SLIDE 1
World's Fastest Machine Learning With GPUs
http://github.com/h2oai/h2o4gpu
Speaker: Jonathan C. McKinney
H2O4GPU TEAM
Mateusz Erin Navdeep Rory Terry Karen Arno Jonathan S teve
World's Fastest Machine Learning With GPUs - - PowerPoint PPT Presentation
World's Fastest Machine Learning With GPUs - - PowerPoint PPT Presentation
World's Fastest Machine Learning With GPUs http://github.com/h2oai/h2o4gpu Speaker: Jonathan C. McKinney Mateusz Erin Navdeep Rory Terry Karen Arno Jonathan S teve H2O4GPU TEAM Machine Learning Deep Learning c RIS
SLIDE 2
SLIDE 3
Machine Learning
c
Deep Learning
SLIDE 4
SLIDE 5
6
RIS E OF GPU COMPUTING
GPU-Computing perf 1.5X per year 1000X by 2025 102 103 104 105 106 107 S ingle-threaded perf 1.5X per year 1.1X per year APPLICATIONS S YS TEMS ALGORITHMS CUDA ARCHITECTURE
SLIDE 6
SLIDE 7
8
github.com/ gpuopenanalytics
GPU Data Frame (GDF)
Ingest/ Parse Exploratory Analysis Feature Engineering ML/DL Algorithms Grid Search Scoring Model Export
SLIDE 8
H2O4GPU
/ Open-Source: http://github.com/h2oai/h2o4gpu / Used within our own Driverless AI Product to boost performance 30X / Scikit-Learn Python API (and now R API) / All Scikit-Learn algorithms included / Important algorithms ported to GPU
SLIDE 9
SLIDE 10
11
Driverless AI
https://www.youtube.com/watch?v=KkvWX3FD7yI
SLIDE 11
12
Driverless AI
SLIDE 12
13
Model Accuracy & Speed
SLIDE 13
Generalized Linear Model
/ Algorithm: ‒ A solver for convex optimization problems in graph form using Alternating Direction Method of Multipliers (ADMM) / Solvers: Lasso, Ridge Regression, Logistic Regression, and Elastic Net Regularization / Improvements to original POGS: ‒ Full alpha search ‒ Cross Validation ‒ Early Stopping + Warm Start ‒ Added Scikit-learn like API ‒ Supports multiple GPUs
SLIDE 14
15
https://www.youtube.com/watch?v=LrC3mBNG7WU
https:/ / github.com/ h2oai/ h2o4gpu/ blob/ master/ exa mples/ py/ demos/ Multi-GPU-H2O-GLM-simple.ipynb
SLIDE 15
16
https://www.youtube.com/watch?v=4RKSXNfreLE
SLIDE 16
K-Means
- Significantly faster than scikit-learn implementation (50x)
- Significantly faster than other GPU implementations (5x-10x)
- Supports kmeans++/kmeans|| initialization
- Supports multiple GPUs
- Supports batching data if exceeds GPU memory
SLIDE 17
https://github.com/h2oai/h2o4gpu/blob/master/examples/py/demos/H2O4GPU_KMeans_Images.ipynb
K-Means
SLIDE 18
19
10 with latest solver
SLIDE 19
SLIDE 20
Principle Component Analysis (PCA)
SLIDE 21
SLIDE 22
SLIDE 23
SLIDE 24
Generate faces from PCA
SLIDE 25
SLIDE 26
Gradient Boosting Machines
/ Based upon XGBoost / Raw floating point data -> Binned into Quantiles / Quantiles are stored as compressed instead of floats / Compressed Quantiles are efficiently transferred to GPU / Sparsity is handled directly with highly GPU efficiency / Multi-GPU by sharding rows using NVIDIA NCCL AllReduce
SLIDE 27
Tree Growth Algorithms
SLIDE 28
29
https://www.youtube.com/watch?v=NkeSDrifJdg
171 with latest solver 87 51
SLIDE 29
SLIDE 30
31
Driverless AI on GPUs
https://www.youtube.com/watch?v=KkvWX3FD7yI
SLIDE 31
32
SLIDE 32
33
Driverless AI — Competitive with Kagglers!
Top 8 position in Kaggle with zero manual labor! (ranked above multiple Kaggle Grandmasters)
https://www.kaggle.com/c/mercedes- benz-greener-manufacturing/leaderboard
SLIDE 33