freenode

← freenode

tensor

AI & ML desk

AI & ML24h ago

PyTorch Inductor corrupts heap on dynamic-shape index_fill path

A fill-order versus stride-order mixup in copy_strided lowering overruns CPU buffers under torch.compile with dynamic shapes.

AI & ML2d ago

PyTorch Dynamo silently drops closure mutations in local class methods

Compiled functions under fullgraph mode lost in-place list and dict updates performed by methods on classes defined inside the region.

AI & ML2d ago

PyTorch CPU fold path eyed for parallel col2im speedups

A proposed change would process output channels concurrently and cut redundant writes on common Fold shapes.

AI & ML5d ago

Transformers adds early Gemma 4 logits and hidden-state support

A broad patch set enables boolean control over retained logits and exposes last hidden states for the model family.

AI & ML6d ago

PyTorch Inductor split-scan sizing bug can silently corrupt GPU memory

Coordinate-descent autotune can pick an R0_BLOCK smaller than workspace allocation assumes, overrunning the buffer under torch.compile.

AI & ML6d ago

llama.cpp rejects Apptainer tools support over host access risks

The proposed server runtime skipped container checks and would let remote clients read host files or pull arbitrary images.

AI & ML7d ago

PyTorch split rewrite breaks vLLM QK-norm RoPE fusion

A functionalization change that turns multi-output splits into slices leaves vLLM’s pattern matcher with nothing to fuse under Inductor graph partition.

AI & ML7d ago

llama.cpp gains opt-in codec to cut recurrent checkpoint RAM

An optional compression path can roughly halve host memory for recurrent-state context checkpoints in the llama.cpp server, remaining off by default.

AI & ML8d ago

llama.cpp Vulkan backend adds Intel Xe flash attention kernels

New kernels for Xe-LPG Plus, Xe2, and Xe3 target faster long-context token generation on recent Intel GPUs.

AI & ML8d ago

PyTorch Inductor returns NaNs on zero-size interpolate

Compiled F.interpolate skips eager's size checks and reads past empty buffers, handing callers silent garbage instead of an error.

AI & ML9d ago

llama.cpp adds SYCL graph record and replay for Intel GPUs

Opt-in graphs on the oneAPI backend show modest decode gains in early Arc tests, with timeouts still under review.

AI & ML9d ago

llama.cpp fixes M-RoPE batch position buffer overread

AddressSanitizer exposed a heap read past an undersized position buffer when batches auto-generate multi-dimensional rotary embeddings.

AI & ML9d ago

PyTorch AI PR review can treat attacker filenames as system text

Untrusted paths from fork pull requests were fed into the model as trusted hook context, enabling prompt injection without model cooperation.

AI & ML9d ago

PyTorch batch-norm stats update overruns undersized tensors

The CPU path indexes running mean and variance by channel count without checking buffer length, causing heap out-of-bounds access.

AI & ML9d ago

PyTorch inductor op builds out-of-bounds CPU tensor views

Unchecked offsets in _reinterpret_tensor let callers produce views that read past storage and trigger heap buffer overflows.

AI & ML9d ago

llama.cpp M-RoPE path overreads batch positions past API docs

Callers who sized the position array to the documented n_tokens still hit a multi-kilobyte overread and silent corruption on multimodal decode.

AI & ML10d ago

Transformers gains DeepSeek-V4.1-Flash text backbone

The addition brings compressed sparse attention, n-gram memory tables, and baked-in activation quantization to the library’s DeepSeek line.

AI & ML11d ago

llama.cpp maintainer blocks new IQ2_NL and IQ3_NL types

Johannes Gaessler rejects a pull request adding CPU quantization formats, citing maintenance burden and machine-generated code.

AI & ML12d ago

Ollama ROCm on Strix Halo returns prior answers after context overflow

On gfx1151 hardware, truncated over-window prompts can replay another request's completion instead of generating a new one.

AI & ML12d ago

PyTorch Inductor to auto-route small decode GEMMs via swap_ab

NVGEMM will match vLLM and SGLang on M<=64 shapes instead of losing to cuBLAS by default.

