Alex Chen-Nexa AI
@alexchen4ai
Followers
690
Following
74
Media
6
Statuses
64
Founder and CEO at Nexa AI, Stanford Phd. At the same time, I am also a blogger to share my favorite part of AI: https://t.co/mosYU37y6q
Palo Alto, CA, USA
Joined February 2024
RT @alanzhuly: Today we launch the Public Beta of Hyperlink—an Offline, Private AI Agent built for local files. From hundreds of user conv….
0
2
0
Next year, 100M+ AI PCs will ship—yet most of their AI silicon will sit idle. For months I’ve asked myself: What's the first use-case that proves these chips matter?. My answer: Local-first file search. Knowledge workers waste ~500 hrs/yr chasing insights buried in scattered.
AI PCs are here - yet we still lose 500+hrs/yr chasing scattered files and buried insights. Cloud AI risks your sensitive data. On device AI doesn't. Meet Hyperlink - the fully offline AI agent that instantly searches local folders and unlocks missed ideas with in-text
1
2
5
RT @nexa_ai: AI PCs are here - yet we still lose 500+hrs/yr chasing scattered files and buried insights. Cloud AI risks your sensitive data….
0
32
0
RT @nexa_ai: 👀 Sneak peek: new Nexa SDK Beta. Ship on‑device AI in one command:.• Run MLX models locally.• One‑line GGUF quant selection fr….
0
4
0
RT @nexa_ai: We've been quiet—but busy building. Previously, we created Nexa-SDK and on-device AI models like Octopus-v2, OmniVLM, and Omn….
0
5
0
Excited to combine Google's powerful Gemma model for the on-device audio model with optimized performance and latency. @googledevs.
🎙️📱 Make your apps more conversational and responsive while keeping data private with OmniAudio's on-device audio understanding. Explore the full breakdown → #GoogleDeveloperNews
0
0
1
RT @nexa_ai: 📦 DeepSeek-R1 is making headlines, but quantized DeepSeek-R1-Distill models often lose up to 22% accuracy, significantly under….
0
8
0
Excited to share that we’ve enabled efficient edge AI inference on Qualcomm NPUs, including Snapdragon® X Elite, X Plus, and Gen3! 🚀. Pushing the boundaries of on-device AI with industry-leading hardware partners. More to come—stay tuned! ⚡🙌 #EdgeAI #Innovation.
🎉Nexa Inference Framework now supports deploying LLMs on the Qualcomm NPU—including Snapdragon chipsets!👏Check out our demo video showcasing fast, efficient on-device generative AI experiences.⚡️. Ready to bring the latest Gen AI models to @Qualcomm-powered devices (PC💻,
0
0
0
RT @nexa_ai: Introducing NexaQuant: 🦙 Llama.cpp-Compatible Model Compression with 100%+ Accuracy Recovery. Perfect for bringing Gen AI to….
0
3
0
Couldn't agree more! We Nexa AI have been engaged in the development of on-device generative AI in terms of infrastructure and models for a considerable period. Additionally, we have been involved in the on-device AI developments of numerous companies. A plethora of interesting.
It's that time again - Big Ideas for 2025! My take is that we will see small but mighty on-device AI take off next year. Smaller AI models will dominate in volume and usage. Why? 🧵
0
0
1
RT @nexa_ai: 🔊 Meet OmniAudio-2.6B, the world's fastest and most efficient audio-language model that processes both text and audio inputs.….
0
17
0
Alibaba Group made SOTA open-source models Qwen, and we're the first to bring Qwen2-Audio to run on-device with Nexa SDK. $ nexa run qwen2audio . Use cases, instructions 👉 . Qwen2-Audio processes both audio and text, enabling direct voice interaction.
nexa.ai
Run Qwen2-Audio on edge devices with Nexa SDK
For the first time, you can run @Alibaba_Qwen's Qwen2-Audio on your local device for Voice Chat & Audio Analysis with Nexa SDK - various quantization options available. 🎉. 👉 🔊 Qwen2-Audio is a SOTA small-scale multimodal model that handles audio and
0
0
3
Excited to see more developers using our on-device toolkit! We will releasing more exciting features in the future!!.
We just reached 4,000 stars on GitHub! ⭐️🎉. A huge thanks to our amazing community—your support means the world. Let's keep pushing the boundaries of on-device AI! 🚀. Run GGUF & ONNX models, convert HuggingFace to GGUF, and evaluate models—all with Nexa SDK:
0
0
3
Our omnivision is now trending on huggingface, 3rd on the list. Most importantly, we will iterate this model continuously to improve its capability!.
Experience OmniVision, our sub-billion-parameter multimodal model, delivering seamless AI capabilities on @AMD-powered devices!. Running on the @AMDRyzen AI 9 HX 370, OmniVision showcases:.⚡ Tiny Model Size: only 968M parameters.⚡ Efficient Processing: Powered by AMD iGPU
0
0
3
RT @mervenoyann: OmniVision-968M: a new local VLM for edge devices, fast & small but performant 👏 . it's based on SigLIP-so-400M and Qwen-2….
0
131
0
RT @nexa4ai: 🙌 Nexa SDK now supports Qwen2.5, Qwen2.5-code, Qwen2.5 Math. See changelog for instructions： https://t….
0
4
0