News and insights on edge AI
Releases and notes from the team.
ReleasesQwen 3.5 Optimized with FlashHeadDiscover how Qwen 3.5, optimized with FlashHead, achieves faster inference speeds without sacrificing quality, making it ideal for edge computing.
Embedl··1 min read
ReleasesIntroducing hfviewerDiscover hfviewer, a tool that simplifies model architecture exploration in the Hugging Face ecosystem, enhancing understanding and collaboration in AI development.
Embedl··2 min read
ReleasesFlashHead for vLLM, made simpleDiscover how FlashHead simplifies integration with vLLM, enhancing model performance and reducing latency without complex setups. Explore the benefits now.
Embedl··8 min read
ReleasesLightning-Fast Multimodal Edge Inference with Under 8GB RAMEmbedl is releasing Cosmos Reason2 with FlashHead, a new optimized model that delivers fast multi-modal reasoning on devices with less than 8GB of RAM while maintaining strong reasoning performance.
Embedl··3 min read
ReleasesCosmos Reason 2 Without the Quantization Trade-OffDiscover how our latest model, W4A16-Edge2, achieves full baseline reasoning accuracy for edge deployment without the usual trade-offs in memory and latency.
Embedl··3 min read
ReleasesBlackwell-optimized Cosmos Reason 2New Cosmos Reason 2 variant optimized for NVIDIA Jetson AGX Thor offers efficient NVFP4 quantization, boosting edge AI performance and maintaining reasoning quality. Discover more in our latest release.
Embedl··2 min read
ReleasesCosmos Reason 2, Quantized for the EdgeRelease of optimized Cosmos Reason 2 model for Jetson Nano Super, enhancing multimodal reasoning on constrained edge devices with improved performance and memory efficiency.
Embedl··1 min read
ReleasesUltra-Efficient SLMs: Embedl’s Breakthrough for On-Device AIEmbedl's FlashHead revolutionizes efficient on-device AI by optimizing classification heads in LLMs, leading to faster, scalable, and more accurate models for edge devices.
Embedl··5 min read
ReleasesEDGE AI Talks: Faster Time-To-Device with Embedl HubLearn how Embedl Hub streamlines AI deployment to edge devices by overcoming hardware fragmentation, reducing trial-and-error, and enabling faster, reliable performance profiling.
Embedl··7 min read
ReleasesAmazon releases AZ1 Neural Edge ProcessorOn September 24th, Amazon held its annual hardware event, introducing their new generation of smart home hubs, speakers and other gadgets.
Elaheh Malekzadeh··2 min read