SlideShare una empresa de Scribd logo
1 de 23
Descargar para leer sin conexión
© 2019 Texas Instruments Inc
Machine Learning at the Edge in
Smart Factories Using TI Sitara
Processors
Manisha Agrawal
Texas Instruments Incorporated
May 2019
© 2019 Texas Instruments Inc
Agenda
• Smart factory overview
• Sitara processors for smart factories
• Sitara machine learning overview
• Getting started with Sitara machine learning
• Third-party ecosystem
2
© 2019 Texas Instruments Inc
Smart factory overview
© 2019 Texas Instruments Inc
Industry 4.0 / IIOT / Smart Factory
4
Smart factory overview
Industry 4.0, the Industrial Internet of
Things (IIoT) or Smart Factory are all terms
describing a network of devices connected
via communications technologies to form
systems that monitor, collect, exchange
and analyze data.
Based on four design principles –
Interconnection: Machines, devices, sensors, and people to connect and communicate with each other via the
internet of things (IoT) or the Internet of People (IoP).
Information transparency: the systems create a virtual copy of the physical world through sensor data in order to
contextualize information
Technical assistance: Ability of the systems to support humans in making decisions and solving
problems and the ability to assist humans with tasks that are too difficult or unsafe for humans.
Decentralized decisions: The ability of cyber-physical systems to make simple decisions on their own and
become as autonomous as possible
© 2019 Texas Instruments Inc
Industry 4.0 / IIOT / Smart Factory
Key technology considerations for smart factories:
• Industrial communication protocols
• Real time control of cyber physical system
• Security and functional safety requirements
• Machine learning
• Enhanced reliability
• Device longevity
5
Smart factory overview
Industry 4.0, the Industrial Internet of
Things (IIoT) or Smart Factory are all terms
describing a network of devices connected
via communications technologies to form
systems that monitor, collect, exchange
and analyze data.
© 2019 Texas Instruments Inc
Sitara processors for smart
factories
© 2019 Texas Instruments Inc
Sitara processors enabling smart factories
7
• Industrial robots for automated sorting
• Vision computer & optical Inspection
• Automated Guided Vehicles (AGVs)
• Smart HMI (teaching operators)
• Optimize equipment settings
• Predictive Maintenance (PdM)
• Classify and recognize specific
sounds/audio patterns
• Track/identify/count people and objects
• Extracting useful information gathered
from the data aggregator
Sitara processors for smart factories
© 2019 Texas Instruments Inc
Sitara processors: Key differentiation for smart factories
• Support for multiple industrial communication
protocols on one platform
• Real-time control and processing with
incorporated PRU-ICSS and C66x DSP
• Accelerated machine learning and multimedia
capabilities with integrated hardware accelerators
• Advanced integration for enhanced functionality
• Enables machine learning inference at the edge
for many vision applications
• Predictive Maintenance at the edge at the field,
controller or operator level devices
• CSI-2 and parallel interface for camera
connection
• Support for PCIe, USB3, SATA, etc. connectivity
standards 8
Sitara processors for smart factories
Highly integrated processors supporting industrial peripherals and connectivity needed for
smart factories products
Highly integrated processors for smart
factory applications
© 2019 Texas Instruments Inc
Sitara Processors: Key differentiation for smart factories
• Enhanced Reliability
• Extended temperature & high voltage I/O
• Low Failure in Time (FIT)
• High POH (Power On Hours)
• Enabling lower system power
• Low Latency
• Scalable family of devices
• Software scalability across the entire portfolio
‒ Reuse software across platforms with TI’s processor SDK
• TI.com support
• Technical documentation, including reference designs, on machine learning with Sitara
processors
• Technical support by subject matter experts on e2e.ti.com
9
Sitara processors for smart factories
© 2019 Texas Instruments Inc
Sitara Arm processor overview
10
Sitara processors for smart factories Layout Compatible
AM43x
•1x Cortex-A9, 1 GHz
•Secure Boot, Tamper, QSPI
•Camera, 2x PRU-ICSS, GBE, 3D
•17x17mm
AM335x
•1x Cortex-A8, 1 GHz
•1x PRU-ICSS, 3D, GbE Switch
•DDR3, CAN
•13x13, 15x15mm
AM572x
• 2x Cortex- A15, 1.5 GHz
• 2x C66x DSP, 750 MHz
• 4x Machine Learning Accelerators
• 2x 3D, 4x M4, 2x PRU-ICSS
• 23x23mm HD
AM571x
• 1x Cortex-A15, 1.5 GHz
• 1x C66x DSP, 750 MHz
• 3D, 4x M4, 2x PRU-ICSS
• 23x23mm HD
AM574x
• 2x Cortex-A15, 1.5 GHz
• 2x C66x DSP, 750 MHz
• 2x Machine Learning Accelerators
• 2x 3D, 4x M4, 2x PRU-ICSS
• 23x23mm, ECC HD
AM570x
•1x Cortex-A15, 1.0 GHz
•1x C66x DSP, 750 MHz
•3D, 4x M4, 2x PRU-ICSS
•17x17mm HD
Performance&Integration
AM654x
• 4x A53, up to 1.1 GHz
• 2x R5F MCU/Safety Processor
• 32-bit DDR3/4, PCIe, USB3, 3D
• 3x PRU-ICSS (Gb) , ECC
• 23x23mm
• Highly integrated solution
• Brings Deep Learning to the edge
• Dedicated machine learning accelerators
(Embedded Vision Engine) on AM5729 and
AM5749 device for high performance and low
power
• AM5729 is the highest performing device
• Up to 120 GOPS
© 2019 Texas Instruments Inc
Sitara machine learning overview
© 2019 Texas Instruments Inc
Sitara machine learning
12
Sitara machine learning
Sitara machine learning solutions bring machine learning to the edge by enabling deep learning
inference on all Sitara devices.
Sitara machine learning
solutions consists of:
• TI Deep Learning (TIDL)
software framework
• Available on
AM57x device only,
runs on C66x DSP
cores and/or on
Embedded Vision
Engine (EVE)
subsystems.
• ArmNN: Runs on all
Sitara devices.
© 2019 Texas Instruments Inc
TI Deep Learning (TIDL) software framework on AM57x
13
Sitara machine learning
• TIDL software framework consists of:
– Converter tool to convert TensorFlow, Caffe and
ONNX formats to TIDL network format
– Quantize weight with 4-12 bit accuracy
– Dynamic quantization of activation layer outputs
