Buy anything from 5,000+ international stores. One checkout price. No surprise fees. Join 2M+ shoppers on Desertcart.
Desertcart purchases this item on your behalf and handles shipping, customs, and support to Argentina.
🚀 Power your edge AI revolution with the Hailo-8 M.2 — where speed meets efficiency!
The Waveshare Hailo-8 M.2 AI Accelerator Module is a compact, high-performance AI processor delivering 26 TOPS of inferencing power at just 2.5W. Compatible with Raspberry Pi 5 and supporting Linux and Windows, it enables real-time, multi-model AI processing on edge devices. Designed for industrial reliability with a wide temperature range, it supports major AI frameworks like TensorFlow and PyTorch, making it the go-to choice for next-gen AI applications.
| ASIN | B0D928WG5L |
| Best Sellers Rank | #50,335 in Computers ( See Top 100 in Computers ) #692 in Single-Board Computers |
| Brand | Waveshare |
| Built-In Media | Hailo-8 AI M.2 Module ×1 |
| Compatible Devices | Raspberry Pi 5 |
| Compatible devices | Raspberry Pi 5 |
| Customer Reviews | 4.4 out of 5 stars 24 Reviews |
| Included components | Hailo-8 AI M.2 Module ×1 |
| Manufacturer | Waveshare |
| Mfr Part Number | Hailo-8 |
| Model Name | Hailo-8 |
| Model Number | Hailo-8 |
| Model name | Hailo-8 |
| RAM Memory Technology | LPDDR4 |
| Ram memory technology | LPDDR4 |
| Unit Count | 1 Count |
D**G
Die Karte läuft wirklich schnell und ist super für kleinere lokale Anwendungen. Ich nutze diesen Chip zum Beispiel für lokale Objekt Erkennung mit mehreren Kameras unter Frigate. Beachtet allerdings dass ihr aktuell noch die Treiber teilweise von Hand Kompiliert werden müssen unter einigen Linux Derivaten. Ebenfalls solltet ihr eine M.2 Adapter Platte / Platine mitbestellen, da die Karte sehr kurz ist und in den meisten Mainboard nur mit Verlängerung bzw. Adapter in die Standoffs der M.2 Slots passt. Die Leistungsaufnahme ist super gering für die Rechenleistung
A**S
Todo perfecto
J**N
So far I am happy with this and Frigate. I was running BlueIris on a PC with a GTX1050 where inference times were around 120-175ms on my cameras. I wanted to switch to Frigate and use the GPU for something else. It's reporting 10-20ms inference times and is noticably faster on notifications. Thats even using the full 2k stream on 2 of my cameras and it identifies people much further away. I had read in forums that people were able to use this with a usb-c adapter but I wasn't able to get that working. I had to swap drives since the mini PC only had 1 nvme and 1 SATA m.2 slot, and this will only work in the nvme slot. Overall I'm happy with it so far.
G**G
Works great
W**Y
Hab die kurze Variante bekommen anstatt die lange, Installation unter proxmox auch extrem schwierig
Trustpilot
1 month ago
5 days ago