generative AIabhishek36461
 
GDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentationGDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentationGDSCAESB
 
Current Transformer Drawing and GTP for MSETCL
Current Transformer Drawing and GTP for MSETCLCurrent Transformer Drawing and GTP for MSETCL
Current Transformer Drawing and GTP for MSETCLDeelipZope
 
chaitra-1.pptx fake news detection using machine learning
chaitra-1.pptx  fake news detection using machine learningchaitra-1.pptx  fake news detection using machine learning
chaitra-1.pptx fake news detection using machine learningmisbanausheenparvam
 

Recently uploaded (20)

Microscopic Analysis of Ceramic Materials.pptx
Microscopic Analysis of Ceramic Materials.pptxMicroscopic Analysis of Ceramic Materials.pptx
Microscopic Analysis of Ceramic Materials.pptx
 
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
 
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCRCall Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
 
POWER SYSTEMS-1 Complete notes examples
POWER SYSTEMS-1 Complete notes  examplesPOWER SYSTEMS-1 Complete notes  examples
POWER SYSTEMS-1 Complete notes examples
 
What are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptxWhat are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptx
 
Concrete Mix Design - IS 10262-2019 - .pptx
Concrete Mix Design - IS 10262-2019 - .pptxConcrete Mix Design - IS 10262-2019 - .pptx
Concrete Mix Design - IS 10262-2019 - .pptx
 
young call girls in Rajiv Chowk🔝 9953056974 🔝 Delhi escort Service
young call girls in Rajiv Chowk🔝 9953056974 🔝 Delhi escort Serviceyoung call girls in Rajiv Chowk🔝 9953056974 🔝 Delhi escort Service
young call girls in Rajiv Chowk🔝 9953056974 🔝 Delhi escort Service
 
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
 
Sachpazis Costas: Geotechnical Engineering: A student's Perspective Introduction
Sachpazis Costas: Geotechnical Engineering: A student's Perspective IntroductionSachpazis Costas: Geotechnical Engineering: A student's Perspective Introduction
Sachpazis Costas: Geotechnical Engineering: A student's Perspective Introduction
 
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptxDecoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
 
Oxy acetylene welding presentation note.
Oxy acetylene welding presentation note.Oxy acetylene welding presentation note.
Oxy acetylene welding presentation note.
 
Introduction to Microprocesso programming and interfacing.pptx
Introduction to Microprocesso programming and interfacing.pptxIntroduction to Microprocesso programming and interfacing.pptx
Introduction to Microprocesso programming and interfacing.pptx
 
Architect Hassan Khalil Portfolio for 2024
Architect Hassan Khalil Portfolio for 2024Architect Hassan Khalil Portfolio for 2024
Architect Hassan Khalil Portfolio for 2024
 
Artificial-Intelligence-in-Electronics (K).pptx
Artificial-Intelligence-in-Electronics (K).pptxArtificial-Intelligence-in-Electronics (K).pptx
Artificial-Intelligence-in-Electronics (K).pptx
 
Past, Present and Future of Generative AI
Past, Present and Future of Generative AIPast, Present and Future of Generative AI
Past, Present and Future of Generative AI
 
GDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentationGDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentation
 
Exploring_Network_Security_with_JA3_by_Rakesh Seal.pptx
Exploring_Network_Security_with_JA3_by_Rakesh Seal.pptxExploring_Network_Security_with_JA3_by_Rakesh Seal.pptx
Exploring_Network_Security_with_JA3_by_Rakesh Seal.pptx
 
Current Transformer Drawing and GTP for MSETCL
Current Transformer Drawing and GTP for MSETCLCurrent Transformer Drawing and GTP for MSETCL
Current Transformer Drawing and GTP for MSETCL
 
chaitra-1.pptx fake news detection using machine learning
chaitra-1.pptx  fake news detection using machine learningchaitra-1.pptx  fake news detection using machine learning
chaitra-1.pptx fake news detection using machine learning
 
Design and analysis of solar grass cutter.pdf
Design and analysis of solar grass cutter.pdfDesign and analysis of solar grass cutter.pdf
Design and analysis of solar grass cutter.pdf
 

PyTorch 04 What's New in PyTorch Land

  • 1. What’s New in PyTorch Land By Sam Witteveen @sam_witteveen
  • 2. About me • Google Developer Expert for Machine Learning & Deep Learning • Deep Learning - Language & Dialogue • Multiple Startups- B2B, B2C • Red Dragon AI
  • 3. • Deep Learning Consulting and Prototyping • Education & Training • Products: • Conversational Computing • Natural Voice Generation - multiple languages • Knowledge Base Creation & Reasoning
  • 4. PyTorch - What’s New • PyTorch 0.4 • PyTorch Road Map • ONNX • Fast.ai • Skorch • Others
  • 5. PyTorch 0.4 • “Tensors” and “Variables” have merged • Device agnostic code - no need for .cuda() • Support for scalar Tensors • Windows support
  • 7. PyTorch Roadmap • Path to 1.0 and to Production • Imperative and declarative • Easy for research, easy for production • Folding Caffe2 into PyTorch • ONNX/Caffe for mobile support
  • 8. PyTorch Roadmap • Fused kernels for declarative graphs • Tools, pre-trained models & Datasets • ELF (game playing), Glow (compiler), FAISS etc • Currently in use in Facebook • Azure & AWS Support • Coming in the next few months.
  • 9. ONNX (Open Neural Network Exchange) • Converting models to Caffe, CNTK, others • 0.4 adds RNN support for PyTorch • TensorFlow support is experimental currently
  • 12. Fast.ai Library • Wrapper for PyTorch • Fast.ai - Jeremy Howard & Rachel Thomas • Adds features beyond PyTorch Vision and Text • Abstracts a lot of the lower level PyTorch code • Wraps Pandas, Spacy, TorchVision, TorchText to create a set of often used data and ML functions • Focuses on data handling and cutting edge techniques
  • 13. Fast.ai Library • date handling, text look ups and embeddings, tabular data, categorical data etc • Latest things like Stochastic Weight Averaging • Very opinionated lib, often well explained in the courses • Function names are not always the easiest to decipher • Documentation ….
  • 15. • A scikit-learn compatible neural network library that wraps pytorch • Supports 0.4 • Allows you to use SKL pipelines, grid search etc • https://github.com/dnouri/skorch
  • 16. Other interesting projects • Pyro 0.2 - https://github.com/uber/pyro • Wisdom - https://github.com/facebookresearch/ visdom • Tensor Board - https://github.com/lanpa/ tensorboard-pytorch
  • 17. Links • https://pytorch.org/2018/04/22/0_4_0-migration- guide.html • https://pytorch.org/2018/05/02/road-to-1.0.html • https://github.com/onnx/onnx • https://github.com/onnx/tensorflow-onnx • https://github.com/fastai/fastai • https://github.com/dnouri/skorch
  • 18. How to Choose? • Worth learning both • If your are building for production go with TensorFlow • If you are doing research or trying to make something new work go with PyTorch
  • 19. The End Contact email: s@maclea.ai twitter: sam_witteveen https://github.com/samwit/