with 8-bit accuracy
– Accumulation is done with 40-bit accuracy
– C++ based TIDL API on Arm/Linux
– TIDL Library to offload processing on C66x DSP
cores and EVE subsystem
• Ease of use
– Easily integrate TIDL APIs into other frameworks
such as OpenCV
– Provides a common host abstraction for user
applications across multiple compute engines
(EVEs and C66x DSPs)
TIDL software framework enables machine learning inference on Sitara AM57x processors by
leveraging integrated EVE and/or C66x DSP hardware accelerators to run CNN models specifically.
Format Conversion
(desktop / Arm Linux)
Inference – embedded
deployment
(Sitara AM57x)
TIDL-InferenceTIDL-translator
TIDL Import
TIDL API
OpenCL
TIDL Library
© 2019 Texas Instruments Inc
TI Deep Learning (TIDL) software framework
• Provides several TI network models:
• JacintoNet11 for image classification
• JDetNet for classification and detection
• JSeg21 for pixel level semantic segmentation
• TI network models leverages the following for embedded performance optimization:
• Complexity reductions
• Sparse convolutions
• fixed point dynamic quantization
• These tools result in minor decrease in accuracy but result in major computational and
bandwidth reductions
14
Sitara machine learning
Classification
+ Localization
Object detection +
classification
Semantic
Segmentation
© 2019 Texas Instruments Inc
Performance of validated network models on AM57x
15
Sitara machine learning
Network
Model
Type of
Model
ROI
size
Performance in ROIs/Sec
AM5729
(2x Arm A15 @1.5GHz, 4x
EVE @650MHz,
2x C66x DSP @750MHz)
AM5749
(2x Arm A15 @1.5GHz, 2x
EVE @650MHz,
2x C66x DSP @750MHz)
AM572x/AM574x
(2x Arm A15@1.5GHz,
2x C66x DSP @750MHz)
MobileNet_v1 Image
Classification
224x224 7.3 5.4 2.6 1.6
SqueezeNet_v1 Image
Classification
227x227 19.7 9.6 4.6 5.2
InceptionNet_v1 Image
Classification
224x224 6.3 3.2 1.4 1.6
JacinctoNet11 Embedded
Image
Classification
224x224 64.6 33.4 16.6 NA
JSeg21 Embedded
Semantic
Segmentation
1024x512 11.8 3.4 1.8 NA
JDetNet Embedded
Object detection
768x320 22.6 11.2 NA NA
*A15 is available for a real-time control loop or
communication protocol or run additional NN models
Utilizing 4xEVE+1xDSP* Utilizing 2xEVE+1xDSP* Utilizing DSP* Utilizing ARM**
• Power efficient embedded solution at the edge, < 5W
**Arm benchmarking done using Arm NN.
© 2019 Texas Instruments Inc
Getting started with
Sitara machine learning
© 2019 Texas Instruments Inc
beagleboard.zorg/ai
Getting started with Sitara machine learning
Features: 4 EVEs with built-in AM5729
processor in addition to the following
development support:
⚫ Zero-download training workshops
⚫ Tool for everyday automation
⚫ Definitive AM5279 reference design
17
Available from 3Q, 2019
BeagleBone AI (AM5729 SoC)TMDSIDK574 AM574x Industrial Development
Kit (IDK) Board
Features: 2 EVEs, 4 Ethernet ports with concurrent
operation including 2 from PRU-ICSS DDR3 with ECC
Profibus connection, EtherCAT and RS485 Headers On-
board eMMC Mini PCIe, USB3, and HDMI connectors
http://www.ti.com/tool/SITARA-MACHINE-LEARNING
© 2019 Texas Instruments Inc
Third-party ecosystem
© 2019 Texas Instruments Inc
Third-party ecosystem
• Cross Compass: Cross Compass has a strong
reputation in the manufacturing industry, provide
AI consulting services, and develops and provides
algorithms and neural networks.
• Path Partner: Provides services for development,
optimization and integration of AI solutions on TI
platforms
• D3 Engineering: Provides engineering services,
software development and platforms on AI
solutions with TI technology.
• Ittiam Systems: Provides services for high-
performance computer vision, machine learning
and analytics
19
www.cross-compass.com
www.d3engineering.com
www.pathpartnertech.com
www.ittiam.com
© 2019 Texas Instruments Inc
Resources
Sitara Arm processor overview
http://www.ti.com/processors/sitara
-arm/overview.html
Sitara machine learning
http://www.ti.com/tool/SITARA-
MACHINE-LEARNING
Software development Kit
http://software-dl.ti.com/processor-sdk-
linux/esd/AM57X/latest/index_FDS.html
TI Design on TI Deep Learning
http://www.ti.com/tool/TIDEP-01004
TI Deep Learning document
http://software-dl.ti.com/processor-
sdk-
linux/esd/docs/latest/linux/Foundati
onal_Components_TIDL.html
Embedded Vision Summit
“Sitara Machine learning for smart factories”
(Wednesday, May 22 from 11:20 – 11:50 am)
20
© 2019 Texas Instruments Inc
Backup
© 2019 Texas Instruments Inc
Sitara processors enable machine learning inference at
the edge
22
Data transmission
cost
Network latency
Security
Reliability
System power
Network Bandwidth
Network connectivity
Data uploaded to
cloud
Prediction result
returned
Machine learning inferenceSensor data
Action
The Cloud
Inference on the cloud
Sitara machine learning addresses
following concerns from cloud processing
ActionSensor data Machine learning inference
Local hardware
Inference at the edge
© 2019 Texas Instruments Inc
TIDL software framework reference design
TIDEP-01004 Design: available now
• Highly integrated solution bringing deep learning
inference to the edge.
• Guide through dense and sparse model training,
importing and deploying the model for inference run
on AM57x SoC
• API to run multiple (same or different) networks in
parallel on different EVE subsystem and DSP cores.
• Example trained networks ready to run
Benefits
• TIDEP-01004
• Design Guide
• Design Files: Schematics, BOM, Gerbers,
Software, etc.
• Device Datasheets:
‒ AM5749 IDK
‒ AM5749 SoC
• This TI Design demonstrates TI Deep Learning (TIDL)
inference running on an AM5749 IDK.
• Shows running TIDL inference both on the EVE subsystems
and on the C66x DSP cores
• Includes benchmarks of several popular networks on AM57x
• Walk-through of deep learning development flow including
training for sparse network, import and inference deployment
Features
• Machine Vision: Vision Computer, Code Readers
• Automated Machinery: Automated Sorting Equipment,
Optical Inspection
• Logistics Robots: Logistics Robots CPU board
• EPOS: ATMs, Currency Counter
• Many others…
Applications
Tools & Resources

