Popular repositories Loading
-
pascallama.cpp
pascallama.cpp PublicForked from madsteph74/pascallama.cpp
Patch set for llama.cpp — mmq_y=96, #pragma unroll, optimized for Pascal GPUs (SM 6.1)
Shell
-
koboldcpp-avx2-vulkan-numa
koboldcpp-avx2-vulkan-numa PublicForked from LostRuins/koboldcpp
KoboldCpp / llama.cpp speed and performance patches for AVX2 Haswell CPUs and AMD RX580 Polaris (gfx803): Vulkan, NUMA, PGO+LTO Windows builds and measured benchmarks.
C++
-
koboldcpp-media-pipeline
koboldcpp-media-pipeline PublicKoboldCpp portable photo/music API pipeline for RX580 Polaris gfx803 and AVX2: OpenWebUI, YuNet/GPEN/SFace, Gater Restore, Helaman ATD 4x, settings and measured performance.
Python
-
polaris-gfx803-rx580-zluda-rocm-hip-comfyui-windows
polaris-gfx803-rx580-zluda-rocm-hip-comfyui-windows PublicUniversal Windows AI backend toolkit for AMD Polaris / gfx803 / RX580: automatic ZLUDA, HIP SDK, patched rocBLAS + gfx803 blobs, PyTorch, DirectML and OpenVINO CPU; optional ComfyUI with desktop sh…
Python
-
strata-avx1-p40
strata-avx1-p40 PublicForked from Niko1221/Strata
Strata for AVX1-only Xeons and Tesla P40 (sm_61): reviewed performance patches, measured results, and portable Windows releases without model weights.
C++
-
open-webui-ivybridge-avx1
open-webui-ivybridge-avx1 PublicForked from open-webui/open-webui
Open WebUI for Ivy Bridge AVX1: reviewed fixes, LTO experiments, agent compatibility; extended Russian UI translation / дополненный русский перевод.
Python
If the problem persists, check the GitHub status page or contact support.