#edgeai — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #edgeai, aggregated by home.social.
-
Forlinx AM62L32 Local EVM – A low-power industrial SBC powered by TI’s AM62L32 Cortex-A53/M4F SoC
-
NXP FRDM-IMX95-PRO i.MX 95 board features 10GbE, faster 6400 MT/s LPDDR5 memory, dual M.2 expansion
-
Muse Glimmer puts a 30B open-weight agent on consumer hardware. Local execution keeps files off cloud APIs and shifts cost from tokens to hardware.
https://huggingface.co/meta-models/Muse-Glimmer-30B
#EdgeAI #DigitalSovereignty -
Muse Glimmer puts a 30B open-weight agent on consumer hardware. Local execution keeps files off cloud APIs and shifts cost from tokens to hardware.
https://huggingface.co/meta-models/Muse-Glimmer-30B
#EdgeAI #DigitalSovereignty -
Thanks to all the AI haters, many AI providers are cutting prices, making heavy AI users like me smile everyday.
Besides, these AI haters are actually driving AI research and development out of the US into countries like China, India, and even Vietnam. They are also speeding up the development of much smaller #agenticAI and #edgeAI that can even be run on PCs and very soon all smart phones.
AI will replace AI haters sooner, not later.
-
Thanks to all the AI haters, many AI providers are cutting prices, making heavy AI users like me smile everyday.
Besides, these AI haters are actually driving AI research and development out of the US into countries like China, India, and even Vietnam. They are also speeding up the development of much smaller #agenticAI and #edgeAI that can even be run on PCs and very soon all smart phones.
AI will replace AI haters sooner, not later.
-
🤖 #LiquidAI released #LFM2_5 2.6B, an #agentic model that runs entirely on-device: planning, tool calling & multi-step tasks without any cloud API #AI #LLM #EdgeAI #opensource
🧵👇⚡ Decodes 220 tokens/s on an M5 Max CPU, 113 tokens/s on a Ryzen AI Max+ 395 and 30 tokens/s on a phone, staying under 2.5 GB memory
-
🤖 #LiquidAI released #LFM2_5 2.6B, an #agentic model that runs entirely on-device: planning, tool calling & multi-step tasks without any cloud API #AI #LLM #EdgeAI #opensource
🧵👇⚡ Decodes 220 tokens/s on an M5 Max CPU, 113 tokens/s on a Ryzen AI Max+ 395 and 30 tokens/s on a phone, staying under 2.5 GB memory
-
reComputer Mini J501 Edge AI computer features NVIDIA Jetson AGX Orin 64GB/32GB module, two FAKRA camera connectors
-
クラウドだけではないAI推論 ～なぜAIのハードウェア実装が求められているのか～
https://qiita.com/KA026/items/23b0952b181807b5dabd?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items -
28.9M-parameter LLM runs locally on ESP32-S3 at 9 tokens/s
-
28.9M-parameter LLM runs locally on ESP32-S3 at 9 tokens/s
-
$299.99 ASUS UGen300 USB AI accelerator combines 40 TOPS Hailo-10H chip with 8GB LPDDR4
-
NightRun UEFI application boots a local LLM on Raspberry Pi 5 and x86 PCs without an OS
-
#Jetson Orin Nano #Ubuntu is very unfriendly with no good native apps and is relatively slow running most programs. So, it is only good for doing some small LLM inferencing and #edgeAI stuff like #LiDAR which I don't care cause I will be using #esp32 for that purpose.
But the good news is, I can use Pi Apps which uses #Flatpak, allowing some of my favorite apps. like #Obsidian notes, to run on this Jetson board that runs on an ARM 64 version of Ubuntu.
-
#Jetson Orin Nano #Ubuntu is very unfriendly with no good native apps and is relatively slow running most programs. So, it is only good for doing some small LLM inferencing and #edgeAI stuff like #LiDAR which I don't care cause I will be using #esp32 for that purpose.
But the good news is, I can use Pi Apps which uses #Flatpak, allowing some of my favorite apps. like #Obsidian notes, to run on this Jetson board that runs on an ARM 64 version of Ubuntu.
-
Luxonis M8 Controller Box adds industrial I/Os, CAN Bus, and relay control to OAK4 AI vision cameras
-
Every AI API call is a copy of your data leaving the building.
The vivibit E-series keeps it in-house: an E1001 hub plus up to four NVIDIA DGX Spark nodes (1 PFLOPS each) — a private cluster running DeepSeek & Qwen on hardware you own.
No tokens metered. No data shipped out. Just your models, on your rack.
-
Introducing the vivibit E-series — AI compute you own, from desktop to cluster.
Start with one E1001 node: 8-core ARMv9, 48GB LPDDR5, up to 366TB storage — runs 14B models locally at ≤100W.
Scale to a full cluster (E1001 + up to 4 compute nodes) over built-in 50GbE, no data-center switch:
▪ up to 4 PFLOPS AI compute
▪ 512GB unified memory
▪ up to 366TB storage
▪ runs DeepSeek & Qwen on-premYour models. Your data. Your rack.
-
Globalscale Case8 – A MediaTek Genio 520/720 cyberdeck for gaming, home automation, and education (Crowdfunding)
-
reCamera Pro “Open AI Camera” supports computer vision, LLM, VLM, STT, and TTS workloads