Más contenido relacionado

Más de Edge AI and Vision Alliance

“Computer Vision in Sports: Scalable Solutions for Downmarkets,” a Presentati...
“Computer Vision in Sports: Scalable Solutions for Downmarkets,” a Presentati...“Computer Vision in Sports: Scalable Solutions for Downmarkets,” a Presentati...
“Computer Vision in Sports: Scalable Solutions for Downmarkets,” a Presentati...Edge AI and Vision Alliance
 
“Detecting Data Drift in Image Classification Neural Networks,” a Presentatio...
“Detecting Data Drift in Image Classification Neural Networks,” a Presentatio...“Detecting Data Drift in Image Classification Neural Networks,” a Presentatio...
“Detecting Data Drift in Image Classification Neural Networks,” a Presentatio...Edge AI and Vision Alliance
 
“Deep Neural Network Training: Diagnosing Problems and Implementing Solutions...
“Deep Neural Network Training: Diagnosing Problems and Implementing Solutions...“Deep Neural Network Training: Diagnosing Problems and Implementing Solutions...
“Deep Neural Network Training: Diagnosing Problems and Implementing Solutions...Edge AI and Vision Alliance
 
“AI Start-ups: The Perils of Fishing for Whales (War Stories from the Entrepr...
“AI Start-ups: The Perils of Fishing for Whales (War Stories from the Entrepr...“AI Start-ups: The Perils of Fishing for Whales (War Stories from the Entrepr...
“AI Start-ups: The Perils of Fishing for Whales (War Stories from the Entrepr...Edge AI and Vision Alliance
 
“A Computer Vision System for Autonomous Satellite Maneuvering,” a Presentati...
“A Computer Vision System for Autonomous Satellite Maneuvering,” a Presentati...“A Computer Vision System for Autonomous Satellite Maneuvering,” a Presentati...
“A Computer Vision System for Autonomous Satellite Maneuvering,” a Presentati...Edge AI and Vision Alliance
 
“Bias in Computer Vision—It’s Bigger Than Facial Recognition!,” a Presentatio...
“Bias in Computer Vision—It’s Bigger Than Facial Recognition!,” a Presentatio...“Bias in Computer Vision—It’s Bigger Than Facial Recognition!,” a Presentatio...
“Bias in Computer Vision—It’s Bigger Than Facial Recognition!,” a Presentatio...Edge AI and Vision Alliance
 
