jdelony/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-06-28 04:30:15 -05:00
Code Issues Packages Projects Releases Wiki Activity

Added workflow to build container images (#1279)

Browse Source 
* ci: implement build matrix for CUDA/CPU containers with dynamic tagging

* fix: Updated Docker images/build-container.yml

* fix: Updated the documentation about Docker

* fix: Set Arch for 3090s

* fix: Updated build step name.

* fix: Set target ARCH as a variable

* feat: Added cleanup step

* feat: Added docker-bake and updated workflow

* fix: Issue with REPO_OWNER variable

* fix: Updated workflow to solve errors

* fix: Updated branch format

* fix: Wrong naming

* Update docker-bake.hcl

* Update build-container.yml

* Update ik_llama-cuda.Containerfile

* Update ik_llama-cpu.Containerfile

* Update docker-bake.hcl

* Update build-container.yml

* Removed action/cache

* added -sSL for reliability and fixed the URL path

* added -sSL for reliability and fixed the URL path CUDA containerfile

* fix: correct Dockerfile RUN command syntax errors

- Combine split apt-get install commands in both Containerfiles
- Fix broken cmake command continuation in ik_llama-cuda.Containerfile

* fix: correct llama-swap download URL in Containerfiles

- Fix broken line continuation in curl download URL for llama-swap

* perf: improve ccache configuration in Containerfiles

- Add CCACHE_UMASK=000 for cache accessibility across stages
- Add CCACHE_MAXSIZE=1G to prevent unbounded growth
- Initialize ccache with ccache -i during build stage

* fix: remove problematic ccache initialization from Containerfiles

- ccache -i fails because CCACHE_DIR mount doesn't exist yet at build time

* fix: add git to CPU Containerfile build dependencies

- Resolves CMake warning about missing Git for build info

* chore: optimize Containerfile with smaller images and better healthchecks

- Add --no-install-recommends to all apt-get commands for smaller image size
- Add ca-certificates to base stage for HTTPS support
- Merge redundant build copy commands from 3 layers to 1
- Fix llama-swap version from 198 to v199 (latest release)
- Add HEALTHCHECK configuration with interval/timeout/retries to server and swap stages
- Copy /app/lib in server stage to fix container startup

* chore: fix CUDA Containerfile healthchecks and swap version

- Add /app/lib copy in server stage to fix container startup
- Fix llama-swap version from 198 to v199 (latest release)
- Add HEALTHCHECK configuration with interval/timeout/retries

* chore: fix indentation in Containerfiles and add LD_LIBRARY_PATH for server target

* fix: add --break-system-packages flag for pip in CPU Containerfile

* feat: add git bind mount for build info and NCCL support for CUDA

* fix: remove libnccl-dev from CUDA build (already included in base image)

* fix: added Markdown files to ignore files

* feat: use BUILD_NUMBER-COMMIT pattern for docker image tags

- Add BUILD_NUMBER and LLAMA_COMMIT to build workflow
- Update docker-bake.hcl to use version tag format matching llama-server --version output
- Format: VARIANT-BUILD_NUMBER-COMMIT (e.g., cu12-full-4406-3bc90dfd)

* fix: fetch full git history for accurate BUILD_NUMBER

- Add fetch-depth: 0 to actions/checkout to get all commits
- This ensures git rev-list --count HEAD returns correct total commit count

* fix: fetch full git history in Dockerfile for accurate BUILD_NUMBER

- Add git fetch --unshallow to get complete commit history during build
- This ensures build-info.cpp is generated with correct LLAMA_BUILD_NUMBER

* chore: update GitHub Actions to latest versions for Node.js 24 compatibility

- docker/setup-buildx-action@v3 -> v4
- docker/login-action@v3 -> v4

* chore: update all GitHub Actions to Node.js 24 compatible versions

- actions/checkout@v4 -> v6
- docker/setup-buildx-action@v3 -> v5
- docker/login-action@v3 -> v6
- docker/bake-action@v5 -> v7

* fix: use CI-passed BUILD_NUMBER and LLAMA_COMMIT in Dockerfile

- Add BUILD_NUMBER and LLAMA_COMMIT as build args
- Fall back to git commands if not provided
- Pass values explicitly to cmake for accurate build info

