- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Register here for the free webinar: https://software.seek.intel.com/Edge_Devices_Webinar_Reg
Tuesday, March 13, 2018
9:00 am - 10:00 am PST
Deploying Image Classifiers on Edge Devices
Learn how to profile, optimize, and deploy image classifiers on edge devices. In this webinar, Ashwin Vijayakumar will walk through the process of profiling pre-trained neural networks designed for image classification, identify a good balance between accuracy and real-time performance, and write a simple Python* script to deploy these classifiers on the Intel® Movidius™ Neural Compute Stick.
What You Will Learn:
How to deploy pre-trained neural networks on edge devices.
How to use Intel® Movidius™ Neural Compute SDK to profile a pre-trained neural network.
How to write a simple Python* app that performs classification on a live camera feed.
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Where is the link? I registered on that page but never received actual instructions on how to join the webinar!
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I couldnt join because there were no link for broadcast nor anything been sent in email
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
@GoldenWings , @smineyes , sorry to hear you didn't get the confirmation email. Can you please check your spam or re register @ https://software.seek.intel.com/Edge_Devices_Webinar_Reg? The webinar is available on-demand, so you should get an email from "Intel Software Developer Zone", titled "Your webinar is now available on demand".
I registered from my personal account as a test; I got the email.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Ashwin,
Thanks for this Webinar. I liked it a lot and it helped sorting some things out.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
@jfey, glad to hear that the webinar was helpful. Please do let us know if you have topic suggestions for future webinars.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Is this available for download or view somewhere?
Interested in watching this.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
@chicagobob123 , all of our webinars are available on-demand.
Deploying image classifiers (Caffe & TensorFlow) - https://software.intel.com/en-us/videos/deploying-image-classifiers-on-intel-movidius-neural-compute-stick
Introduction to NCS - https://software.seek.intel.com/NeuralComputeStickWebinar_Reg
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I looked for the blog post where you showed how you built your network and could not find it. Do you have direct link?
When I typed in movidius.com/blogs or movidius.com/blog I went to the news page. I understood that you removed several layers to get a speed up and went to use a frozen mobile network.
I am having real problems just getting my first retrained network to work. I have only 10 to 20 items to classify and none of them are in the generic databases except for people so I need to retrain.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
@chicagobob123,
If you are referring to https://movidius.github.io/blog/deploying-custom-caffe-models/, please use the dogsvscats project on my fork - https://github.com/ashwinvijayakumar/ncappzoo/tree/dogs-vs-cats/apps/dogsvscats. I am in the process of creating a pull request into the main AppZoo repo.
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page