“Sensor Fusion Techniques for Accurate Perception of Objects in the Environme...
“Sensor Fusion Techniques for Accurate Perception of Objects in the Environme...“Sensor Fusion Techniques for Accurate Perception of Objects in the Environme...
“Sensor Fusion Techniques for Accurate Perception of Objects in the Environme...Edge AI and Vision Alliance
 
“Updating the Edge ML Development Process,” a Presentation from Samsara
“Updating the Edge ML Development Process,” a Presentation from Samsara“Updating the Edge ML Development Process,” a Presentation from Samsara
“Updating the Edge ML Development Process,” a Presentation from SamsaraEdge AI and Vision Alliance
 
“Combating Bias in Production Computer Vision Systems,” a Presentation from R...
“Combating Bias in Production Computer Vision Systems,” a Presentation from R...“Combating Bias in Production Computer Vision Systems,” a Presentation from R...
“Combating Bias in Production Computer Vision Systems,” a Presentation from R...Edge AI and Vision Alliance
 
“Developing an Embedded Vision AI-powered Fitness System,” a Presentation fro...
“Developing an Embedded Vision AI-powered Fitness System,” a Presentation fro...“Developing an Embedded Vision AI-powered Fitness System,” a Presentation fro...
“Developing an Embedded Vision AI-powered Fitness System,” a Presentation fro...Edge AI and Vision Alliance
 
“Navigating the Evolving Venture Capital Landscape for Edge AI Start-ups,” a ...
“Navigating the Evolving Venture Capital Landscape for Edge AI Start-ups,” a ...“Navigating the Evolving Venture Capital Landscape for Edge AI Start-ups,” a ...
“Navigating the Evolving Venture Capital Landscape for Edge AI Start-ups,” a ...Edge AI and Vision Alliance
 
“Advanced Presence Sensing: What It Means for the Smart Home,” a Presentation...
“Advanced Presence Sensing: What It Means for the Smart Home,” a Presentation...“Advanced Presence Sensing: What It Means for the Smart Home,” a Presentation...
“Advanced Presence Sensing: What It Means for the Smart Home,” a Presentation...Edge AI and Vision Alliance
 
“Tracking and Fusing Diverse Risk Factors to Drive a SAFER Future,” a Present...
“Tracking and Fusing Diverse Risk Factors to Drive a SAFER Future,” a Present...“Tracking and Fusing Diverse Risk Factors to Drive a SAFER Future,” a Present...
“Tracking and Fusing Diverse Risk Factors to Drive a SAFER Future,” a Present...Edge AI and Vision Alliance
 
“MIPI CSI-2 Image Sensor Interface Standard Features Enable Efficient Embedde...
“MIPI CSI-2 Image Sensor Interface Standard Features Enable Efficient Embedde...“MIPI CSI-2 Image Sensor Interface Standard Features Enable Efficient Embedde...
“MIPI CSI-2 Image Sensor Interface Standard Features Enable Efficient Embedde...Edge AI and Vision Alliance
 
“Introduction to the CSI-2 Image Sensor Interface Standard,” a Presentation f...
“Introduction to the CSI-2 Image Sensor Interface Standard,” a Presentation f...“Introduction to the CSI-2 Image Sensor Interface Standard,” a Presentation f...
“Introduction to the CSI-2 Image Sensor Interface Standard,” a Presentation f...Edge AI and Vision Alliance
 
“Practical Approaches to DNN Quantization,” a Presentation from Magic Leap
“Practical Approaches to DNN Quantization,” a Presentation from Magic Leap“Practical Approaches to DNN Quantization,” a Presentation from Magic Leap
“Practical Approaches to DNN Quantization,” a Presentation from Magic LeapEdge AI and Vision Alliance
 
"Optimizing Image Quality and Stereo Depth at the Edge," a Presentation from ...
"Optimizing Image Quality and Stereo Depth at the Edge," a Presentation from ..."Optimizing Image Quality and Stereo Depth at the Edge," a Presentation from ...
"Optimizing Image Quality and Stereo Depth at the Edge," a Presentation from ...Edge AI and Vision Alliance
 
“Using a Collaborative Network of Distributed Cameras for Object Tracking,” a...
“Using a Collaborative Network of Distributed Cameras for Object Tracking,” a...“Using a Collaborative Network of Distributed Cameras for Object Tracking,” a...
“Using a Collaborative Network of Distributed Cameras for Object Tracking,” a...Edge AI and Vision Alliance
 
“A Survey of Model Compression Methods,” a Presentation from Instrumental
“A Survey of Model Compression Methods,” a Presentation from Instrumental“A Survey of Model Compression Methods,” a Presentation from Instrumental
“A Survey of Model Compression Methods,” a Presentation from InstrumentalEdge AI and Vision Alliance
 
“Reinventing Smart Cities with Computer Vision,” a Presentation from Hayden AI
“Reinventing Smart Cities with Computer Vision,” a Presentation from Hayden AI“Reinventing Smart Cities with Computer Vision,” a Presentation from Hayden AI
“Reinventing Smart Cities with Computer Vision,” a Presentation from Hayden AIEdge AI and Vision Alliance
 

Más de Edge AI and Vision Alliance (20)

“Computer Vision in Sports: Scalable Solutions for Downmarkets,” a Presentati...
“Computer Vision in Sports: Scalable Solutions for Downmarkets,” a Presentati...“Computer Vision in Sports: Scalable Solutions for Downmarkets,” a Presentati...
“Computer Vision in Sports: Scalable Solutions for Downmarkets,” a Presentati...
 