* fix: pass BUILD_NUMBER and LLAMA_COMMIT as Docker build args

- Add BUILD_NUMBER and LLAMA_COMMIT to docker bake args
- These will be used by the Containerfile for accurate build info

* fix: revert docker actions to v4 (latest available versions)

* fix: calculate BUILD_NUMBER and LLAMA_COMMIT directly in Containerfile

- Removed ARG defaults since we calculate from git during build
- Use git rev-list --count HEAD and git rev-parse for accurate version info
- Falls back to 0/unknown if git commands fail

* feat: calculate BUILD_NUMBER and LLAMA_COMMIT in Containerfiles

- Add git-based version calculation in both CPU and CUDA Containerfiles
- Remove .git bind mount (git is copied with COPY .)
- Pass build info to CMake for accurate llama-server --version output

* feat: calculate BUILD_NUMBER and LLAMA_COMMIT in Containerfiles

- Add git-based version calculation using git rev-list and git rev-parse
- Copy .git directory separately to ensure git commands work during build
- Pass build info to CMake for accurate llama-server --version output

* fix: cache improvements for CUDA and CPU builds

* fix: "/.git": not found

* fix: Unnecessary mv llama-swap

* fix: Remove BUILD_NUMBER and LLAMA_COMMIT from docker file, calculated by cmake proc

* fix: remove .git from dockerignore for local and CI builds

- Enables cmake to access .git directory during Docker build
- Required for version calculation in llama-server binary
- GitHub Actions uses explicit mount via bake action set parameter

* fix: Remove mounts key from Build and Push step in gh workflow

* ci: add .git verification step before build

* refactor: standardize Containerfile structure and remove .git mount dependency

- Remove --mount=type=bind,source=.git,target=.git from both Containerfiles
- Replace COPY . . with git clone for cleaner build context
- Add CUSTOM_COMMIT ARG for optional custom commit switching
- Standardize ARG/ENV ordering and comment formatting across CPU/CUDA variants
- Install ca-certificates before git clone to fix SSL verification issues
- Rename 'Structured artifact collection' to 'Collect build artifacts'

* ci: remove broken cache pruning step

* ci: remove broken prune-cache job

- Remove prune-cache job that was failing due to missing .git directory
- The job required a checkout step and the cache pruning logic was non-critical

* chore: Removed step for Verifying .git existance in GH workflow

* fix: ensure build always proceeds even if git switch fails

- Add '|| true' to git switch command so build continues on failure
- This prevents the entire RUN step from failing when CUSTOM_COMMIT is invalid

* fix: resolve Docker build pipeline issues

- Remove external git clone from Containerfiles, use build context directly
- Add BUILD_NUMBER and BUILD_COMMIT as CMake cache variables in build-info.cmake
- Fix .devops/tools.sh inclusion by using explicit COPY for hidden directories
- Set USE_CCACHE=true for CI builds
- Clean up unused SHA_SHORT variable from docker-bake.hcl

Fixes: Build steps were cached incorrectly due to external git clone ignoring the actual build context source.

* fix: include .git in Docker build context and add verification

* ci: add .git directory verification step after checkout

* build: fix .git mount path for Docker build context compatibility

* build: fix .git mount path for Docker build context compatibility

* docker: include .git in build context for version calculation

* ci: add .git directory verification step after checkout

* chore: Removed unecessary Verify .git step (It was a test)

* docs: update README with docker-bake and build-local.sh instructions

* docs: remove build-local.sh reference (not in repo)

* ci: optimize disk usage by limiting fetch depth and cleaning workspace

---------

Co-authored-by: HP Prodesk <sourceupdev@gmail.com>
This commit is contained in:
Yadir Hernandez Batista 2026-04-10 02:06:47 -04:00 committed by GitHub
parent 13d7178db9
commit db31e7d803
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
7 changed files with 422 additions and 172 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
.dockerignore
View File

