Knowledge catalogue
LayerStoRm: GLM-5.3-Flash UD-Q4_K_XL (186 GiB) at 1M context on 2× RTX 5090 + 2× RTX 5080 (96 GB VRAM total) using RAM for the pinned experts. LayerStoRm is a (still experimental) MIT-licensed continu
I've built and open-sourced ParaRNN, a PyTorch/Triton library that parallelizes the training of non-linear recurrent architectures across time using Newton-Raphson iterations mapped to associative pre
arXiv:2609.05235v1 Announce Type: cross Abstract: Cryptocurrency markets exhibit extreme volatility and non-stationary dynamics that challenge conventional forecasting methods. Although Large Language
I recently ported VDN-H3 (Video DeltaNet for MiniMax H3) to Vpipe and tested it on an M5 Pro 24GB. The benefit scales with both sequence length and resolution: at 832×480, VDN-H3 starts to pull ahead
I've been tinkering with local LLMs since the beginning of the year when I had an Intel Arc B580 and 32 GB of DDR5. Curiosity got the best of me and I bought the first R9700 about half a year ago, als
TL;DR: switching KV cache to f16 may give a boost in speed if using MTP and ngrams. I have a self-built 'AI mega-cluster' with 2x P40s on a cheap Chinese motherboard and a Xeon CPU (around $1,100 to b
This comparison was requested by a few people, and certainly we were all eager to see the final results. The comparison had taken 11 days and the GPU was crunching numbers for ~167 hours. We've been c
Hi everyone I'm running four rtx 4090, 64GB ram, on a threadripper pro motherboard so all PCIe x16 ports, as a homelab machine for coding. I was migrating from vllm + qwen 3.8 27B (fp8+256k kv cache)
Hiya, NInfer is amazng for Qwen, but lacking for real-world-use. As adoption of issues/pr's was not really what I needed, I created a fork and hit it for this week with 3 concurrent claude code sessio
I've been running Qwen3.8-27B as a local inference server for a production content intelligence pipeline (HVAC industry stuff, lots of long-context retrieval and structured extraction). I have been wa
First, I'd like to thank the Qwen and Unsloth teams for the Qwen3.8 27b UD Q4_K_XL. Fits the poor 24GB of 3090 VRAM with 100k context at Q8 and works phenomenally well! Imho if theres anything that ca
I was running into a lot of posts that praised both the Fixed template and the Sharp template in comparison to the stock one, so I put them to the test. It's not as extensive as it should be for a pap
arXiv:2609.03716v1 Announce Type: new Abstract: Data centers are increasingly optimized by artificial intelligence and, at the same time, increasingly loaded by it. The literature treats these as two
arXiv:2609.03362v1 Announce Type: new Abstract: Grasping and holding tools while using them presents a considerable challenge not only for robots but also for humans. Such a challenge is particularly
arXiv:2609.03497v1 Announce Type: cross Abstract: Developing humanoid robots capable of leveraging human behavioral data is essential for general-purpose embodiment, yet conventional development remai
arXiv:2609.03807v1 Announce Type: cross Abstract: A free pause token gives a language model extra compute to form each next-token prediction (as a pause, or thinking, token does) but carries that comp
arXiv:2609.03889v1 Announce Type: cross Abstract: Contact-rich loco-manipulation requires a bridge between semantic action generation and physical interaction control. Existing Vision-language-action
arXiv:2609.03193v1 Announce Type: cross Abstract: Nested sampling (NS) resolves the thermodynamics of an atomistic system from a single simulation, but its practical reach is limited by the Markov-cha
It sounds like the shift back toward open weight models that I noticed among the startups in the summer batch is a genuine trend. 🦙 @ollama is used by 9 million developers and 85% of the Fortune 500,
I've got an refurb Dell R740 running Proxmox that I put a Tesla T4 in, mainly to run some CTC local transcription work, but thought it would be fun to try DS4 when it came out, and it was appalling at
arXiv:2510.01984v4 Announce Type: replace Abstract: Quadruped mammals coordinate sagittal spinal bending with axial extension and compression during dynamic locomotion. Yet most robotic quadrupeds use
AI factory management is expanding beyond software and servers to encompass the physical systems supporting production workloads. Broadcom Inc. and Super Micro Computer Inc. are integrating their tech
Two weeks, one rule, and hundreds of entries from nearly 50 countries. Here's who took the four titles, and a look at everyone who made the final ten! Entries opened August 20 and closed September 1.