“Detecting Data Drift in Image Classification Neural Networks,” a Presentatio...
“Detecting Data Drift in Image Classification Neural Networks,” a Presentatio...“Detecting Data Drift in Image Classification Neural Networks,” a Presentatio...
“Detecting Data Drift in Image Classification Neural Networks,” a Presentatio...
 
“Deep Neural Network Training: Diagnosing Problems and Implementing Solutions...
“Deep Neural Network Training: Diagnosing Problems and Implementing Solutions...“Deep Neural Network Training: Diagnosing Problems and Implementing Solutions...
“Deep Neural Network Training: Diagnosing Problems and Implementing Solutions...
 
“AI Start-ups: The Perils of Fishing for Whales (War Stories from the Entrepr...
“AI Start-ups: The Perils of Fishing for Whales (War Stories from the Entrepr...“AI Start-ups: The Perils of Fishing for Whales (War Stories from the Entrepr...
“AI Start-ups: The Perils of Fishing for Whales (War Stories from the Entrepr...
 
“A Computer Vision System for Autonomous Satellite Maneuvering,” a Presentati...
“A Computer Vision System for Autonomous Satellite Maneuvering,” a Presentati...“A Computer Vision System for Autonomous Satellite Maneuvering,” a Presentati...
“A Computer Vision System for Autonomous Satellite Maneuvering,” a Presentati...
 
“Bias in Computer Vision—It’s Bigger Than Facial Recognition!,” a Presentatio...
“Bias in Computer Vision—It’s Bigger Than Facial Recognition!,” a Presentatio...“Bias in Computer Vision—It’s Bigger Than Facial Recognition!,” a Presentatio...
“Bias in Computer Vision—It’s Bigger Than Facial Recognition!,” a Presentatio...
 
“Sensor Fusion Techniques for Accurate Perception of Objects in the Environme...
“Sensor Fusion Techniques for Accurate Perception of Objects in the Environme...“Sensor Fusion Techniques for Accurate Perception of Objects in the Environme...
“Sensor Fusion Techniques for Accurate Perception of Objects in the Environme...
 
“Updating the Edge ML Development Process,” a Presentation from Samsara
“Updating the Edge ML Development Process,” a Presentation from Samsara“Updating the Edge ML Development Process,” a Presentation from Samsara
“Updating the Edge ML Development Process,” a Presentation from Samsara
 
“Combating Bias in Production Computer Vision Systems,” a Presentation from R...
“Combating Bias in Production Computer Vision Systems,” a Presentation from R...“Combating Bias in Production Computer Vision Systems,” a Presentation from R...
“Combating Bias in Production Computer Vision Systems,” a Presentation from R...
 
“Developing an Embedded Vision AI-powered Fitness System,” a Presentation fro...
“Developing an Embedded Vision AI-powered Fitness System,” a Presentation fro...“Developing an Embedded Vision AI-powered Fitness System,” a Presentation fro...
“Developing an Embedded Vision AI-powered Fitness System,” a Presentation fro...
 
“Navigating the Evolving Venture Capital Landscape for Edge AI Start-ups,” a ...
“Navigating the Evolving Venture Capital Landscape for Edge AI Start-ups,” a ...“Navigating the Evolving Venture Capital Landscape for Edge AI Start-ups,” a ...
“Navigating the Evolving Venture Capital Landscape for Edge AI Start-ups,” a ...
 
“Advanced Presence Sensing: What It Means for the Smart Home,” a Presentation...
“Advanced Presence Sensing: What It Means for the Smart Home,” a Presentation...“Advanced Presence Sensing: What It Means for the Smart Home,” a Presentation...
“Advanced Presence Sensing: What It Means for the Smart Home,” a Presentation...
 
“Tracking and Fusing Diverse Risk Factors to Drive a SAFER Future,” a Present...
“Tracking and Fusing Diverse Risk Factors to Drive a SAFER Future,” a Present...“Tracking and Fusing Diverse Risk Factors to Drive a SAFER Future,” a Present...
“Tracking and Fusing Diverse Risk Factors to Drive a SAFER Future,” a Present...
 
“MIPI CSI-2 Image Sensor Interface Standard Features Enable Efficient Embedde...
“MIPI CSI-2 Image Sensor Interface Standard Features Enable Efficient Embedde...“MIPI CSI-2 Image Sensor Interface Standard Features Enable Efficient Embedde...
“MIPI CSI-2 Image Sensor Interface Standard Features Enable Efficient Embedde...
 
“Introduction to the CSI-2 Image Sensor Interface Standard,” a Presentation f...
“Introduction to the CSI-2 Image Sensor Interface Standard,” a Presentation f...“Introduction to the CSI-2 Image Sensor Interface Standard,” a Presentation f...
“Introduction to the CSI-2 Image Sensor Interface Standard,” a Presentation f...
 
