We are off to a good start and have already shot past 60% thanks to your support. As word gets around and more backers learn about K210 AI Accelerator, we full expect this campaign to succeed!
Firstly, I would like to clarify that as of now, the face detection model based on RetinaNet and custom object detection based on Yolo-like models are working and is available for free. We continue to focus on object detection as this fulfills most of your project requirements. We are also working hard to release the TensorFlow project on Google Colab to help you train your own models. That is why other models such as age/gender estimation, voice, and anomaly detection are categorized as something that we might release in the future. But there is no stopping you from developing it yourself and we would be glad to help!
This year’s Lunar New Year falls in February. During this period, the factories in China would be shut down for a minimum of a week. In anticipation, we had placed extra orders and received them for the earlier batch. This allows us to create some buffer. During this time we will also be developing our test jig for automated testing while the factory is closed.
Depending on the response, we will also be placing orders for key components with a long lead time so that we would be able to fulfill your orders the soonest after the end of the campaign.
The PI hat is quite a small PCB. And to efficiently produce and assemble them, the factory packed 8 of them onto a single PCB panel. They are then broken up into individual pieces before testing and packaging. This process is commonly known as Panelization.
To end this update, here are some photos of the boards we currently have in hand. We hope to get our test jig up soon and put these through all the tests.
So, that’s all for this update. Please help to spread the word and get your friends to joins us in bringing AI to the edge.
Help fund this project! Spread the word! Thank you!!!
A 100% open source dev board for the EOS S3 low-power MCU with embedded FPGA
The open-source stereoscopic camera based on Raspberry Pi with Wi-Fi, Bluetooth, and an advanced powering system
An easy-to-use LiDAR camera that acts like a webcam for real-time 3D depth data