Search: “hardware”

I keep testing harness & model setups on real tasks: basic math, vision, computer use (reading emails, navigating stores) and coding. I created a benchmark and a leaderboard. You can see it here: http

Help with this list. Give me your recommendation on 'not supported anymore' GPUs. Looking for budget and Vulkan friendly options. Most of the GPU are not supported by latest CUDA / ROCm. Often with so

Auto-generated synthesis of 2154 entries about r-stablediffusion

I wanted to see how far I could push a fairly ordinary laptop with a huge MoE model. Turns out, Qwen3.8 Flash Next 176B can run on: RTX 3080 Laptop — 16GB VRAM 32GB system RAM SSD No 128GB/256GB RAM w

There seems to be a whole category of extremely narrow inference runtimes appearing: Strata, ninfer, DwarfStar, Splash, llamAmpere, gufo, etc. They deliberately give up the thing llama.cpp/vLLM are gr

arXiv:2610.00053v1 Announce Type: new Abstract: Four-bit floating-point (FP4) Tensor Cores accelerate matrix multiplication, but scale computation, operand packing, layout construction, and saved back

arXiv:2610.00007v1 Announce Type: cross Abstract: Named-entity recognition (NER) is increasingly wanted on-device (no API, low latency, data kept local). The practitioner's question is not the leaderb

arXiv:2610.00649v1 Announce Type: cross Abstract: Synthetic speech detection is critical for audio security, but performance can degrade when labeled data are scarce and evaluation conditions differ f

arXiv:2610.00067v1 Announce Type: new Abstract: Reliable visual inspection is essential for quality assurance in aero-engine blade manufacturing, where defect appearance may vary across production lin

arXiv:2609.39090v1 Announce Type: cross Abstract: Data-driven machine learning (ML) techniques have become an essential tool in many domains of science. Their application to atomistic simulations of m

An AI agent uses compute differently than a traditional application. Over a single session, it writes and runs code, calls a model, and waits on a tool or a person before picking the work back up. As

arXiv:2609.39755v1 Announce Type: new Abstract: Safe and efficient quadruped navigation over unfamiliar terrain requires predicting terrain-robot interaction before contact: geometry and visual appear

So everyone has been yelling about how I should be using Gufo instead of halogen because it's open source and it's 'just as good or better'. Checking in on their GitHub (GitHub.com/gufo-org/gufo) got

arXiv:2609.39350v1 Announce Type: cross Abstract: As model sizes continue to scale, distributed training has become inevitable. Automatic parallelization techniques can derive efficient training paral

arXiv:2609.38202v1 Announce Type: new Abstract: Soft continuum robots can exploit distributed compliance for whole-body manipulation, but synthesizing behavior through changing contacts remains diffic

arXiv:2310.11379v2 Announce Type: replace-cross Abstract: Voice-based interfaces rely on a wake-up word mechanism to initiate communication with devices. However, achieving a robust, energy-efficient,

arXiv:2607.10891v2 Announce Type: replace Abstract: Large language models (LLMs) are rapidly shifting toward agents that solve tasks through diverse interfaces, including web and graphical user interf

40 year tech gap? No problem! The Tandy runs DeskMind, a native DOS program. It talks over WiFi (a PicoMEM 2 card with mTCP) to a small Python server on my PC. That server drives Qwen3.8-27B (NInfer o

arXiv:2609.36993v1 Announce Type: new Abstract: Complex tasks for underwater robots remain limited by the capabilities of their controllers. Learning a better one for a soft, underactuated robotic fis

arXiv:2609.37887v1 Announce Type: new Abstract: Activation and key-value cache precision change what a quantized language model computes without altering its stored weights. Direct weight-code bounds,

arXiv:2505.12683v2 Announce Type: replace Abstract: Key feature fields need bigger embedding dimensionality, others need smaller. This demands automated dimension allocation. Existing approaches, such

arXiv:2609.37072v1 Announce Type: new Abstract: Aerial seed broadcasting can reach remote restoration sites, but provides limited control over seed placement within the soil. This paper presents a geo

arXiv:2601.11433v2 Announce Type: replace Abstract: Deep Differentiable Logic Gate Networks (LGNs) and Lookup Table Networks (LUTNs) are demonstrated to be suitable for the automatic classification of

arXiv:2609.36584v1 Announce Type: new Abstract: Analog in-memory computing (AIMC) offers an alternative for model training by executing matrix operations directly where weights are stored. However, sc

arXiv:2609.37070v1 Announce Type: cross Abstract: Rare but consequential failures can persist in learned locomotion policies for legged robots even when average task performance is high, in part becau

arXiv:2609.35783v1 Announce Type: cross Abstract: Large-scale recommender systems, particularly short-form video platforms, are often bottlenecked by massive popularity feedback loops. In such environ