“Practical Approaches to DNN Quantization,” a Presentation from Magic Leap
“Practical Approaches to DNN Quantization,” a Presentation from Magic Leap“Practical Approaches to DNN Quantization,” a Presentation from Magic Leap
“Practical Approaches to DNN Quantization,” a Presentation from Magic Leap
 
"Optimizing Image Quality and Stereo Depth at the Edge," a Presentation from ...
"Optimizing Image Quality and Stereo Depth at the Edge," a Presentation from ..."Optimizing Image Quality and Stereo Depth at the Edge," a Presentation from ...
"Optimizing Image Quality and Stereo Depth at the Edge," a Presentation from ...
 
“Using a Collaborative Network of Distributed Cameras for Object Tracking,” a...
“Using a Collaborative Network of Distributed Cameras for Object Tracking,” a...“Using a Collaborative Network of Distributed Cameras for Object Tracking,” a...
“Using a Collaborative Network of Distributed Cameras for Object Tracking,” a...
 
“A Survey of Model Compression Methods,” a Presentation from Instrumental
“A Survey of Model Compression Methods,” a Presentation from Instrumental“A Survey of Model Compression Methods,” a Presentation from Instrumental
“A Survey of Model Compression Methods,” a Presentation from Instrumental
 
“Reinventing Smart Cities with Computer Vision,” a Presentation from Hayden AI
“Reinventing Smart Cities with Computer Vision,” a Presentation from Hayden AI“Reinventing Smart Cities with Computer Vision,” a Presentation from Hayden AI
“Reinventing Smart Cities with Computer Vision,” a Presentation from Hayden AI
 

Último

Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsMark Billinghurst
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreternaman860154
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)Gabriella Davis
 
[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdfhans926745
 
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024BookNet Canada
 
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024BookNet Canada
 
Breaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountBreaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountPuma Security, LLC
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxMalak Abu Hammad
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Allon Mureinik
 
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptxHampshireHUG
 
Maximizing Board Effectiveness 2024 Webinar.pptx
Maximizing Board Effectiveness 2024 Webinar.pptxMaximizing Board Effectiveness 2024 Webinar.pptx
Maximizing Board Effectiveness 2024 Webinar.pptxOnBoard
 
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking MenDelhi Call girls
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsMaria Levchenko
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024Scott Keck-Warren
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024Rafal Los
 
Scaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationScaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationRadu Cotescu
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking MenDelhi Call girls
 
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationFrom Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationSafe Software
 
Install Stable Diffusion in windows machine
Install Stable Diffusion in windows machineInstall Stable Diffusion in windows machine
Install Stable Diffusion in windows machinePadma Pradeep
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptxLBM Solutions
 

Último (20)

Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR Systems
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreter
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)
 
[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf
 
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
 
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
 
Breaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountBreaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path Mount
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptx
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)
 
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
 
Maximizing Board Effectiveness 2024 Webinar.pptx
Maximizing Board Effectiveness 2024 Webinar.pptxMaximizing Board Effectiveness 2024 Webinar.pptx
Maximizing Board Effectiveness 2024 Webinar.pptx
 
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed texts
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024
 
Scaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationScaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organization
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
 
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationFrom Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
 
Install Stable Diffusion in windows machine
Install Stable Diffusion in windows machineInstall Stable Diffusion in windows machine
Install Stable Diffusion in windows machine
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptx
 

"Machine Learning at the Edge in Smart Factories Using TI Sitara Processors," a Presentation from Texas Instruments

