WeAreDevelopers Logo WeAreDevelopers
Sign in
For You LIVE
JavaScript Congress • Nov 25, 2021

Making neural networks portable with ONNX

Ron Dagdag

Struggling to deploy Python models in Java or JavaScript? ONNX acts as a universal translator for neural networks. Decouple training from deployment and ship AI to the edge seamlessly.

Pause
Mute Enter Fullscreen
#1 about 3 min

Establishing introductory context regarding artificial intelligence deployments

Framing context around artificial intelligence deployment capabilities establishes foundational concepts for developers transitioning operational logic structures.

#2 about 3 min

Differences between traditional programming and machine learning algorithms

Providing structured data as training examples allows programmatic pipelines to automatically write an algorithm solving a core problem.

#3 about 4 min

Bridging application frameworks with the universal ONNX format

Creating universal model formats allows artificial intelligence capabilities to run seamlessly across completely disconnected server architectures.

#4 about 3 min

Ideal system architectures and use cases for portability

Removing heavy dependencies reduces computational latency making models ideal for constrained environments running directly on hardware endpoints.

#5 about 5 min

Sourcing and generating pre-trained machine learning models

Bootstrapping software capabilities via public model repositories accelerates immediate project value avoiding prolonged dataset configuration overhead.

#6 about 2 min

Visualizing model operation graphs with the Netron application

Inspecting loaded payload architectures illuminates proper array dimensions and variable names needed when coding inference execution scripts.

#7 about 6 min

Converting existing application architectures into interoperable runtime graphs

Recompiling localized scripts into universally accessible payloads unifies continuous integration infrastructure when managing experimental operational variants.

#8 about 7 min

Comparing cloud compute infrastructure against local edge deployment

Distributing computation targets closer to client networks lowers cloud ingestion fees and resolves critical bandwidth dependency factors.

#9 about 2 min

Integrating capabilities through high performance ONNX runtime engines

Abstracting execution paths ensures analytical computations fall gracefully back on existing hardware graphic processors determining automatic render speeds.

#10 about 4 min

Writing data payload mappings inside Node backend instances

Translating primitive numbers into valid session instances enables raw backend environments securely interpreting loaded runtime regression logic definitions.

#11 about 3 min

Running independent inference graphs securely on client browsers

Isolating execution evaluations natively within standard browser sessions protects client transmission events and avoids continuous architectural hosting limits.

#12 about 7 min

Processing browser image states into acceptable neural formats

Programmatically resizing source images matching specified training dimensions accurately transforms raw inputs correctly predicting contextual emotion markers.

#13 about 2 min

Optimizing graph layers targeting constrained mobile execution environments

Optimizing parameter weight scales tightens application boundaries preventing bloated bundle downloads restricting underlying runtime mobile execution capabilities.

#14 about 10 min

Addressing audience questions on entering data science domains

Exploring straightforward remote interfaces builds comfortable developer understanding before studying complex foundational computation architecture initially required completely.

Matching moments

2:57 min

Standardizing interoperable model deployments with the ONNX framework

Jose Luis Latorre Millas · LIVE

1:55 min

Exporting framework independent representations for edge processing platforms

Nico Schmidt · LIVE

2:15 min

Open-source community and machine learning frameworks

Gian Marco Iodice Gian Marco Iodice · World Congress 2025

2:46 min

Leveraging ONNX Runtime Web for local model execution

Maxim Salnikov Maxim Salnikov · World Congress 2025

4:47 min

Exporting and evaluating learned ONNX models

Marco Zamana · LIVE

57 sec

Comparing ONNX runtime web and TensorFlow deployments

Jason Mayes · Coffee With Developers

More from JavaScript Congress

The Eternal Sunshine of the Zero Build Pipeline thumbnail
50:10
almost 5 years ago

The Eternal Sunshine of the Zero Build Pipeline

Teini - an extremely small webshop leveraging awesome and free tech thumbnail
40:45
almost 5 years ago

Teini - an extremely small webshop leveraging awesome and free tech

Christian K

Native Web Apps: Are We There Yet? thumbnail
44:40
almost 5 years ago

Native Web Apps: Are We There Yet?

Maxim Salnikov

Modern Web Development with Nuxt3 thumbnail
41:43
almost 5 years ago

Modern Web Development with Nuxt3

Alexander Lichter

JavaScript the Grumpy Parts thumbnail
46:08
almost 5 years ago

JavaScript the Grumpy Parts

Rob Richardson

How Your Bundle Size Affects The Climate thumbnail
53:16
almost 5 years ago

How Your Bundle Size Affects The Climate

Roy Derks

Upcoming sessions on this topic

Open session

World Congress 2026 North America

September 24, 2026 · 14:10–14:40

Stage 5

Edge AI: Running Agentic Intelligence Where Internet Can't Reach

Nitin Eusebius

AWS - Principal Solutions Architect

Nitin Eusebius
Open session

World Congress 2026 North America

September 25, 2026 · 13:30–14:00

Stage 1

The State of Local AI in 2026

Kirah Sapong

Co-founder & CTO of Aquaduck AI

Kirah Sapong
Open session

World Congress 2026 North America

September 24, 2026 · 17:30–18:00

Stage 6

No Single Model to Rule Them All: Building Resilient AI Agents Across Open & Closed LLMs

Emmanuel Acheampong

Senior Manager Developer Relations at Crusoe AI

Emmanuel Acheampong
Open session

World Congress 2026 North America

September 23, 2026 · 10:00–17:00

Stage 11

Building Stuff with GenAI - The Open Minded Workshop beyond OpenAI

Andreas Erben

CTO for Applied AI and Metaverse at daenet

Andreas Erben
Open session

World Congress 2026 North America

September 24, 2026 · 14:10–14:40

Stage 1

Anatomy of an AI Request: Where Latency and Cost Are Really Born

Dan Fu

VP of Kernels at Together AI

Dan Fu
Open session

World Congress 2026 North America

September 25, 2026 · 16:50–17:20

Stage 7

You Can’t Re-Run Sunlight: Designing ML Data Architectures for Physical AI

An Phan

Senior Data Infrastructure Engineer @ Hippo Harvest

An Phan

Related videos

Privacy-first in-browser Generative AI web apps: offline-ready, future-proof, standards-based thumbnail
30:13
World Congress 2025 • about 1 year ago

Privacy-first in-browser Generative AI web apps: offline-ready, future-proof, standards-based

Maxim Salnikov

Localized Open Models in Production: What Builders Need to Know thumbnail
31:45
World Congress 2026 Europe • 2 months ago

Localized Open Models in Production: What Builders Need to Know

Ankit Patel, Pierre-Louis Cedoz, Jamie Madden, Stephen Batifol

RTX AI PC: Developing local and edge AI applications thumbnail
26:06
World Congress 2026 Europe • 2 months ago

RTX AI PC: Developing local and edge AI applications

Joerg Krall

Nemotron: NVIDIA's open model strategy for developers thumbnail
27:47
World Congress 2026 Europe • 2 months ago

Nemotron: NVIDIA's open model strategy for developers

Sergio Perez

How Web AI Can Power the Agentic Web - Jason Mayes (Google) thumbnail
20:57
Coffee With Developers • 4 months ago

How Web AI Can Power the Agentic Web - Jason Mayes (Google)

Jason Mayes

Trends, Challenges and Best Practices for AI at the Edge thumbnail
29:50
World Congress 2023 • about 3 years ago

Trends, Challenges and Best Practices for AI at the Edge

Ekaterina Sirazitdinova

Related articles

From learning to earning

Jobs that call for the skills explored in this talk.