Search: “model-releases”

hexagon: fix IM2COL patch-embed DMA ring overflow (#30189) hexagon: fix IM2COL patch-embed DMA ring overflow The exact-tiling (stride == kernel, no pad/dilation) IM2COL DMA kernel issues ICKH DDR->VTC

CUDA: pass src1 precision to host MMQ config helpers (#30168) CUDA: pass src1 precision to host MMQ config helpers Signed-off-by: ynankani ynankani@nvidia.com make src1 prec explicit Signed-off-by: yn

metal: add the 128/96 flash attention kernels, accept every compiled pair (#30209) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/54260937 macOS/iOS: macOS

server: use port 9931 by default (#30159) server: use port 9931 by default revert unrelated changes Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/54269123

ci: fix Models Backend webgpu by supporting GGML_OP_DUP (#30216) The mixed batch path of PR-29622 writes the token rows with set_rows into a dup of the embeddings. WebGPU did not support DUP, so the d

musa: drop mp_21 from the default architectures (#30203) musa: drop mp_21 from the default architectures Signed-off-by: Xiaodong Ye xiaodong.ye@mthreads.com Address review comments Signed-off-by: Xiao

vulkan: fix rms_norm workgroup count overflow (#30145) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/54337638 macOS/iOS: macOS Apple Silicon (arm64) macOS

webgpu: use 2D workgroup dispatch for all the ops which use 1D dispatch (e.g., rms_norm) (#30219) remove 1D workgroups dispatching formatting Website: https://llama.app Attestations: https://github.co

meta : handle host views (#30217) meta : handle views of tensors allocated on the host A view shares the memory of its view_src, so ggml-alloc never allocates a view in the buffer of the split it land

ggml: fix OOB write in ggml_acc with negative offset (#30135) ggml_acc_impl narrowed a size_t offset to int32_t without checking that it fits, so a large offset could truncate to a negative int32_t. T

llama: keep the backend sampling graph static across ubatches (#30223) The reserve builds n_outputs_max_per_seq sampling chains per sampler, while a decode built one per output row, so the graph chang

chat : refactor API (#30210) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/54392536 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, Kle

model : use exact GELU for ModernBERT encoders (#30108) model : use exact GELU for ModernBERT encoders Assisted-by: Codex model : keep tanh GELU aliases on ggml_geglu Assisted-by: Claude Opus 5.5 mode

opencl: fix kernel compilation for a6x GPUs (#30176) opencl: skip kernel_cpy_f32_f32_pack on A6X to avoid shader compiler crash The A6x compiler backend found in iot device with a623 (E031.50.31.01) c

CUDA: Remove redundant CUDA copies after SSM_SCAN (#29807) CUDA: fuse copy of updated state snapshots into recurrent cache with ssm_scan CUDA: remove redundant cuda copies with K==1 (non spec-dec) sce

llama-bench: respect -fitc if bigger than required benchmark size (#28331) Assisted-By: opencode,llama.cpp,Qwen3.6-35B-A3B,Qwen3.8-27B Website: https://llama.app Attestations: https://github.com/ggml-

graph : reorder get_rows for embeddings (#30160) graph : reorder get_rows for embeddings cont : fix gemma4 and improve input embedding construction logic cont : add TODO for lora cont : fix raw embedd

CUDA: fix round issue, under MSVC the CPU and GPU agree (#30229) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/54453988 macOS/iOS: macOS Apple Silicon (ar

vendor: apply deep nested json patch from upstream (#30253) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/54488539 macOS/iOS: macOS Apple Silicon (arm64)

Banger paper from Meta Superintelligence Labs on self-improving agents. (bookmark it) It's hard to know exactly what drives self-improvement, since so many variables are at play (notes, skills, tool c