  • 1. © 2019 Texas Instruments Inc Machine Learning at the Edge in Smart Factories Using TI Sitara Processors Manisha Agrawal Texas Instruments Incorporated May 2019
  • 2. © 2019 Texas Instruments Inc Agenda • Smart factory overview • Sitara processors for smart factories • Sitara machine learning overview • Getting started with Sitara machine learning • Third-party ecosystem 2
  • 3. © 2019 Texas Instruments Inc Smart factory overview
  • 4. © 2019 Texas Instruments Inc Industry 4.0 / IIOT / Smart Factory 4 Smart factory overview Industry 4.0, the Industrial Internet of Things (IIoT) or Smart Factory are all terms describing a network of devices connected via communications technologies to form systems that monitor, collect, exchange and analyze data. Based on four design principles – Interconnection: Machines, devices, sensors, and people to connect and communicate with each other via the internet of things (IoT) or the Internet of People (IoP). Information transparency: the systems create a virtual copy of the physical world through sensor data in order to contextualize information Technical assistance: Ability of the systems to support humans in making decisions and solving problems and the ability to assist humans with tasks that are too difficult or unsafe for humans. Decentralized decisions: The ability of cyber-physical systems to make simple decisions on their own and become as autonomous as possible
  • 5. © 2019 Texas Instruments Inc Industry 4.0 / IIOT / Smart Factory Key technology considerations for smart factories: • Industrial communication protocols • Real time control of cyber physical system • Security and functional safety requirements • Machine learning • Enhanced reliability • Device longevity 5 Smart factory overview Industry 4.0, the Industrial Internet of Things (IIoT) or Smart Factory are all terms describing a network of devices connected via communications technologies to form systems that monitor, collect, exchange and analyze data.
  • 6. © 2019 Texas Instruments Inc Sitara processors for smart factories
  • 7. © 2019 Texas Instruments Inc Sitara processors enabling smart factories 7 • Industrial robots for automated sorting • Vision computer & optical Inspection • Automated Guided Vehicles (AGVs) • Smart HMI (teaching operators) • Optimize equipment settings • Predictive Maintenance (PdM) • Classify and recognize specific sounds/audio patterns • Track/identify/count people and objects • Extracting useful information gathered from the data aggregator Sitara processors for smart factories
  • 8. © 2019 Texas Instruments Inc Sitara processors: Key differentiation for smart factories • Support for multiple industrial communication protocols on one platform • Real-time control and processing with incorporated PRU-ICSS and C66x DSP • Accelerated machine learning and multimedia capabilities with integrated hardware accelerators • Advanced integration for enhanced functionality • Enables machine learning inference at the edge for many vision applications • Predictive Maintenance at the edge at the field, controller or operator level devices • CSI-2 and parallel interface for camera connection • Support for PCIe, USB3, SATA, etc. connectivity standards 8 Sitara processors for smart factories Highly integrated processors supporting industrial peripherals and connectivity needed for smart factories products Highly integrated processors for smart factory applications
  • 9. © 2019 Texas Instruments Inc Sitara Processors: Key differentiation for smart factories • Enhanced Reliability • Extended temperature & high voltage I/O • Low Failure in Time (FIT) • High POH (Power On Hours) • Enabling lower system power • Low Latency • Scalable family of devices • Software scalability across the entire portfolio ‒ Reuse software across platforms with TI’s processor SDK • TI.com support • Technical documentation, including reference designs, on machine learning with Sitara processors • Technical support by subject matter experts on e2e.ti.com 9 Sitara processors for smart factories
  • 10. © 2019 Texas Instruments Inc Sitara Arm processor overview 10 Sitara processors for smart factories Layout Compatible AM43x •1x Cortex-A9, 1 GHz •Secure Boot, Tamper, QSPI •Camera, 2x PRU-ICSS, GBE, 3D •17x17mm AM335x •1x Cortex-A8, 1 GHz •1x PRU-ICSS, 3D, GbE Switch •DDR3, CAN •13x13, 15x15mm AM572x • 2x Cortex- A15, 1.5 GHz • 2x C66x DSP, 750 MHz • 4x Machine Learning Accelerators • 2x 3D, 4x M4, 2x PRU-ICSS • 23x23mm HD AM571x • 1x Cortex-A15, 1.5 GHz • 1x C66x DSP, 750 MHz • 3D, 4x M4, 2x PRU-ICSS • 23x23mm HD AM574x • 2x Cortex-A15, 1.5 GHz • 2x C66x DSP, 750 MHz • 2x Machine Learning Accelerators • 2x 3D, 4x M4, 2x PRU-ICSS • 23x23mm, ECC HD AM570x •1x Cortex-A15, 1.0 GHz •1x C66x DSP, 750 MHz •3D, 4x M4, 2x PRU-ICSS •17x17mm HD Performance&Integration AM654x • 4x A53, up to 1.1 GHz • 2x R5F MCU/Safety Processor • 32-bit DDR3/4, PCIe, USB3, 3D • 3x PRU-ICSS (Gb) , ECC • 23x23mm • Highly integrated solution • Brings Deep Learning to the edge • Dedicated machine learning accelerators (Embedded Vision Engine) on AM5729 and AM5749 device for high performance and low power • AM5729 is the highest performing device • Up to 120 GOPS
  • 11. © 2019 Texas Instruments Inc Sitara machine learning overview
  • 12. © 2019 Texas Instruments Inc Sitara machine learning 12 Sitara machine learning Sitara machine learning solutions bring machine learning to the edge by enabling deep learning inference on all Sitara devices. Sitara machine learning solutions consists of: • TI Deep Learning (TIDL) software framework • Available on AM57x device only, runs on C66x DSP cores and/or on Embedded Vision Engine (EVE) subsystems. • ArmNN: Runs on all Sitara devices.