While AI has made plenty of inroads on people's phones and computers, it's largely failed in dedicated devices. But over the next year, two major AI companies, Meta and OpenAI, will attempt to change

arXiv:2609.31168v1 Announce Type: new Abstract: My doctoral work aims to explore Audio Emotion Recognition (AER) in the context of atypical listening. This research focuses on auditory hypersensitivit

arXiv:2609.30884v1 Announce Type: new Abstract: Large language models rely on KV caching to reduce repeated prefill computation in long context and interactive applications. As lightweight adapters ev

arXiv:2609.31211v1 Announce Type: new Abstract: Underwater robotic inspection depends on acquiring views that reveal task-relevant structure. For a structurally complex coral colony, recognising the t

arXiv:2609.30270v1 Announce Type: new Abstract: On-device inference with small language models keeps user data local, works offline, and incurs no per-query cost, so the on-device tier is preferred wh

arXiv:2603.05069v3 Announce Type: replace Abstract: Personal AI agents face a deployment paradox on mobile: persistent background execution drains the battery and conflicts with platform background li

I've been experimenting with ways to run MoE models that don't fit comfortably in RAM, and I ended up making Overspill, a disk tier for FreeToken. The basic idea came from looking at how Colibri handl

For instance whenever a model comes out, what’s the best engine to run it, the best harness and absolute minimum you need to get same or near same re results that the benchmark of that model claims. A

I've been playing around with Ling 3.0 Tiny, which is an 8 billion parameter model (MoE, 1B active). And I've had a lot of poignant thoughts as a result. Just for fun, I got it running with llama.cpp

Open-source RL envs for the win. On Hugging Face of course! was looking for a quiet weekend but xiaomi dropped their rl envs repo last night to put in perspective, if you have to buy some tasks like t

For the last few days, I've been using two computers to run multiple agents. My 4x3090 machine is running Qwen 3.8 27B with parallel=2, so I can run two agents at the same time. My pi instances are ru

arXiv:2609.29017v1 Announce Type: new Abstract: Most robots manipulate objects solely with their end effectors, whereas humans flexibly leverage different body parts, such as the forearm and elbow, es

arXiv:2609.29163v1 Announce Type: new Abstract: This paper presents an Edge AI-based system for detecting sleep and wake states in non-stationary mobile environments using resource-constrained embedde

The jump from Qwen3 Coder 30B A3B to current-day Qwen 3.6 35B A3B is crazy, especially with all the fine-tunes, and that was around 6 months (I didn't care for local AI back then, or AI at all, apart

Compute (EC2) opened the front door to the first generation cloud. Agents need something different: compute, inference, and data working as one. We built the cloud that delivers all three as one simpl

arXiv:2609.27828v1 Announce Type: cross Abstract: Legacy smart HVAC controllers rely on vendor-cloud TLS secured by ECDH and RSA, both broken by Shor's algorithm, and typical 10-15 year lifespans mean

arXiv:2609.28182v1 Announce Type: new Abstract: Coordinating large population of flexible grid-edge devices can alleviate the need for time-consuming and capital-intensive network upgrades, and AI-bas

The availability of resources for AI workloads can be challenging across the industry, especially accelerators. This can slow your AI workload deployment if it’s built around a specific type of accele

I am looking to hire PhD students, post-docs, and ML researchers in our Paris office interested in collaborating on fundamental research in AI model safety, not safety rules, not AI policy, not post h

arXiv:2609.25047v1 Announce Type: new Abstract: Autonomous agents that automatically build artificial intelligence (AI) models could broaden access to AI across science and engineering. A popular line