Average rating of 5 out of 5 stars
Raspberry Pi AI HAT+ 26 TOPS, Hailo-8 Accelerator
Product number: RPI-AIHAT-26
CHF 110.75
Available · 25 piece · 2-5 Tage
CHF 110.75
Customers also bought
Available · 100+ piece · 2-5 Tage
Average rating of 5 out of 5 stars
Available · 100+ piece · 2-5 Tage
Available · 100+ piece · 2-5 Tage
Average rating of 4.8 out of 5 stars
Available · 100+ piece · 2-5 Tage
Average rating of 4 out of 5 stars
Available · 33 piece · 2-5 Tage
Average rating of 4 out of 5 stars
Available · 77 piece · 2-5 Tage
Available · 100+ piece · 2-5 Tage
Available · 66 piece · 2-5 Tage
Available · 69 piece · 2-5 Tage
Available · 100+ piece · 2-5 Tage
Available · 100+ piece · 2-5 Tage
Available · 100+ piece · 2-5 Tage
Average rating of 4.3 out of 5 stars
Available · 41 piece · 2-5 Tage
Available · 100+ piece · 2-5 Tage
Average rating of 4.8 out of 5 stars
Available · 100+ piece · 2-5 Tage
Average rating of 4.7 out of 5 stars
Available · 100+ piece · 2-5 Tage
Average rating of 5 out of 5 stars
Available · 100+ piece · 2-5 Tage
Available · 100+ piece · 2-5 Tage
Available · 100+ piece · 2-5 Tage
Description
Product information "Raspberry Pi AI HAT+ 26 TOPS, Hailo-8 Accelerator"
Diese Variante des AI HAT+ mit 26 Tera-Operationen pro Sekunde (TOPS) basiert auf dem Hailo-8 Beschleuniger für neuronale Netzwerk-Inferenz. Ideal für leistungsstarke Anwendungen wie Objekterkennung, Segmentierung und Pose-Schätzung. Darüber hinaus kommuniziert es über die PCIe Gen 3 Schnittstelle des Raspberry Pi 5 und wird automatisch erkannt, wenn das aktuelle Raspberry Pi OS verwendet wird.
Eigenschaften
- Hailo-8 Beschleuniger mit 26 TOPS Inferenzleistung
- Vollständig in die Kamerasoftware von Raspberry Pi integriert
- Entspricht der Raspberry Pi HAT+ Spezifikation
- Mit 16mm Stapelhülsen, Abstandshaltern und Schrauben geliefert
- Kompatibel mit dem Raspberry Pi 5 und Active Cooler
- 20 mm x 65 mm Abmessungen
- Arbeitstemperatur: 0°C bis 50°C (Umgebung)
- Unterstützt PCIe Gen 3 Schnittstelle
- Raspberry Pi AI HAT+
- 16mm Stapelhülsen
- Abstandshalter
- Schrauben
Properties
| Zolltarifnummer: | 84733020 |
|---|---|
| Brand: | Raspberry Pi |
| Compatibility: | Raspberry Pi 5 |
| Country of Origin: | China |
| EAN: | 5056561804156 |
| Gross Weight (kg): | 0.034 |
| Manufacturer ID: | SC1791 |
| Product ID: | RPI-AIHAT-26 |
| VPE: | 1 / 200 |
Manufacturer
Raspberry Pi Foundation
Raspberry Pi: A real success story
Hardly any of the developers of the Raspberry Pi would have guessed at the beginning what gigantic popularity their mini computer would one day enjoy. The original motivation of the creators was to counteract the declining number of computer science students at Cambridge University. The previous knowledge among the decreasing number of first-year students was also decreasing noticeably. The remedy was to be a computer that was easy to use and as cheap as possible, with which programming could be learned in a playful way. According to the developers, young people interested in programming are reluctant to experiment with more expensive and complex home computers for fear of possible damage. In 2006, for example, a prototype was already produced with an Atmel ATmega644 microcontroller. However, the actual success story of the Raspberry Pi only began at the beginning of 2012, when the first official Raspberry Pi model came onto the market. By the end of 2019, around 30 million copies of the popular mini computer had been sold.
Raspberry Pi Foundation
Behind the success is, on the one hand, the Raspberry Pi Foundation, a British foundation that is also registered as a charity. The aim of the foundation, whose trustees to this day come at least in part from the Cambridge University environment, is to promote the study of computer science - at universities as well as at schools. Production and distribution of the hardware, on the other hand, is carried out by Raspberry Pi Trading, a company wholly owned by the Raspberry Pi Foundation. The fact that the name "Raspberry" was ultimately chosen for the company is related to the trend in the IT industry at the time to name themselves after fruits - see Apple or Acorn. The suffix "Pi", on the other hand, stands for "Python interpreter" - a reference to the fact that the mini-computer was originally intended to be equipped with a built-in interpreter for the Python programming language. It is no coincidence that the Raspberry Pi is so popular because the compact single-board computer has many talents: on the one hand, it is suitable as an energy-saving desktop thanks to its 700 Mhz ARM CPU. But the Raspberry Pi is also suitable as an inexpensive media centre for the living room: thanks to its HDMI output and the XBMC interface, videos with 1080p resolution can be displayed on a screen as well as on a TV without any problems.
