dan

Repositories

1

Projects Packages

29

Public activity Starred repositories

llama-diffusion-cli-spike (pr24423-c3fb9724-avx)

Published 2026-07-04 20:24:36 +00:00 by dan

Installation

Pull the image from the command line:

docker pull forge.coffee-anon.com/dan/llama-diffusion-cli-spike:pr24423-c3fb9724-avx

Digest

sha256:c54f44f9478a51171d09f4b678183f251a2f8f96d19cc7af758965819215696e

For more information on the Container registry, see the documentation.

Image layers

KIWI 10.3.0

RUN /bin/sh -c microdnf -y --nodocs --setopt=install_weak_deps=0 install bash ca-certificates libatomic libstdc++ libgcc vulkan-loader vulkan-loader-devel vulkaninfo mesa-vulkan-drivers radeontop procps-ng && microdnf clean all && rm -rf /var/cache/dnf/* # buildkit

COPY /usr/ /usr/ # buildkit

COPY /usr/local/ /usr/local/ # buildkit

COPY /opt/llama.cpp/build/bin/rpc-* /usr/local/bin/ # buildkit

RUN /bin/sh -c echo "/usr/local/lib" > /etc/ld.so.conf.d/local.conf && echo "/usr/local/lib64" >> /etc/ld.so.conf.d/local.conf && ldconfig && cp -n /usr/local/lib/libllama*.so* /usr/lib64/ 2>/dev/null || true && ldconfig # buildkit

COPY gguf-vram-estimator.py /usr/local/bin/gguf-vram-estimator.py # buildkit

RUN /bin/sh -c chmod +x /usr/local/bin/gguf-vram-estimator.py # buildkit

CMD ["/bin/bash"]

ARG LLAMA_CPP_PR_SHA

LABEL maintainer=citizendaniel

LABEL description=SPIKE ROUND 2: llama-diffusion-cli from llama.cpp PR #24423 (Vulkan-RADV, GGML_CPU_ALL_VARIANTS runtime dispatch) — DiffusionGemma feasibility probe, not a service

LABEL backend=vulkan-radv-diffusion-spike-avx

LABEL gpu.target=gfx1151

LABEL llamacpp.pr=https://github.com/ggml-org/llama.cpp/pull/24423

LABEL llamacpp.sha=

LABEL cpu.dispatch=GGML_CPU_ALL_VARIANTS+GGML_BACKEND_DL runtime-selected (ivybridge on node13tower i7-3770)

COPY /staging/usr/bin/ /usr/bin/ # buildkit

COPY /staging/usr/lib64/ /usr/lib64/ # buildkit

RUN |1 LLAMA_CPP_PR_SHA= /bin/sh -c ldconfig && echo "=== Spike verification (build-time smoke, may still fail on a foreign builder CPU — real check is on node13) ===" && ls -la /usr/bin/llama-diffusion-cli && (/usr/bin/llama-diffusion-cli --version 2>&1 | head -20 || true) # buildkit

ENTRYPOINT []

CMD ["sleep" "infinity"]

Labels

Key	Value
backend	vulkan-radv-diffusion-spike-avx
cpu.dispatch	GGML_CPU_ALL_VARIANTS+GGML_BACKEND_DL runtime-selected (ivybridge on node13tower i7-3770)
description	SPIKE ROUND 2: llama-diffusion-cli from llama.cpp PR #24423 (Vulkan-RADV, GGML_CPU_ALL_VARIANTS runtime dispatch) — DiffusionGemma feasibility probe, not a service
gpu.target	gfx1151
io.buildah.version	1.43.2
license	MIT
llamacpp.pr	https://github.com/ggml-org/llama.cpp/pull/24423
llamacpp.sha
maintainer	citizendaniel
name	fedora-minimal
org.opencontainers.image.license	MIT
org.opencontainers.image.licenses	MIT
org.opencontainers.image.name	fedora-minimal
org.opencontainers.image.title	fedora-minimal
org.opencontainers.image.url	https://fedoraproject.org/
org.opencontainers.image.vendor	Fedora Project
org.opencontainers.image.version	43
vendor	Fedora Project
version	43

Details

Container

2026-07-04 20:24:36 +00:00

9

OCI / Docker

linux/amd64

MIT

Project website

614 MiB

Versions (2) View all

pr24423-c3fb9724-avx 2026-07-04

pr24423-c3fb9724 2026-07-04