@ -1,13 +1,15 @@
*.o
*.a
*.md
.cache/
.git/
# Ensure .git is NOT ignored so it can be mounted/copied
!.git
.github/
.gitignore
.vs/
.vscode/
.DS_Store
build*/
models/*
@ -18,3 +20,5 @@ models/*
arm_neon.h
compile_commands.json
Dockerfile
**/*.md

118
.github/workflows/build-container.yml vendored Normal file
View File

@ -0,0 +1,118 @@
name: Build and Push Docker Image
on:
push:
branches:
- main
permissions:
contents: read
packages: write
actions: read
jobs:
build-and-push:
runs-on: ubuntu-latest
strategy:
fail-fast: false
matrix:
include:
- variant: "cu12"
cuda_version: "12.6.2"
containerfile: "ik_llama-cuda.Containerfile"
- variant: "cu13"
cuda_version: "13.1.1"
containerfile: "ik_llama-cuda.Containerfile"
- variant: "cpu"
cuda_version: "none"
containerfile: "ik_llama-cpu.Containerfile"
steps:
- name: Checkout
uses: actions/checkout@v4
with:
fetch-depth: 100 # Enough for rev-list, but saves GBs of history
- name: Free Disk Space (Ubuntu)
run: |
echo "Listing initial disk usage..."
df -h
sudo rm -rf /usr/share/dotnet
sudo rm -rf /usr/local/lib/android
sudo rm -rf /opt/ghc
sudo rm -rf "/usr/local/share/boost"
sudo rm -rf /usr/lib/jvm
sudo docker image prune -af
echo "Listing disk usage after cleanup..."
df -h
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v4
- name: Log in to GHCR
uses: docker/login-action@v4
with:
registry: ghcr.io
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: Prepare Environment
id: prep
run: |
echo "BUILD_NUMBER=$(git rev-list --count HEAD)" >> $GITHUB_ENV
echo "LLAMA_COMMIT=$(git rev-parse --short HEAD)" >> $GITHUB_ENV
echo "REPO_LOWER=$(echo ${{ github.repository_owner }} | tr '[:upper:]' '[:lower:]')" >> $GITHUB_ENV
# 5.1 Restore the cache from GitHub's storage to a host folder
- name: Cache ccache
uses: actions/cache@v4
with:
path: .buildkit-cache
key: ccache-${{ matrix.variant }}-${{ github.run_id }}
restore-keys: |
ccache-${{ matrix.variant }}-
# 5.2. "Inject" that host folder into BuildKit's internal mount system
- name: Inject ccache into BuildKit
uses: reproducible-containers/buildkit-cache-dance@v3
with:
cache-map: |
{
".buildkit-cache": "/ccache"
}
skip-extraction: ${{ github.event_name == 'pull_request' }}
# 5.3 Build and push using the cache
- name: Build and Push
uses: docker/bake-action@v7
env:
REPO_OWNER: ${{ env.REPO_LOWER }}
VARIANT: ${{ matrix.variant }}
BUILD_NUMBER: ${{ env.BUILD_NUMBER }}
LLAMA_COMMIT: ${{ env.LLAMA_COMMIT }}
CUDA_VERSION: ${{ matrix.cuda_version }}
GGML_NATIVE: "OFF" # Force OFF for CI portability
USE_CCACHE: "true"
with:
push: true
files: ./docker-bake.hcl
set: |
*.context=.
*.dockerfile=./docker/${{ matrix.containerfile }}
*.cache-from=type=gha,scope=ccache-${{ matrix.variant }}
*.cache-to=type=gha,mode=max,scope=ccache-${{ matrix.variant }}
source: .
cleanup:
runs-on: ubuntu-latest
needs: build-and-push
if: success()
steps:
- name: Delete untagged images
uses: vlaurin/action-ghcr-prune@v0.6.0
with:
token: ${{ secrets.GITHUB_TOKEN }}
organization: ${{ github.repository_owner }}
container: ik-llama-cpp
keep-younger-than: 0
untagged-only: true