Tech projects with Raspberry Pi
Above all, however, the mini computer is extremely popular with tinkerers and programmers who like to experiment. A wide variety of tech projects can be carried out simply and easily with the help of the Raspberry Pi: For example, the mini computer can be used to create a game console, a robot or a weather station. Perhaps the most characteristic feature of the Raspberry Pi is its compactness: the dimensions of the small computer - depending on the model - are only up to 93.0 mm x 63.5 mm x 20.0 mm. At the same time, the powerful circuit board is only as big as a credit card. The Raspberry Pi is available in various versions and is constantly being optimised and re-released by its creators. At Berrybase, the in-house Raspberry Pi shop, you will not only find all Raspberry Pi models - you will also receive a huge range of Raspberry electronic products for countless areas of application - both from our own production and from established top brands. So nothing stands in the way of your next tech project!
Hersteller- und Sicherheitsinformationen
Raspberry Pi Ltd.
194 Cambridge Science Park, Milton Road
GB CB4 0AB Cambridge
https://www.raspberrypi.com
compliance@raspberrypi.com
WEEE-Nummer: 20453810
Verantwortliche Person für die EU
Raspberry Pi Ireland Limited
3 Dublin Landings
IE D01 C4E0 North Wall Quay, Dublin 1
https://www.raspberrypi.com
compliance@raspberrypi.com
Sicherheitshinweise
- Verwenden Sie das Produkt nur gemäß den Anweisungen des Herstellers.
- Schalten Sie den Raspberry Pi und alle verbundenen Geräte aus, bevor Sie das Produkt installieren oder entfernen.
- Achten Sie darauf, dass keine leitenden Objekte (z. B. Schraubendreher) die Kontakte berühren, um Kurzschlüsse zu vermeiden.
- Verwenden Sie das Produkt in einer trockenen Umgebung und schützen Sie es vor Feuchtigkeit.
- Üben Sie keinen übermäßigen Druck auf Stecker oder Bauteile aus, um Schäden zu vermeiden.
- Lassen Sie Kinder nicht unbeaufsichtigt mit dem Produkt hantieren.
- Beachten Sie die zulässigen Spannungs- und Stromwerte, um Überlastung und Schäden zu vermeiden.
- Lüften Sie das System ausreichend, um Überhitzung zu vermeiden, besonders bei intensiven Anwendungen.
- Entsorgen Sie das Produkt umweltgerecht gemäß den geltenden Vorschriften für Elektronikschrott.
- Akku- oder Batteriekomponenten, sofern vorhanden, müssen separat entsorgt werden.
- Verpackungsmaterialien recyclen, sofern möglich.
- Keine unsachgemäße Entsorgung über den Hausmüll vornehmen.
Reviews 1
1 of 1 reviews
Average rating of 5 out of 5 stars
5 out of 5 stars
Login
1 review
1 October 2025 22:00
Review with rating of 5 out of 5 stars
Sehr gutes Produkt für Computervision
Sehr schnelles Edge-Inferencing bei minimalem Strombedarf, Toolchain (Hailo AI Suite) ist geradezu angenehm simpel. Für produktive Pipelines klar empfehlenswert.
Setup (Use-Case):
Raspberry Pi 5 (PCIe) + Hailo-8; HDMI-Capture 1080p60; ONNX-Modelle (MobileNet-V4-Variante) → Hailo AI Suite (Parser/Optimize/Compiler) → HEF. Einsatz im Echtzeit-Szenario mit strikten Latenzanforderungen (4ms Inferenzzeit).
Wo es hakte (und wie es lösbar war):
Python-Bindings der Runtime: Außerhalb des Containers fehlen teils die HailoRT-Bindings. Lösung: Runtime-Libs/Bindings aus dem Docker extrahieren oder innerhalb des Containers laufen lassen.
Parser-Details: Start/End-Node-Namen müssen eindeutig sein (z. B. input → /classifier/Conv). Ist schnell korrigiert, aber ohne Doku-Blick stolpert man einmal.
Kalibrier-Set-Formate: NPY muss Form und Datentyp exakt treffen (NHWC, uint8). Wenn nicht, vorher konvertieren/splitten – danach ist die Quantisierung unproblematisch.
Versions-Skew: Nicht passende Suite/Runtime-Versionen führen zu wenig hilfreichen Fehlermeldungen. Mit konsistentem Tooling (ein Build-Script) ist das Thema erledigt.
Wenn man sich erstmal eine Struktur geschaffen hat, ist es super simpel eine hef zu generieren. In einem Wochenende hat man sich da komplett reingearbeitet.
Similar Items
Available · 48 piece · 2-5 Tage
Customers also viewed
Available · 100+ piece · 2-5 Tage
Available · 100+ piece · 2-5 Tage
Average rating of 5 out of 5 stars
Available · 100+ piece · 2-5 Tage
Available · 67 piece · 2-5 Tage
Available · 1 piece · 2-5 Tage
Available · 41 piece · 2-5 Tage
Available · 20 piece · 2-5 Tage
Average rating of 5 out of 5 stars
Available · 27 piece · 2-5 Tage
Average rating of 5 out of 5 stars
Available · 48 piece · 2-5 Tage
Available · 55 piece · 2-5 Tage
Available · 11 piece · 2-5 Tage
Available · 100+ piece · 2-5 Tage
Available · 54 piece · 2-5 Tage
Available · 8 piece · 2-5 Tage
Average rating of 0.5 out of 5 stars
Available · 6 piece · 2-5 Tage
Available · 10 piece · 2-5 Tage
Available · 9 piece · 2-5 Tage
Available · 72 piece · 2-5 Tage
Available · 48 piece · 2-5 Tage
Average rating of 4 out of 5 stars
Available · 77 piece · 2-5 Tage