Skip to content

Commit

Permalink
Add video and details about Raspberry Pi's new AI Kit.
Browse files Browse the repository at this point in the history
  • Loading branch information
geerlingguy committed Jun 4, 2024
1 parent 6d92fb5 commit 6df8fe9
Show file tree
Hide file tree
Showing 8 changed files with 20 additions and 6 deletions.
3 changes: 2 additions & 1 deletion _cards_m2/coral-accelerator-ae-key.md
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,8 @@ functionality_pi5: "Full"
driver_required: "Yes"
github_issue: "https://github.com/geerlingguy/raspberry-pi-pcie-devices/issues/44"
buy_link: https://amzn.to/3jWowRg
videos: []
videos:
- https://www.youtube.com/watch?v=HgIMJbN0DS0
---
Coral TPU accelerators offer an easy way to add advanced (and fast!) ML and AI coprocessors to embedded devices. You can use them for things like image recognition, and they are integrated into popular software like Frigate.

Expand Down
3 changes: 2 additions & 1 deletion _cards_m2/coral-m2-accelerator-dual-edge-tpu.md
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,8 @@ functionality_pi5: "Full"
driver_required: "Yes"
github_issue: "https://github.com/geerlingguy/raspberry-pi-pcie-devices/issues/318"
buy_link: https://coral.ai/products/m2-accelerator-dual-edgetpu/
videos: []
videos:
- https://www.youtube.com/watch?v=HgIMJbN0DS0
---
The [Coral M.2 Accelerator with Dual Edge TPU](https://coral.ai/products/m2-accelerator-dual-edgetpu/) uses an interesting feature of M.2 E-key slots—it uses both lanes that are in the spec to the slot (though most board manufacturers only implement one lane per slot).

Expand Down
5 changes: 4 additions & 1 deletion _cards_m2/hailo-8-ai-module.md
Original file line number Diff line number Diff line change
Expand Up @@ -7,8 +7,11 @@ functionality_pi5: "Full"
driver_required: "Yes"
github_issue: "https://github.com/geerlingguy/raspberry-pi-pcie-devices/issues/164"
buy_link: https://hailo.ai/products/ai-accelerators/hailo-8-m2-ai-acceleration-module/
videos: []
videos:
- https://www.youtube.com/watch?v=HgIMJbN0DS0
---
The Hailo-8 is an AI accelerator module that is keyed for multiple M.2 pinouts, including M, B+M, and A+E, in sizes ranging from 2230 to 2242, and you can get a model with breakable extensions out to 60 and 80mm deep.

It uses PCIe Gen 3 x2 or x4, and boasts 26 TOPS performance. It's compatible with TensorFLow, TensorFlow Lite, ONNX, Keras, Pytorch, and it should work with Compute Module 4 and Pi 5, though testing will be documented in the linked GitHub issue.

The smaller cousin, the Hailo-8L, is included in Raspberry Pi's [AI Kit](https://www.raspberrypi.com/products/ai-kit/), and use with the Raspberry Pi is documented here: [AI Kit Documentation](https://www.raspberrypi.com/documentation/accessories/ai-kit.html). Hailo maintains example code and more documentation on their GitHub: [hailo-rpi5-examples](https://github.com/hailo-ai/hailo-rpi5-examples).
5 changes: 4 additions & 1 deletion _cards_m2/hailo-8l-ai-module.md
Original file line number Diff line number Diff line change
Expand Up @@ -7,8 +7,11 @@ functionality_pi5: "Untested"
driver_required: "Yes"
github_issue: "https://github.com/geerlingguy/raspberry-pi-pcie-devices/issues/650"
buy_link: https://hailo.ai/products/ai-accelerators/hailo-8l-ai-accelerator-for-ai-light-applications/
videos: []
videos:
- https://www.youtube.com/watch?v=HgIMJbN0DS0
---
The [Hailo-8L](https://hailo.ai/products/ai-accelerators/hailo-8l-ai-accelerator-for-ai-light-applications/) is Hailo's 'Entry Level' AI accelerator, meant for edge and lighter use cases.