58
docker-bake.hcl Normal file
View File

@ -0,0 +1,58 @@
variable "REPO_OWNER" { default = "local" }
variable "VARIANT" { default = "cpu" }
variable "BUILD_NUMBER" { default = "0" }
variable "LLAMA_COMMIT" { default = "local-dev" }
variable "CUDA_VERSION" {}
variable "CUDA_DOCKER_ARCH" { default = "86;90" }
variable "USE_CCACHE" { default = "true" }
variable "GGML_NATIVE" { default = "ON" }
# Common cache configuration for GitHub Actions
target "cache_settings" {
cache-from = ["type=gha,scope=ccache-${VARIANT}"]
cache-to = ["type=gha,mode=max,scope=ccache-${VARIANT}"]
}
group "default" {
targets = ["server", "full", "swap"]
}
target "settings" {
context = "."
inherits = ["cache_settings"]
args = {
BUILD_NUMBER = "${BUILD_NUMBER}"
BUILD_COMMIT = "${LLAMA_COMMIT}"
CUDA_VERSION = "${CUDA_VERSION}"
CUDA_DOCKER_ARCH = "${CUDA_DOCKER_ARCH}"
GGML_NATIVE = "${GGML_NATIVE}"
USE_CCACHE = "${USE_CCACHE}"
}
}
target "server" {
inherits = ["settings"]
target = "server"
tags = [
"ghcr.io/${REPO_OWNER}/ik-llama-cpp:${VARIANT}-server-${BUILD_NUMBER}",
"ghcr.io/${REPO_OWNER}/ik-llama-cpp:${VARIANT}-server"
]
}
target "full" {
inherits = ["settings"]
target = "full"
tags = [
"ghcr.io/${REPO_OWNER}/ik-llama-cpp:${VARIANT}-full-${BUILD_NUMBER}-${LLAMA_COMMIT}",
"ghcr.io/${REPO_OWNER}/ik-llama-cpp:${VARIANT}-full"
]
}
target "swap" {
inherits = ["settings"]
target = "swap"
tags = [
"ghcr.io/${REPO_OWNER}/ik-llama-cpp:${VARIANT}-swap-${BUILD_NUMBER}-${LLAMA_COMMIT}",
"ghcr.io/${REPO_OWNER}/ik-llama-cpp:${VARIANT}-swap"
]
}

15
docker-bake.override.hcl Normal file
View File

@ -0,0 +1,15 @@
# Local development override - automatically sets BUILD_NUMBER and BUILD_COMMIT
variable "BUILD_NUMBER" { default = "0" }
variable "BUILD_COMMIT" { default = "local-dev" }
target "server" {
dockerfile = "./docker/ik_llama-cpu.Containerfile"
}
target "swap" {
dockerfile = "./docker/ik_llama-cpu.Containerfile"
}
target "full" {
dockerfile = "./docker/ik_llama-cpu.Containerfile"
}

