Pull Requests ggerganov/llama.cpp

server : fix division by zero when reporting stats examples server

#16501 opened 2025-10-10 15:18 by ggerganov

vendor : sync minja

#16500 opened 2025-10-10 14:00 by CISC

Switch to using Ubuntu 25.10 vulkan/mesa devops

#16497 opened 2025-10-10 11:09 by ericcurtin

Properly handle bracket delimiters for LaTeX formulas in chat message output server/webui examples bugfix server

#16496 opened 2025-10-10 10:20 by allozaur

metal : fix mul-mm condition + fix mul-mv permuted kernels ggml Apple Metal

#16494 opened 2025-10-10 07:28 by ggerganov

CUDA: faster tile FA, add oob checks, more HSs Nvidia GPU python ggml

#16492 opened 2025-10-09 22:03 by JohannesGaessler

graph : reuse SSM graphs

#16490 opened 2025-10-09 16:58 by ggerganov

Remove Legacy Copy-OP Pointer Indirection Code Nvidia GPU ggml

#16485 opened 2025-10-09 12:43 by anavp-nvidia

Add AfmoeForCausalLM support python

#16477 opened 2025-10-08 20:58 by bartowski1182

fix: convert_hf_to_gguf - change Jamba non-sentencepiece mode (tokeni… python

#16470 opened 2025-10-08 10:20 by amirai21

opencl: add q8_0 mm support testing ggml OpenCL

#16469 opened 2025-10-08 05:45 by lhez

vulkan: Add State Space Model (SSM) Operations Support Vulkan ggml

#16463 opened 2025-10-07 13:28 by giuseppe

Add hipblasLt implementation for batched gemm to improve performance for CDNA3 only Nvidia GPU ggml

#16457 opened 2025-10-07 07:27 by peizhang56

vulkan: Handle FA with all -inf mask values Vulkan ggml

#16447 opened 2025-10-06 19:00 by jeffbolznv

Metal Pool 1D Kernel testing ggml Apple Metal

#16429 opened 2025-10-05 09:06 by ThoreKoritzius

fix: add generic fallback to detect trailing <think> tags in Jinja templates and handle forced-open reasoning blocks testing

#16426 opened 2025-10-04 19:08 by ServeurpersoCom

Implement llama-pull tool examples

#16423 opened 2025-10-04 17:02 by ericcurtin

contrib : add fish completions via --completion-fish

#16404 opened 2025-10-03 06:46 by g0t4

server / ranking : add sorting and management of top_n examples server

#16403 opened 2025-10-03 06:37 by YannFollet

Add ARANGE Operator to SYCL Backend (Small & Focused Changes) ggml SYCL

#16362 opened 2025-09-30 23:35 by GittyBurstein

feat: render user content as markdown option examples server

#16358 opened 2025-09-30 19:20 by ServeurpersoCom

SYCL SET operator optimized for F32 tensors ggml SYCL

#16350 opened 2025-09-30 12:56 by GittyBurstein

Update build.md documentation

#16346 opened 2025-09-30 06:19 by refine360-debug

Svelte webui model selector examples server

#16335 opened 2025-09-29 16:29 by ServeurpersoCom

ggml-cpu : inspect -march and -mcpu to found the CPU ggml

#16333 opened 2025-09-29 14:05 by angt

Enable per-conversation loading states to allow having parallel conversations examples server

#16327 opened 2025-09-29 10:19 by allozaur

ggml : remove KQ mask padding ggml

#16309 opened 2025-09-28 15:21 by ggerganov

Add a deepwiki badge to auto-refresh the wiki-in-deepwiki weekly.

#16296 opened 2025-09-28 06:14 by 0400H

hip : substituted bpermute ops with swizzle ops (gfx906, maybe all AMD) Nvidia GPU ggml

#16291 opened 2025-09-27 16:10 by iacopPBK

Update convert_hf_to_gguf_update.py python

#16280 opened 2025-09-26 14:04 by cpumaxx