It runs at 13 TOPS, and 8 TOPS/W, and was announced late last year. It [comes in two M.2 form factors](https://hailo.ai/files/hailo-8l-product-brief-en/), a 2242-size B+M-key version, and a 2230-size A+E-key version.

The Hailo-8L is included in Raspberry Pi's [AI Kit](https://www.raspberrypi.com/products/ai-kit/), and use with the Raspberry Pi is documented here: [AI Kit Documentation](https://www.raspberrypi.com/documentation/accessories/ai-kit.html). Hailo maintains example code and more documentation on their GitHub: [hailo-rpi5-examples](https://github.com/hailo-ai/hailo-rpi5-examples).
5 changes: 4 additions & 1 deletion _hats/pi-nvme-hat.md
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,10 @@ github_issue: "https://github.com/geerlingguy/raspberry-pi-pcie-devices/issues/5
link: "https://www.raspberrypi.com/products/m2-hat-plus/"
videos:
- https://www.youtube.com/watch?v=GYN3ub8Qb_I
- https://www.youtube.com/watch?v=HgIMJbN0DS0
---
The Raspberry Pi NVMe HAT allows you to add on 2230 and 2242-size NVMe SSDs, and fits inside the official Pi 5 case, though with little room for airflow.
The Raspberry Pi M.2 NVMe HAT allows you to add on 2230 and 2242-size NVMe SSDs, and fits inside the official Pi 5 case, though with little room for airflow.

There is enough room underneath to fit an Active Cooler.

The M.2 HAT is also used as the base for the AI Kit, which attaches a Hailo-8L AI Accelerator to the Raspberry Pi 5.
3 changes: 2 additions & 1 deletion _hats/pineboards-hatai-dual-edge-tpu.md
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,8 @@ status: production
picture: "/images/hat-pineboards-hatai-dual-edge-tpu.jpg"
github_issue: "https://github.com/geerlingguy/raspberry-pi-pcie-devices/issues/648"
link: "https://pineboards.io/products/hat-ai-dual-edge-coral-tpu-bundle-for-raspberry-pi-5"
videos: []
videos:
- https://www.youtube.com/watch?v=HgIMJbN0DS0
---
Pineboards offers a [Hat AI! Dual Edge Coral TPU Bundle for Raspberry Pi 5](https://pineboards.io/products/hat-ai-dual-edge-coral-tpu-bundle-for-raspberry-pi-5), which unites a Dual Edge Coral TPU for AI/ML/inference to the Raspberry Pi 5, through a PCIe Switch to support both PCIe lanes for both TPUs.

Expand Down
1 change: 1 addition & 0 deletions _hats/pineboards-hatdrive-commander.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,7 @@ link: "https://pineboards.io/products/hatbrick-commander-2-ports-gen2-for-raspbe
videos:
- https://www.youtube.com/watch?v=l30sADfDiM8
- https://www.youtube.com/watch?v=GYN3ub8Qb_I
- https://www.youtube.com/watch?v=HgIMJbN0DS0
redirect_from: /hats/pineberry-pi-hatdrive-commander
---
The Pineboards HatBRICK! Commander splits the Pi's single PCIe Gen 2.0 lane into two PCIe Gen 2.0 ports, via an integrated ASMedia ASM8211e PCIe Gen 2.0 switch chip.
1 change: 1 addition & 0 deletions _hats/pineboards-hatdrive-top.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,7 @@ link: "https://pineboards.io/products/hat-top-2230-2240-for-rpi5"
videos:
- https://www.youtube.com/watch?v=EXWu4SUsaY8
- https://www.youtube.com/watch?v=jjzvh-bfV-E
- https://www.youtube.com/watch?v=HgIMJbN0DS0
redirect_from:
- /hats/mirko-hat5m1-hat
- /hats/pineberry-pi-hatdrive-top
Expand Down

0 comments on commit 6df8fe9

Please sign in to comment.