Knowledge catalogue
I maintain xberg, an open-source (MIT) document extraction engine, and v1 is out. Sharing here because a common piece of a local Ollama RAG setup is 'get clean text out of my PDFs/Office files/images,
GPT-Live is so good that I use it almost every day. I've been wanting to replicate it since it was released. My first attempt was to fine-tune Gemma 4 12B to behave like a full-duplex model. Something
TL;DR: I spent 9 days developing a new quantization method for MLX models and measured 18 variants against each other on a single M5 Max MacBook Pro (128 GB). The result is the best-measuring MLX quan
# Running DeepSeek-V4-Flash-0731 (155 GB MoE) on a DGX Spark with vLLM-Moet 2-bit quantization I used Deepseek-v4-Flash-0731 cloud API settig up vllm-moet to run deepseek-v4-flash with MTP locally on
March 6th, 2026 the highest intelligence index score was 51 for frontier models. deepseek-ai/DeepSeek-V4-Flash-0731 that has an intelligence score of 50. If these benchmarks are accurate, models avail
SYCL: add oneMKL GEMM flash attention for XMX-accelerated prompt proc… (#25025) SYCL: add oneMKL GEMM flash attention for XMX-accelerated prompt processing fattn-mkl: fix interleaved dst layout in nor
arXiv:2505.04999v2 Announce Type: replace-cross Abstract: Learning robot control policies from demonstrations typically requires action-labeled expert data, which is expensive to collect through teleo
arXiv:2607.27784v1 Announce Type: new Abstract: Scalable collection of dexterous manipulation demonstrations remains a major bottleneck for robot learning. High-fidelity interfaces often require costl
arXiv:2607.10099v2 Announce Type: replace Abstract: Gust disturbances, dynamic vertical inflow and ground effect are key adverse aerodynamic phenomena that induce variations in the forces acting on a
arXiv:2607.27990v1 Announce Type: cross Abstract: Spiking Neural Networks (SNNs) communicate through sparse binary spike events rather than dense activations, enabling energy-efficient inference on ne
arXiv:2607.28571v1 Announce Type: new Abstract: Operational Earth observation increasingly calls for answering queries such as ``find the image pairs where a new building appeared.'' This means search
arXiv:2607.28097v1 Announce Type: new Abstract: Mathematically equivalent expert-reduction orders can produce observably different sparse-MoE executions. We isolate this effect in native DeepSeek-V4-F
arXiv:2607.27251v1 Announce Type: new Abstract: Transformer-based surrogate models are increasingly used to replace expensive first-principles simulation in engineering design. But conventional transf
A few weeks ago I posted about if it was worth using a V620 for Comfyui, and was told it likely wouldn't work, at least in Windows 11. And if it did, it would be far too slow and unusable. I decided t
arXiv:2607.27713v1 Announce Type: new Abstract: Mobile robots can infer local flow structure from onboard sensing, but a locally plausible estimate is not always safe to write into a global map. Simil
arXiv:2607.26980v1 Announce Type: new Abstract: Sensing ego-velocity estimation is fundamental to state estimation in visually degraded environments, where camera- and LiDAR-based pipelines can become
In today’s agentic era, modern cloud applications are evolving from a set of passive tools to fleets of autonomous digital workers that reason, plan, and take action across a wide range of tasks. For
arXiv:2607.26515v1 Announce Type: new Abstract: We present, to our knowledge, the first end-to-end FP4 RL post-training, in which both the rollout and training policies, including their forward and ba
arXiv:2607.26766v1 Announce Type: cross Abstract: Code search in large-scale ecosystems is often hindered by the lexical gap between user queries and implementation details, alongside the trade-off be
arXiv:2607.26279v1 Announce Type: new Abstract: Collaborative robots are well-suited to maritime missions that benefit from coordination, such as the exploration of unknown reef structures, inspection
arXiv:2607.26370v1 Announce Type: cross Abstract: We propose a self-adaptive online learning for control method for tracking unknown target dynamics. The target dynamics can exhibit switching behavior
arXiv:2607.26121v1 Announce Type: new Abstract: Embodied intelligence integrates learned perception and decision making with real-time computation, control, and physical interaction. Because failures
arXiv:2607.26412v1 Announce Type: new Abstract: Tracking dense, homogeneous targets like schooling fish remains a major challenge for multiple object tracking due to extreme inter-individual homogenei
arXiv:2607.25066v1 Announce Type: new Abstract: Long-horizon LLM agents accumulate reasoning traces, actions, and tool observations that can eventually exceed a model's fixed context window. Existing
arXiv:2607.25145v1 Announce Type: cross Abstract: We implement an agentic AI workflow built around a large language model (LLM) agent for autonomous experiments with nitrogen-vacancy (NV) centers in d
arXiv:2607.24814v1 Announce Type: new Abstract: Access to specialist clinical expertise remains severely limited across sub-Saharan Africa, where physician-to-patient ratios can fall below 1:25,000 in
arXiv:2607.25852v1 Announce Type: new Abstract: Speculative decoding accelerates large language model inference without changing the target distribution, but no single drafting structure performs best
arXiv:2607.26055v1 Announce Type: cross Abstract: Generalist manipulation policies increasingly take the form of action-chunking flow policies built on large pretrained backbones. Such chunks run open
arXiv:2607.24770v1 Announce Type: new Abstract: Procedural tasks such as furniture assembly and home repair impose substantial cognitive demands because users must interpret instructions, track task p
we're experimenting with our own dynamic GGUF quants of kimi k3, made from the original weights with our llama.cpp fork. Q3_K_S is done and works 1114.76 GiB on disk. Q1 and Q2 are in progress, result
Today’s data lakehouse is no longer mere data repository, but increasingly a system of action, actively executing tasks via always-on, autonomous AI agents. Rather than waiting for static reports, the
arXiv:2607.24568v1 Announce Type: new Abstract: Learned feature reweighting can improve automatic modulation classification (AMC) in software, but the same operation introduces additional arithmetic a
arXiv:2607.22720v1 Announce Type: cross Abstract: Existing adaptive inference methods for Large Language Models rely on observational heuristics, such as hidden-state similarity or activation magnitud
arXiv:2607.23044v1 Announce Type: cross Abstract: Quantum devices are open systems whose dynamics interleave coherent evolution with dissipation, and benchmarking, error mitigation, and error correcti
arXiv:2607.23972v1 Announce Type: new Abstract: Color Fundus Photography (CFP) is a primary non-invasive imaging modality for large-scale screening of ophthalmic and systemic diseases. Existing survey
arXiv:2607.23755v1 Announce Type: new Abstract: Robust and accurate pose estimation with multi-modal sensors is fundamental for autonomous vehicles and mobile robotic systems in complex environments.