140
docker/README.md
View File

@ -4,7 +4,7 @@ Built on top of [ikawrakow/ik_llama.cpp](https://github.com/ikawrakow/ik_llama.c
All commands are provided for Podman and Docker.
CPU or CUDA sections under [Build](#Build) and [Run]($Run) are enough to get up and running.
CPU or CUDA sections under [Build](#Build) and [Run]($Run) are enough to get up and running.
## Overview
@ -14,98 +14,100 @@ CPU or CUDA sections under [Build](#Build) and [Run]($Run) are enough to get up
- [Extra Features](#Extra)
- [Credits](#Credits)
# Build
## Build
Builds two image tags:
### Using docker-bake (Recommended)
- `swap`: Includes only `llama-swap` and `llama-server`.
- `full`: Includes `llama-server`, `llama-quantize`, and other utilities.
The project uses Docker Bake for building multiple targets efficiently.
Start: download the 4 files to a new directory (e.g. `~/ik_llama/`) then follow the next steps.
#### CPU Variant
```
└── ik_llama
├── ik_llama-cpu.Containerfile
├── ik_llama-cpu-swap.config.yaml
├── ik_llama-cuda.Containerfile
└── ik_llama-cuda-swap.config.yaml
```bash
docker buildx bake --builder ik-llama-builder full swap
```
## CPU
Or with custom tags:
```
podman image build --format Dockerfile --file ik_llama-cpu.Containerfile --target full --tag ik_llama-cpu:full && podman image build --format Dockerfile --file ik_llama-cpu.Containerfile --target swap --tag ik_llama-cpu:swap
```bash
REPO_OWNER=yourname docker buildx bake --builder ik-llama-builder \
-f ./docker-bake.hcl \
full swap
```
```
docker image build --file ik_llama-cpu.Containerfile --target full --tag ik_llama-cpu:full . && docker image build --file ik_llama-cpu.Containerfile --target swap --tag ik_llama-cpu:swap .
#### CUDA Variant
First, set the CUDA version and GPU architecture in `ik_llama-cuda.Containerfile`:
- `CUDA_DOCKER_ARCH`: Your GPU's compute capability (e.g., `86` for RTX 30*, `89` for RTX 40*, `12.0` for RTX 50*)
- `CUDA_VERSION`: CUDA Toolkit version (e.g., `12.6.2`, `13.1.1`)
```bash
VARIANT=cu12 docker buildx bake --builder ik-llama-builder full swap
```
## CUDA
### Build Targets
```
podman image build --format Dockerfile --file ik_llama-cuda.Containerfile --target full --tag ik_llama-cuda:full && podman image build --format Dockerfile --file ik_llama-cuda.Containerfile --target swap --tag ik_llama-cuda:swap
```
Builds two image tags per variant:
```
docker image build --file ik_llama-cuda.Containerfile --target full --tag ik_llama-cuda:full . && docker image build --file ik_llama-cuda.Containerfile --target swap --tag ik_llama-cuda:swap .
```
- **`full`**: Includes `llama-server`, `llama-quantize`, and other utilities.
- **`swap`**: Includes only `llama-swap` and `llama-server`.
# Run
### Local Development
- Download `.gguf` model files to your favorite directory (e.g. `/my_local_files/gguf`).
1. Clone the repository: `git clone https://github.com/ikawrakow/ik_llama.cpp`
2. Enter the repo: `cd ik_llama.cpp`
3. Use either docker-bake or build-local.sh as shown above.
## Run
- Download `.gguf` model files to your favorite directory (e.g., `/my_local_files/gguf`).
- Map it to `/models` inside the container.
- Open browser `http://localhost:9292` and enjoy the features.
- API endpoints are available at `http://localhost:9292/v1` for use in other applications.
## CPU
### CPU
```
```bash
podman run -it --name ik_llama --rm -p 9292:8080 -v /my_local_files/gguf:/models:ro localhost/ik_llama-cpu:swap
```
```
docker run -it --name ik_llama --rm -p 9292:8080 -v /my_local_files/gguf:/models:ro ik_llama-cpu:swap
```bash
docker run -it --name ik_llama --rm -p 9292:8080 -v /my_local_files/gguf:/models:ro localhost/ik_llama-cpu:swap
```
## CUDA
### CUDA
- Install Nvidia Drivers and CUDA on the host.
- For Docker, install [NVIDIA Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html)
- For Podman, install [CDI Container Device Interface](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/cdi-support.html)
- Identify for your GPU:
