pull-request

Pull request to add architecture support for the cohere2-MoE model to llama.cpp, enabling inference of this Mixture of Experts model.

A pull request for llama.cpp that removes padding and multiple device-to-device copies for Multi-Token Prediction (MTP), improving performance on GPU.

This pull request adds video input support to llama.cpp, enabling multimodal models to process video data via the new mtmd component.

This pull request by ggerganov optimizes kv-cache in llama.cpp to avoid unnecessary copies of kv cells, improving inference performance. It is a contribution to the open-source LLM inference library llama.cpp.

A pull request for llama.cpp ports multi-column MMVQ from CUDA to SYCL, achieving approximately 45% speculative decoding speedup on Intel Arc GPUs.

This pull request adds support for the Granite4 Vision model to llama.cpp, an open-source LLM inference engine.

Adds support for rendering Mermaid diagrams in chat and an interactive preview within the llama.cpp web UI.

Pull request adding support for StepFun 3.5 MTP model in llama.cpp.

llama.cpp is an open-source LLM inference engine in C/C++ with minimal dependencies and support for various hardware and quantization methods. This pull request limits the maximum outputs of llama_context.

Peter Steinberger used Codex to build a fully automated QA bot that automatically generates tests, runs tests after each code commit, and can automatically fix bugs and submit PRs, greatly improving development efficiency.

This pull request for the llama.cpp inference engine implements using f16 mask for Flash Attention to reduce VRAM usage.

This pull request adds tokenizer support for MiniCPM5 to llama.cpp, extending the tool's compatibility with the MiniCPM family of models.

A rejected PR for llama.cpp provides up to 30% faster prompt processing for MOE models on AMD Strix Halo hardware, with gains diminishing at higher context lengths.

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.

A pull request for llama.cpp fixes the constant prompt processing issue that occurs when using OpenCode or Pi with the library.

This pull request optimizes llama.cpp by avoiding unnecessary copying of logits during prompt decode in multi-token prediction, improving inference performance.

The pull request adding MTP (Multi-Token Prediction) support to llama.cpp has been merged into the master branch.

A pull request for MTP (likely a model training pipeline or similar) related to LLaMA models has been merged, marking a milestone.

Pull request adding Multi-Token Prediction (MTP) support to llama.cpp, enabling speculative decoding for faster inference.

An AI agent built with LangChain continuously monitors its own codebase, flags missing monitors, and automatically opens PRs to fix bugs it finds, as described by Alex Shevchenko from Ramp.