edge-ai.
6 writings found
Latest Archives
EmbeddingGemma 2: Multimodal AI Finally Goes Truly Local
Google's new 740M parameter embedding model unifies text, images, audio, and video for on-device inference. What this means for privacy-first AI development.
EmbeddingGemma 2: Multimodal AI That Actually Runs on Your Phone
Google's new 740M parameter embedding model unifies text, images, audio, and video on-device. What it means for privacy-first AI development.
Speculative Decoding Brings VLMs to the Edge
Liquid AI's DSpark draft model achieves 3x faster vision-language inference on-device and 2.66x on H100s while adding just 8.9% parameters.
Speculative Decoding Hits Vision Models: What Developers Need to Know
Liquid AI's DSpark draft model accelerates vision-language inference 2-3x on edge devices and 20x on H100s. Here's what it means for your stack.
LFM2.5-2.6B: Small Models Win on Agents
The 2.6B parameter LFM2.5 outperforms models 4x its size on instruction following and tool use. What this means for on-device AI.
Moonshine Voice brings AI speech to 80-cent microcontrollers
An open-source AI toolkit that enables real-time voice agents on embedded systems with minimal RAM requirements. What this means for edge computing.