- [CUDA GPU Compute Capability](https://developer.nvidia.com/cuda/gpus) (e.g. `8.6` for `RTX30*0`, `8.9` for `RTX40*0`, `12.0` for `RTX50*0`) then change `CUDA_DOCKER_ARCH` in `ik_llama-cuda.Containerfile` to your GPU architecture (e.g. `CUDA_DOCKER_ARCH=86` for `RTX30*0`, `CUDA_DOCKER_ARCH=89` for `RTX40*0`, `CUDA_DOCKER_ARCH=120` for `RTX50*0`). If you have a mix of different GPUs add them like `CUDA_DOCKER_ARCH=86;89;120`).
- [CUDA Toolkit supported version](https://developer.nvidia.com/cuda-toolkit-archive) then adjust `CUDA_VERSION` in `ik_llama-cuda.Containerfile` to your GPU (e.g. `CUDA_VERSION=13.1` for `RTX50*0`).
- Identify your GPU:
- [CUDA GPU Compute Capability](https://developer.nvidia.com/cuda/gpus) (e.g., `8.6` for RTX30*, `8.9` for RTX40*, `12.0` for RTX50*)
- [CUDA Toolkit supported version](https://developer.nvidia.com/cuda-toolkit-archive)
```
```bash
podman run -it --name ik_llama --rm -p 9292:8080 -v /my_local_files/gguf:/models:ro --device nvidia.com/gpu=all --security-opt=label=disable localhost/ik_llama-cuda:swap
```
```
docker run -it --name ik_llama --rm -p 9292:8080 -v /my_local_files/gguf:/models:ro --runtime nvidia ik_llama-cuda:swap
```bash
docker run -it --name ik_llama --rm -p 9292:8080 -v /my_local_files/gguf:/models:ro --runtime nvidia localhost/ik_llama-cuda:swap
```
# Troubleshooting
## Troubleshooting
- If CUDA is not available, use `ik_llama-cpu` instead.
- If models are not found, ensure you mount the correct directory: `-v /my_local_files/gguf:/models:ro`
- If you need to install `podman` or `docker` follow the [Podman Installation](https://podman.io/docs/installation) or [Install Docker Engine](https://docs.docker.com/engine/install) for your OS.
# Extra
## Extra
- `CUSTOM_COMMIT` can be used to build a specific `ik_llama.cpp` commit (e.g. `1ec12b8`).
- **Custom commit**: Build a specific `ik_llama.cpp` commit by modifying the Containerfile or using build args.
```
podman image build --format Dockerfile --file ik_llama-cpu.Containerfile --target full --build-arg CUSTOM_COMMIT="1ec12b8" --tag ik_llama-cpu-1ec12b8:full && podman image build --format Dockerfile --file ik_llama-cpu.Containerfile --target swap --build-arg CUSTOM_COMMIT="1ec12b8" --tag ik_llama-cpu-1ec12b8:swap
```bash
docker buildx bake --builder ik-llama-builder --set full.args.BUILD_COMMIT=1ec12b8 full
```
```
docker image build --file ik_llama-cuda.Containerfile --target full --build-arg CUSTOM_COMMIT="1ec12b8" --tag ik_llama-cuda-1ec12b8:full . && docker image build --file ik_llama-cuda.Containerfile --target swap --build-arg CUSTOM_COMMIT="1ec12b8" --tag ik_llama-cuda-1ec12b8:swap .
```
- **Using the tools in the `full` image**:
- Using the tools in the `full` image:
```
```bash
$ podman run -it --name ik_llama_full --rm -v /my_local_files/gguf:/models:ro --entrypoint bash localhost/ik_llama-cpu:full
# ./llama-quantize ...
# python3 gguf-py/scripts/gguf_dump.py ...
@ -113,28 +115,40 @@ $ podman run -it --name ik_llama_full --rm -v /my_local_files/gguf:/models:ro --
# ./llama-sweep-bench ...
```
```
docker run -it --name ik_llama_full --rm -v /my_local_files/gguf:/models:ro --runtime nvidia --entrypoint bash ik_llama-cuda:full
```bash
docker run -it --name ik_llama_full --rm -v /my_local_files/gguf:/models:ro --runtime nvidia --entrypoint bash localhost/ik_llama-cuda:full
# ./llama-quantize ...
# python3 gguf-py/scripts/gguf_dump.py ...
# ./llama-perplexity ...
# ./llama-sweep-bench ...
```
- Customize `llama-swap` config: save the `ik_llama-cpu-swap.config.yaml` or `ik_llama-cuda-swap.config.yaml` localy (e.g. under `/my_local_files/`) then map it to `/app/config.yaml` inside the container appending `-v /my_local_files/ik_llama-cpu-swap.config.yaml:/app/config.yaml:ro` to your`podman run ...` or `docker run ...`.
- To run the container in background, replace `-it` with `-d`: `podman run -d ...` or `docker run -d ...`. To stop it: `podman stop ik_llama` or `docker stop ik_llama`.