I've tried both and been having this debate with myself for the last few days, on two Asus Ascent GX10s (effectively the same as 2x DGX Spark): DeepSeek-V4-Flash-0731 (official weights) GLM-5.3-Flash
arXiv:2609.01742v1 Announce Type: new Abstract: Anti-UAV systems increasingly fuse multiple sensors, yet their detection heads provide no per-modality reliability signal. This study evaluates whether
arXiv:2609.02222v1 Announce Type: new Abstract: Foot forward kinematics (FK) is widely used to improve proprioceptive legged odometry by providing reliable velocity constraints during foot support. Ex
arXiv:2511.05187v2 Announce Type: replace Abstract: We study whether otherwise-idle inference resources could reduce the scarce-GPU cost of training. Our analysis uses a simulated compute ledger in wh
If you are like me and use Google flow for NB2/NBP there is hope to finally ditch it. Let me start off by saying I’m not very impressed with any of the current local image edit or Image Reference mode
arXiv:2609.02760v1 Announce Type: new Abstract: On-premise assistants can give factory workers conversational access to machine documentation, but models capable of the task rarely fit shop-floor hard
arXiv:2609.01957v1 Announce Type: cross Abstract: Enterprise wireless access points (APs) are promising platforms for predictive machine learning (ML), but their primary responsibility remains providi
Nvidia is announcing its new Personal AI Router (PAIR), a free tool that syncs up your home computers for tackling local AI inference tasks with tools like Ollama and LM Studio. Let's get the obvious
Portable Computer is now available on Linux for @NVIDIA RTX GPUs with 24GB of VRAM or higher. Today we’re launching Portable Computer on @NVIDIA DGX Spark. Portable Computer is a fully local version o
arXiv:2609.01667v1 Announce Type: cross Abstract: Artificial Intelligence has shown to help improve agricultural practices, yet adoption remains limited: 69% of U.S. farmers have privacy concerns with
Quality: 81.1 → 87.7 (+8%) Speed: 35 → 29 tok/s (−16%) Runtime: 8m51s → 44m39s (5x longer) Output tokens: 18K → 78K (🤯) Noticeably better quality, but you're paying for it with tokens and time. Full b
Frontier intelligence is going local. At IFA 2026, NVIDIA, Microsoft and its partners are teaming up to provide faster inference and new tools that make agents easier to set up and run locally on NVID
Hey everyone, Wanted to share a short paper we just published exploring a simple but surprisingly effective optimization for sparse MoE reasoning models. The problem: MoE reasoning models (especially
arXiv:2609.00718v1 Announce Type: new Abstract: Many automobile and mobility companies deploy learned driving policies on embedded computers with limited memory and power. Pruning, knowledge distillat
arXiv:2609.00356v1 Announce Type: new Abstract: Quantum federated learning (QFL) enables clients to train quantum neural network (QNN) models without sharing private data. We find that aggregation in
arXiv:2609.00641v1 Announce Type: new Abstract: Standardized benchmarks have played a central role in advancing robot manipulation learning, yet most focus on ground-supported manipulation systems, wh
AI factory deployment is becoming a workload-sizing exercise rather than a custom infrastructure project. Enterprises need infrastructure spanning edge inference and large-model training. Broadcom Inc
arXiv:2609.00004v1 Announce Type: new Abstract: This paper studies a finite-horizon multi-item capacitated lot-sizing problem in which demand quantities are deterministic, while demand-arrival periods
arXiv:2609.00097v1 Announce Type: cross Abstract: The development of long-context Large Language Models (LLMs) is constrained by the memory bandwidth bottleneck and quadratic complexity of the attenti
arXiv:2609.00404v1 Announce Type: cross Abstract: Robot capability marketplaces, the 'app store for robot skills,' are emerging as the deployment vector for LLM-driven robot fleets. The default cloud-
This was my first serious attempt at tuning a local LLM. I started because Qwen3.8-27B IQ3 was fast on my RTX 5080 but the coding quality disappointed me, and the Q4 profiles I tried in LM Studio were
arXiv:2609.00475v1 Announce Type: cross Abstract: Angle-encoded quantum kernels on tabular data collapse when the feature map is wider than the intrinsic dimension of the data. We propose the correlat
arXiv:2609.01306v1 Announce Type: cross Abstract: Triangle-based neural rendering bridges neural scene representations and conventional graphics pipelines by optimizing explicit geometric primitives c
Alright, I posted that I had my agent test a bunch of different workflows for over 12 hours and got the 'Bro just wasted 12 hours of credits'. It was obvious the proof should come from the visual data
arXiv:2609.00256v1 Announce Type: new Abstract: LLM-based diagnostic systems achieve high semantic accuracy on benchmarks, but open-ended evaluation on clinically uncommon presentations reveals a syst
arXiv:2609.01024v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) architectures scale Large Language Model (LLM) capacity efficiently by activating a sparse subset of experts per token. However
(Paid 5,849 + tax, which came out to 6,279) TL;DR - Bought this 5090 prebuilt and I want to sanity check if I made the right decision and at the right time. Hey everyone. So I want to start off by say
arXiv:2609.01049v1 Announce Type: new Abstract: This paper introduces QXymb, a general framework for constructing observational declarative twins of quantum circuits, and develops QILP-0, its first co
arXiv:2609.01212v1 Announce Type: new Abstract: With the rapid and continuous growth in the incorporation of machine learning models based on the Transformer architecture, capable deployment is in hig
arXiv:2608.30832v1 Announce Type: new Abstract: To improve torque capacity and energy efficiency of humanoid ankles, this paper proposes a 2-DoF parallel elastic actuator (PEA). The main novelty of th
arXiv:2608.29973v1 Announce Type: new Abstract: Cryptocurrency markets generate high-frequency, multi-source data that is expensive to work with unless a team already has commercial-grade streaming an
Anthropic PBC today debuted Claude Fable 5.1 and Claude Mythos 5.1, its most capable large language models to date. The launch comes a day after the company inked a 35 billion infrastructure deal with
arXiv:2608.28909v1 Announce Type: new Abstract: Large-scale, publicly available datasets have driven advances in deep learning, but privacy and legal restrictions often limit data sharing in medical i
arXiv:2602.00514v3 Announce Type: replace Abstract: Tactile feedback is important for contact-rich robotic manipulation, yet effective use of tactile observations remains challenging when tactile sign
arXiv:2608.28824v1 Announce Type: new Abstract: Currently, in biological systems Neural Correlates of Consciousness (NCCs) are characterized in terms of EEG and FMRI signals. Unfortunately, this chara
arXiv:2608.28995v1 Announce Type: cross Abstract: World models let robots imagine possible futures, but exploiting this capability for real-time control is bottlenecked by a representation misalignmen
arXiv:2608.28913v1 Announce Type: new Abstract: High-resolution 3D radar data is scarce. Commodity mmWave sensors use small antenna arrays that limit angular resolution to several degrees, and existin