  • 13. © 2019 Texas Instruments Inc TI Deep Learning (TIDL) software framework on AM57x 13 Sitara machine learning • TIDL software framework consists of: – Converter tool to convert TensorFlow, Caffe and ONNX formats to TIDL network format – Quantize weight with 4-12 bit accuracy – Dynamic quantization of activation layer outputs with 8-bit accuracy – Accumulation is done with 40-bit accuracy – C++ based TIDL API on Arm/Linux – TIDL Library to offload processing on C66x DSP cores and EVE subsystem • Ease of use – Easily integrate TIDL APIs into other frameworks such as OpenCV – Provides a common host abstraction for user applications across multiple compute engines (EVEs and C66x DSPs) TIDL software framework enables machine learning inference on Sitara AM57x processors by leveraging integrated EVE and/or C66x DSP hardware accelerators to run CNN models specifically. Format Conversion (desktop / Arm Linux) Inference – embedded deployment (Sitara AM57x) TIDL-InferenceTIDL-translator TIDL Import TIDL API OpenCL TIDL Library
  • 14. © 2019 Texas Instruments Inc TI Deep Learning (TIDL) software framework • Provides several TI network models: • JacintoNet11 for image classification • JDetNet for classification and detection • JSeg21 for pixel level semantic segmentation • TI network models leverages the following for embedded performance optimization: • Complexity reductions • Sparse convolutions • fixed point dynamic quantization • These tools result in minor decrease in accuracy but result in major computational and bandwidth reductions 14 Sitara machine learning Classification + Localization Object detection + classification Semantic Segmentation
  • 15. © 2019 Texas Instruments Inc Performance of validated network models on AM57x 15 Sitara machine learning Network Model Type of Model ROI size Performance in ROIs/Sec AM5729 (2x Arm A15 @1.5GHz, 4x EVE @650MHz, 2x C66x DSP @750MHz) AM5749 (2x Arm A15 @1.5GHz, 2x EVE @650MHz, 2x C66x DSP @750MHz) AM572x/AM574x (2x Arm A15@1.5GHz, 2x C66x DSP @750MHz) MobileNet_v1 Image Classification 224x224 7.3 5.4 2.6 1.6 SqueezeNet_v1 Image Classification 227x227 19.7 9.6 4.6 5.2 InceptionNet_v1 Image Classification 224x224 6.3 3.2 1.4 1.6 JacinctoNet11 Embedded Image Classification 224x224 64.6 33.4 16.6 NA JSeg21 Embedded Semantic Segmentation 1024x512 11.8 3.4 1.8 NA JDetNet Embedded Object detection 768x320 22.6 11.2 NA NA *A15 is available for a real-time control loop or communication protocol or run additional NN models Utilizing 4xEVE+1xDSP* Utilizing 2xEVE+1xDSP* Utilizing DSP* Utilizing ARM** • Power efficient embedded solution at the edge, < 5W **Arm benchmarking done using Arm NN.
  • 16. © 2019 Texas Instruments Inc Getting started with Sitara machine learning
  • 17. © 2019 Texas Instruments Inc beagleboard.zorg/ai Getting started with Sitara machine learning Features: 4 EVEs with built-in AM5729 processor in addition to the following development support: ⚫ Zero-download training workshops ⚫ Tool for everyday automation ⚫ Definitive AM5279 reference design 17 Available from 3Q, 2019 BeagleBone AI (AM5729 SoC)TMDSIDK574 AM574x Industrial Development Kit (IDK) Board Features: 2 EVEs, 4 Ethernet ports with concurrent operation including 2 from PRU-ICSS DDR3 with ECC Profibus connection, EtherCAT and RS485 Headers On- board eMMC Mini PCIe, USB3, and HDMI connectors http://www.ti.com/tool/SITARA-MACHINE-LEARNING
  • 18. © 2019 Texas Instruments Inc Third-party ecosystem
  • 19. © 2019 Texas Instruments Inc Third-party ecosystem • Cross Compass: Cross Compass has a strong reputation in the manufacturing industry, provide AI consulting services, and develops and provides algorithms and neural networks. • Path Partner: Provides services for development, optimization and integration of AI solutions on TI platforms • D3 Engineering: Provides engineering services, software development and platforms on AI solutions with TI technology. • Ittiam Systems: Provides services for high- performance computer vision, machine learning and analytics 19 www.cross-compass.com www.d3engineering.com www.pathpartnertech.com www.ittiam.com
  • 20. © 2019 Texas Instruments Inc Resources Sitara Arm processor overview http://www.ti.com/processors/sitara -arm/overview.html Sitara machine learning http://www.ti.com/tool/SITARA- MACHINE-LEARNING Software development Kit http://software-dl.ti.com/processor-sdk- linux/esd/AM57X/latest/index_FDS.html TI Design on TI Deep Learning http://www.ti.com/tool/TIDEP-01004 TI Deep Learning document http://software-dl.ti.com/processor- sdk- linux/esd/docs/latest/linux/Foundati onal_Components_TIDL.html Embedded Vision Summit “Sitara Machine learning for smart factories” (Wednesday, May 22 from 11:20 – 11:50 am) 20
  • 21. © 2019 Texas Instruments Inc Backup
  • 22. © 2019 Texas Instruments Inc Sitara processors enable machine learning inference at the edge 22 Data transmission cost Network latency Security Reliability System power Network Bandwidth Network connectivity Data uploaded to cloud Prediction result returned Machine learning inferenceSensor data Action The Cloud Inference on the cloud Sitara machine learning addresses following concerns from cloud processing ActionSensor data Machine learning inference Local hardware Inference at the edge
  • 23. © 2019 Texas Instruments Inc TIDL software framework reference design TIDEP-01004 Design: available now • Highly integrated solution bringing deep learning inference to the edge. • Guide through dense and sparse model training, importing and deploying the model for inference run on AM57x SoC • API to run multiple (same or different) networks in parallel on different EVE subsystem and DSP cores. • Example trained networks ready to run Benefits • TIDEP-01004 • Design Guide • Design Files: Schematics, BOM, Gerbers, Software, etc. • Device Datasheets: ‒ AM5749 IDK ‒ AM5749 SoC • This TI Design demonstrates TI Deep Learning (TIDL) inference running on an AM5749 IDK. • Shows running TIDL inference both on the EVE subsystems and on the C66x DSP cores • Includes benchmarks of several popular networks on AM57x • Walk-through of deep learning development flow including training for sparse network, import and inference deployment Features • Machine Vision: Vision Computer, Code Readers • Automated Machinery: Automated Sorting Equipment, Optical Inspection • Logistics Robots: Logistics Robots CPU board • EPOS: ATMs, Currency Counter • Many others… Applications Tools & Resources