Tag
A quantization report for LiquidAI's LFM3-1.7B shows it runs losslessly in 3.6 GB VRAM (2.2 GB with modest degradation) and outperforms MiniCPM5-2B on LiquidAI benchmarks while offering 2x context in half the memory.
Fun fact: you can use agents combined with CreateML to convert any HuggingFace model into an iOS-compatible CoreML model and make it roughly 3x faster.
NVIDIA released Nemotron 3.5 Bolt E4B 4GB, a post-trained Gemma 4 E4B model with multi-turn agentic and conversational capabilities optimized for running on RTX, Jetson, and DGX Spark devices, claiming ~40% higher quality and ~80% fewer output tokens than peers.
该论文提出 HRF-Res，一种用谐振放电（HRF）神经元替代传统 LIF 神经元的液态状态机，使脉冲储层具备频率选择性动力学，在十个基准上达到脉冲储层方法的最先进精度，同时比等效非脉冲振荡储层能耗降低多达 19 倍。
Google reported successfully operating four of its TPUs in orbit, confirming the AI hardware could power on, process a prompt, and transmit a response back to Earth after being launched into space in early October.
Google's Project Suncatcher successfully sent a command to a prototype satellite in orbit, booted up a TPU, ran a prompt, and received a complete LLM-generated haiku response back on Earth — a milestone for running AI models in space.
LiquidAI 的 LFM3-Omni-2B 成为首个可在浏览器中通过 WebGPU 100% 本地运行的全模态模型，无需独立 ASR/TTS，在 M4 Max 上实现约 280 token/s、90ms 首 token 延迟。
A zero-retraining REAP-pruned derivative of NVIDIA's Nemotron-3.5-Transcribe-30B-A4B speech LLM shrinks it from 67.5GB to 10GB (4-bit) while keeping WER at 5.16% vs the teacher's 4.12%, and it runs 27% faster in single-stream throughput on an H200 with vLLM.
A developer built a local GPU benchmark server with two RTX 5060 Ti cards (32GB VRAM) running vLLM to test 12 quantized open-weight models from 4B to 35B across full benchmark datasets, adding wall-clock time efficiency scores and publishing sortable results at llmbenchmarks.io.
Liquid AI 工程师展示 LFM3-Audio 模型在本地设备（一盏灯）上运行并进行语音对话，网友戏称"能否在灯上跑模型"正成为新的基准。
Liquid AI released LFM3 on Hugging Face, a ~2B shared-backbone model family spanning text, vision, audio, omni, and decision models, capable of real-time local video captioning in the browser via WebGPU with Transformers.js.
Liquid AI 发布第三代 Liquid Foundation Models（LFM3），首批包含六个 2B 级模型，覆盖文本、视觉、音频、全模态和决策任务，采用开放权重并共享主干架构，可在手机、笔记本和边缘设备上运行，权重已上线 Hugging Face。
This paper analyzes semantic synchronization of operator manuals over ISO 11783 CAN buses to enable offline, multilingual LLM assistance on agricultural machinery, benchmarking Markdown, JSON, and XML transfer and evaluating open-weight LLMs with retrieval-augmented generation across multiple languages.
该论文提出了一套面向离线边缘生成式AI的连通性体制框架（C0-C4）与推理时信任机制分类法，并在Raspberry Pi 5上以Qwen2.5模型进行本地RAG实证研究，发现签名知识包可被投毒攻击绕过，且现有验证机制无法保证证据真实性。
Liquid AI released LFM3-1.7B, a 1.73B-parameter tri-hybrid text model designed for local agents and edge deployment with a 256K-token context window, fitting in about 1 GB of memory at 4-bit precision and supporting single-GPU fine-tuning for domain specialization.
Unsloth 发布 NVIDIA Nemotron 3.5 Bolt E4B 的 GGUF 量化版本（Q4_K_M），这是一个基于 Gemma-4-E4B-it 进一步后训练的小语言模型（SLM），专为低内存、低首字延迟的本地语音助手与嵌入式/边缘设备场景优化。
WhissleAI released SteerTTS-Edge, a steerable small English TTS model compressed to 11.84 MB via int4 quantization-aware training of the acoustic model and a distilled Vocos vocoder, with unchanged intelligibility (WER 0.96%) and only a ~0.1 UTMOS drop while preserving all 12 steering controls.
AuroraAI-Research released AuroraTTS-100K, an extremely small English text-to-speech model with only 95,638 inference parameters, trained from scratch on LJSpeech using a differentiable harmonic+noise DSP synthesizer instead of a neural vocoder, achieving UTMOS 3.53 with very fast inference.
A Virginia Commonwealth University paper explores extreme low-state weight quantization (3, 5, 7, 16 states) for Transformer-based electricity-load forecasting, showing that quantization-aware training and accumulated-update thresholding can retain predictive performance while reducing weight updates by ~99.7%.
一篇系统综述，涵盖了一百多项关于在 FPGA 和集成电路平台上部署强化学习的研究，重点分析架构设计、性能度量与硬件-算法协同设计。该综述统一比较了 FPGA 与 IC 两种实现路径的资源利用率、延迟、吞吐量、功耗及性能增益，并指出了开放挑战与未来方向。