- If you build the image on a diferent machine, change `-DGGML_NATIVE=ON` to `-DGGML_NATIVE=OFF` in the `.Containerfile`.
- If you build only for your GPU architecture and want to make use of more KV quantization types, build with `-DGGML_IQK_FA_ALL_QUANTS=ON`.
- If you experiment with several `CUDA_VERSION`, remember to identify with `podman image ls` or `docker image ls` then delete (e.g.`podman image rm docker.io/nvidia/cuda:12.4.0-runtime-ubuntu22.04 && podman image rm docker.io/nvidia/cuda:12.4.0-devel-ubuntu22.04` or `docker image rm docker.io/nvidia/cuda:12.4.0-runtime-ubuntu22.04 && docker image rm docker.io/nvidia/cuda:12.4.0-devel-ubuntu22.04` the unused images as they have several GB.
- If you want to build without `llama-swap`, change `--target swap` to `--target server` in `ik_llama Containerfiles`, e.g. `docker image build --file ik_llama-cuda.Containerfile --target full --tag ik_llama-cuda:full . && docker image build --file ik_llama-cuda.Containerfile --target server --tag ik_llama-cuda:server .`
- Look for premade quants (and imatrix files) that work well on most standard systems and are designed around ik_llama.cpp (with helpful metrics in the model card) from [ubergarm](https://huggingface.co/ubergarm/models).
- Usefull graphs and numbers on @magikRUKKOLA [Perplexity vs Size Graphs for the recent quants (GLM-4.7, Kimi-K2-Thinking, Deepseek-V3.1-Terminus, Deepseek-R1, Qwen3-Coder, Kimi-K2, Chimera etc.)](https://github.com/ikawrakow/ik_llama.cpp/discussions/715) topic.
- Build custom quants with [Thireus](https://github.com/Thireus/GGUF-Tool-Suite)'s tools.
- Download from [ik_llama.cpp's Thireus fork with release builds for macOS/Windows/Ubuntu CPU and Windows CUDA](https://github.com/Thireus/ik_llama.cpp) if you cannot build.
- For a KoboldCPP experience [Croco.Cpp is fork of KoboldCPP infering GGML/GGUF models on CPU/Cuda with KoboldAI's UI. It's powered partly by IK_LLama.cpp, and compatible with most of Ikawrakow's quants except Bitnet. ](https://github.com/Nexesenex/croco.cpp)
- **Customize `llama-swap` config**: Save the `./docker/ik_llama-cpu-swap.config.yaml` or `./docker/ik_llama-cuda-swap.config.yaml` locally (e.g., under `/my_local_files/`) then map it to `/app/config.yaml` inside the container appending `-v /my_local_files/ik_llama-cpu-swap.config.yaml:/app/config.yaml:ro` to your `podman run ...` or `docker run ...`.
# Credits
- **Run in background**: Replace `-it` with `-d`: `podman run -d ...` or `docker run -d ...`. To stop it: `podman stop ik_llama` or `docker stop ik_llama`.
All credits to the awesome community:
- **GGML_NATIVE**: If you build the image on a different machine, change `-DGGML_NATIVE=ON` to `-DGGML_NATIVE=OFF` in the `.Containerfile`.
- **KV quantization types**: To use more KV quantization types, build with `-DGGML_IQK_FA_ALL_QUANTS=ON`.
- **Cleanup unused CUDA images**: If you experiment with several `CUDA_VERSION`, delete unused images (they are several GB):
```bash
podman image rm docker.io/nvidia/cuda:12.4.0-runtime-ubuntu22.04 && \
podman image rm docker.io/nvidia/cuda:12.4.0-devel-ubuntu22.04
```
- **Build without `llama-swap`**: Change `--target swap` to `--target server` in docker-bake or Containerfiles.
- **Pre-made quants**: Look for premade quants from [ubergarm](https://huggingface.co/ubergarm/models).
- **GGUF tools**: Build custom quants with [Thireus](https://github.com/Thireus/GGUF-Tool-Suite)'s tools.
- **Download prebuilt binaries**: Download from [ik_llama.cpp's Thireus fork with release builds for macOS/Windows/Ubuntu CPU and Windows CUDA](https://github.com/Thireus/ik_llama.cpp).
- **KoboldCPP experience**: [Croco.Cpp is a fork of KoboldCPP inferring GGUF/GGML models on CPU/Cuda with KoboldAI's UI. It's powered partly by IK_LLama.cpp, and compatible with most of Ikawrakow's quants except Bitnet.](https://github.com/Nexesenex/croco.cpp)
## Credits
All credits to the awesome community:
[llama-swap](https://github.com/mostlygeek/llama-swap)

