Llama-b9856 Win Cuda 12.4 - Windows Defender claims it's a trojan
Summary
Windows Defender flags a version of Llama-b9856 running on CUDA 12.4 as a trojan, raising security concerns.
Similar Articles
Build 9254 fixes my TG regression and adds PDL for NVIDIA GPUs
Build 9254 of llama.cpp fixes a token generation regression and adds Programmatic Dependent Launch (PDL) support for NVIDIA GPUs, yielding up to 10% speedup in token generation on newer hardware.
LlamaStation v0.9 — llama.cpp GUI for Windows with multi-backend support, TurboQuant, MTP and more
LlamaStation v0.9 is a Windows GUI for llama.cpp that offers a clean interface with full parameter control, multiple backends (official, TurboQuant, AtomicChat, BeeLlama), real-time VRAM monitoring, per-model profiles, voice mode, and headless mode, all without intermediate layers like Ollama.
Tensor split mode: CUDA error on latest llama.cpp with Qwen-3.6-27b
User reports a CUDA error when using tensor split mode with the latest llama.cpp and Qwen-3.6-27b model on dual RTX 3090s with Ubuntu Server 24.04 and Docker.
I have DeepSeek V4 Pro at home
A user demonstrates successfully running the DeepSeek V4 Pro model on a local workstation using a modified llama.cpp CUDA repository, highlighting performance metrics and hardware requirements.
CUDA: add fast walsh-hadamard transform by am17an · Pull Request #23615 · ggml-org/llama.cpp
This pull request adds a fast Walsh-Hadamard transform implementation for CUDA in llama.cpp, a popular open-source LLM inference engine. The optimization enhances performance for certain computational operations on NVIDIA GPUs.