AI & ML12d ago

llama.cpp GGUF loader accepts tensors whose padded size wraps to zero

Alignment math can overflow so a crafted model file bypasses size checks and understates its true footprint.

AI & ML13d ago

PyTorch Inductor CPU tiled reductions can overflow the heap

Full-width vector stores on tail blocks of 2D-tiled reductions corrupt memory or silently rewrite neighboring output rows.

AI & ML13d ago

Transformers begins FSDP2 and expert parallelism work

The project is adding a two-dimensional device mesh so fully sharded data parallel can run with tensor and expert parallelism.

AI & ML13d ago

PyTorch CUDA cumsum reads past buffer past 2^30 elements

A ones-tensor repro shows an out-of-bounds global load in the GPU prefix-sum path once the length exceeds roughly a billion floats.

AI & ML14d ago

PyTorch feature request seeks decorator API for custom autograd ops

A proposed torch.autograd.op pattern would pair forward math with backward derivatives without class boilerplate.

AI & ML14d ago

llama.cpp adds GLM-5.3-Flash MoE model support

The 320B mixture-of-experts model lands with known decode overhead and no working vision path yet.

AI & ML14d ago

llama.cpp speeds AMD RDNA3/4 inference with Vulkan int8 coopmat

A new cooperative-matrix matmul path in the Vulkan backend lifts prompt and token throughput on Radeon RX 7900-class GPUs.

AI & ML15d ago

llama.cpp WebUI query param enables CSRF tool runs

Crafted links can auto-submit prompts that invoke enabled server tools, including shell execution when confirmation is waived.

AI & ML16d ago

llama.cpp fuses DeepSeek-V4 hyper-connections on Vulkan

Two-node RPC tests show decode and prefill roughly 1.8x faster with lower intermediate memory use.

AI & ML19d ago

llama.cpp adds experimental SM120 CUTLASS MoE prefill

New CUDA kernels target MXFP4 and NVFP4 on SM120 GPUs, while maintainers push for MMVQ refactoring before deeper integration.

AI & ML19d ago

PyTorch adds experimental torch.compile for complex tensors

Complex dtypes can now go through the compiler via a real-valued decomposition path that must be opted in explicitly.

AI & ML19d ago

llama.cpp server gains image-text embeddings for Qwen3-VL

Multimodal embedding and rerank endpoints now support Qwen3-VL models for RAG pipelines that mix pictures and prose.

AI & ML19d ago

PyTorch 2.14 RFC would make CUDA 13.2 the default PyPI build

The proposal keeps CUDA 13.0 as a stable alternate and CUDA 12.6 as legacy while pushing Blackwell-capable 13.2 to pip install torch.

AI & ML19d ago

llama.cpp MoE GPU cache nearly doubles decode on dual Vulkan cards

A host-offloaded expert-weight LRU cache kept hot experts in VRAM and lifted Qwen MoE decode from about 8 to nearly 19 tokens per second on two RX 6950 XTs.

AI & ML19d ago

llama.cpp doubles cold prefill with direct PLE table reads

A direct-read path for lazy PLE tables roughly halves cold-cache prefill on Windows and keeps memory flat on Apple Silicon.

AI & ML19d ago

PyTorch device capability API breaks non-CUDA heuristics

CUDA returns a version tuple while XPU returns a dict, so shared autoheuristic metadata never matches on those backends.

AI & ML20d ago

llama.cpp Docker images lack SemVer tags, miss stable builds

Maintainers will change release packaging so GHCR containers ship with proper version tags when a stable build is cut.

AI & ML20d ago

PyTorch fixes silent out-of-bounds read in compiled kernels

Strided and offset tensor paths in the compiler could read past valid memory without raising an error.

AI & ML20d ago

torch.compile over-reads offset views in Triton kernels

PyTorch can silently return wrong results when compiled kernels skip writing masked output regions.

AI & ML20d ago

PyTorch MPS causal attention leaks future tokens in half precision

On Apple Silicon, scaled_dot_product_attention with is_causal=True can silently ignore the causal mask for float16 and bfloat16.