Banger paper from Meta Superintelligence Labs on user simulators for agent training. Agent RL setups usually let an assistant LLM play the user, so the simulated user is too cooperative and too explic

arXiv:2610.10663v1 Announce Type: cross Abstract: Bioacoustics and ethology encompass a wide range of audio understanding tasks, many of which stand to benefit from recent advances in large audio-lang

arXiv:2501.19065v3 Announce Type: replace-cross Abstract: Long-term time-series forecasting supports a wide range of applications, including weather prediction and electricity demand planning. Frequen

before I joined Anthropic, I spent about 2 weeks hacking on this as a side project with Opus 4. It used the Agent SDK so it needed a constantly running process & didnt work that well but one prompt to

arXiv:2610.11305v1 Announce Type: new Abstract: A language model may revise the same proposition after receiving genuinely relevant evidence or after receiving directional user pressure that adds no r

arXiv:2511.04655v2 Announce Type: replace Abstract: Multimodal LLMs can answer many questions in vision-centric benchmarks without looking at the image, using linguistic priors and skewed answer distr

arXiv:2610.11407v1 Announce Type: cross Abstract: Synthetic tabular generators are commonly optimized for distributional fidelity, but statistical similarity alone does not guarantee preservation of c

arXiv:2610.11087v1 Announce Type: new Abstract: The rapid growth of scientific publishing has strained peer review, particularly in machine learning, raising concerns about declining review quality an

arXiv:2610.11610v1 Announce Type: cross Abstract: Generating ultrasound reports from multiple images requires aggregating clinical evidence across views, yet archived key frames capture only part of t

arXiv:2610.11489v1 Announce Type: new Abstract: Visual instance retrieval often fails when the same object appears at different apparent sizes in the query and gallery. We show that the dominant cause

I made BigMoeOnEdge+ (BigMoeOnEdge-plus), a fork of Helldez's BigMoeOnEdge, the engine that streams MoE experts from flash so models bigger than RAM can run on-device. It adds a full Android chat app

arXiv:2610.11430v1 Announce Type: new Abstract: While domain-specific Large Language Models (LLMs) have encoded vast biomedical knowledge, their limited context windows often hinder a deep understandi

arXiv:2606.22698v2 Announce Type: replace-cross Abstract: As LLM-powered scams proliferate, black-box forensics for conversational LLM agents offers a path to accountability for systems hidden behind

arXiv:2610.11237v1 Announce Type: new Abstract: Group dance generation aims to synthesize coordinated multi-dancer choreography from music, with broad applications in animation and interactive content

arXiv:2610.12452v1 Announce Type: new Abstract: We propose BrickBench, a benchmark for agentic text-conditioned LEGO-set design. Given a prompt, an agent is tasked with producing an assembly that not

Build your own harness, folks. Reading papers like this makes me realize how underexplored harness engineering really is. The authors find that on whole-repository migration, GPT-5.6 Sol goes from 6.5

Bullish on this trend of making post-training more accessible. A new post-training era is upon us. If you work on agentic RL, long-context tasks (a big focus today) are expensive, inefficient, and don

arXiv:2502.17920v2 Announce Type: replace Abstract: Pre-trained visual models have become fundamental in computer vision, but they face challenges in continual learning scenarios where data and tasks

Podcast producer Caloroga Media has launched Linnet, a production studio that runs in a web browser with an artificial intelligence assistant built in. The company developed it from the workflow behin

arXiv:2610.12341v1 Announce Type: new Abstract: Adversarial games have driven advances from heuristic search to reinforcement learning, yet learning and adapting strategies from limited samples remain

arXiv:2610.11901v1 Announce Type: new Abstract: Stance detection requires identifying an author's attitude toward a given target, sometimes based on conversational context. Jev, a specialized decision

arXiv:2610.11692v1 Announce Type: cross Abstract: Foundation models supply reinforcement learning (RL) with priors that mitigate its longstanding weaknesses in sample efficiency and transfer, but thei

arXiv:2610.11963v1 Announce Type: cross Abstract: A no-code fix resolves an invalid bug report by directing the user to change a setting, update to a version where the problem is already fixed, or adj

arXiv:2610.11971v1 Announce Type: cross Abstract: Vision-language-action (VLA) policies assume the embodiment on which they were trained and can fail when a joint fault changes how commanded actions a

arXiv:2610.11165v1 Announce Type: new Abstract: Multivariate long-horizon forecasting is critical to electricity load scheduling and traffic flow management, and to financial risk control. Existing de

arXiv:2610.10597v1 Announce Type: cross Abstract: Public leaderboards for AI models are read continuously, and attackers can see every published standing. Vote rigging, selective disclosure of private