Hey fellow llamas. we have something new for Strix Halo owners we thought would be useful to share. i'll keep it short: We were able to fit DeepSeek V4 Flash plus its speculative draft on a single Ryz
arXiv:2607.24692v1 Announce Type: cross Abstract: Inference systems increasingly combine a fast path that returns predictions within the application's latency deadline together with a higher-accuracy
arXiv:2607.24585v1 Announce Type: new Abstract: We present ELMOD - Efficient Language Model for On-Device Deployment - a compact (2.7B) German language model designed for efficient inference on resour
arXiv:2607.24259v1 Announce Type: new Abstract: Recent work has explored the use of Large Language Models (LLMs) to automate simulation model building, typically by generating executable code directly
Results: prompt eval: 40 tokens / 97.5s → 0.41 tok/s eval: 400 tokens / 1769.9s → 0.23 tok/s total: 440 tokens / 1867s (31 min) Prompt: 'Write a C++ function that reverses a linked list in place. Expl
arXiv:2607.24727v1 Announce Type: new Abstract: Background. Pediatric musculoskeletal trauma represents up to 18% of pediatric ED visits, yet diagnosis still depends on ionizing radiography. Cumulativ
I've been digging into Krea2 to see if there's any way to condition inference in specific regions of pixel -> latent space to implement bbox type prompting in order to apply multiple simultaneous char
arXiv:2607.22568v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly deployed on mobile and embedded devices to improve privacy and reduce network latency. Yet on-device infer
arXiv:2607.24256v1 Announce Type: cross Abstract: As communication networks adopt virtualized and disaggregated architectures, achieving energy efficiency has become increasingly important for both ec
arXiv:2607.22583v1 Announce Type: new Abstract: Large Language Models (LLMs) have achieved widespread adoption because of their strong reasoning and query-response capabilities. However, deploying the
arXiv:2607.23576v1 Announce Type: new Abstract: Conventional frame-based cameras face significant challenges in detecting objects under high-speed motion blur or in low-light environments. Neuromorphi
arXiv:2607.22786v1 Announce Type: cross Abstract: In this work, we explore how the inference time of a Transformer Neural Network can be efficiently optimized with applications to real-time anomaly de
arXiv:2607.22818v1 Announce Type: cross Abstract: We introduce a new class of fully-quantum Metropolis walks in which both the proposal and acceptance steps are intrinsically quantum. Unlike standard
arXiv:2607.24352v1 Announce Type: new Abstract: The aim of this article is to verify whether integrating large language models (LLMs) with the Retrieval-Augmented Generation (RAG) architecture enables
arXiv:2607.24686v1 Announce Type: cross Abstract: Variational quantum circuits have been central to many proposed near-term applications of quantum computing, but a growing body of evidence suggests t
arXiv:2606.18816v2 Announce Type: replace-cross Abstract: Hybrid brain-computer interfaces (BCIs) that integrate motor imagery (MI) and steady-state visual evoked potentials (SSVEP) provide high-dimen
arXiv:2607.24569v1 Announce Type: cross Abstract: Model-based active flow control requires predictive models that are accurate, stable, and fast enough for real-time optimisation. In controlled wake f
arXiv:2607.22587v1 Announce Type: new Abstract: Large language models (LLMs) achieve strong performance across diverse tasks but their deployment is constrained by the memory and compute cost of their
arXiv:2601.17303v2 Announce Type: replace Abstract: As Industrial Internet of Things (IIoT) environments scale to tens of thousands of connected devices, centralized security architectures introduce l
arXiv:2607.21186v1 Announce Type: cross Abstract: Numerous studies have analyzed the use of hybrid quantum-classical convolutional neural networks as a promising alternative to classical deep learning
arXiv:2607.21680v1 Announce Type: new Abstract: Bridge infrastructure deteriorates gradually, yet its root causes---salt intrusion, freezing, fatigue cracking, and others---remain invisible to the nak
arXiv:2509.14617v4 Announce Type: replace Abstract: Efficient medical data classification is essential for modern disease screening, particularly in resource-constrained environments where power budge
arXiv:2607.21897v1 Announce Type: new Abstract: The rapid advancement of generative models has spurred the critical need to evaluate the worst-case robustness of deepfake detectors. In this paper, we
tldr; we are going to host K3 on A100s (yes, thats correct, we'll try to see if it holds up), H200s & B300s - expect results for A100s & H200s this week while we setup the B300 cluster this weekend &