129
docker/ik_llama-cpu.Containerfile
View File

@ -1,73 +1,94 @@
ARG UBUNTU_VERSION=22.04
ARG UBUNTU_VERSION=24.04
# Stage 1: Build
FROM docker.io/ubuntu:$UBUNTU_VERSION AS build
# Build arguments
ARG GGML_NATIVE=ON
ARG GGML_AVX2=ON
ARG USE_CCACHE=true
# Environment variables for portability and GitHub Actions
ENV LLAMA_CURL=1
ENV LC_ALL=C.utf8
ARG CUSTOM_COMMIT
RUN apt-get update && apt-get install -yq build-essential git libcurl4-openssl-dev curl libgomp1 cmake
RUN git clone https://github.com/ikawrakow/ik_llama.cpp.git /app
# ccache configuration
ENV CCACHE_DIR=/ccache
ENV CCACHE_MAXSIZE=1G
ENV CCACHE_COMPRESS=1
ENV CCACHE_COMPRESSLEVEL=6
# This is CRITICAL for GitHub Actions: it ignores the absolute path of the runner
ENV CCACHE_BASEDIR=/app
RUN apt-get update && \
apt-get install -yq --no-install-recommends ca-certificates build-essential libcurl4-openssl-dev curl libgomp1 cmake ccache git && \
rm -rf /var/lib/apt/lists/*
# Copy source code (excluding hidden files/dirs via .dockerignore)
COPY . /app
WORKDIR /app
RUN if [ -n "$CUSTOM_COMMIT" ]; then git switch --detach "$CUSTOM_COMMIT"; fi
RUN cmake -B build -DGGML_NATIVE=ON -DLLAMA_CURL=ON -DGGML_IQK_FA_ALL_QUANTS=ON && \
cmake --build build --config Release -j$(nproc)
RUN mkdir -p /app/lib && \
find build -name "*.so" -exec cp {} /app/lib \;
RUN mkdir -p /app/build/src && \
find build -name "*.so" -exec cp {} /app/build/src \;
RUN mkdir -p /app/full \
&& cp build/bin/* /app/full \
&& cp *.py /app/full \
&& cp -r gguf-py /app/full \
&& cp -r requirements /app/full \
&& cp requirements.txt /app/full \
&& cp .devops/tools.sh /app/full/tools.sh
# Stage 2: Base
# Build using ccache and optional custom commit
RUN --mount=type=cache,target=/ccache \
--mount=type=bind,source=.git,target=.git \
if [ "${USE_CCACHE}" = "true" ]; then \
export PATH="/usr/lib/ccache:$PATH"; \
ccache -z; \
fi && \
cmake -B build \
-DGGML_NATIVE=${GGML_NATIVE} \
-DLLAMA_CURL=ON && \
cmake --build build --config Release -j$(nproc) && \
if [ "${USE_CCACHE}" = "true" ]; then \
ccache -s; \
fi
# Collect build artifacts
RUN mkdir -p /app/dist/lib /app/dist/full /app/dist/bin && \
find build -name "*.so" -exec cp {} /app/dist/lib \; && \
cp build/bin/* /app/dist/bin/ && \
cp build/bin/* /app/dist/full/ && \
cp *.py /app/dist/full/ && \
cp -r gguf-py /app/dist/full/ && \
cp -r requirements /app/dist/full/ && \
cp requirements.txt /app/dist/full/ && \
cp .devops/tools.sh /app/dist/full/
# Stage 2: Base (Shared Runtime)
FROM docker.io/ubuntu:$UBUNTU_VERSION AS base
RUN apt-get update && apt-get install -yq libgomp1 curl \
&& apt-get autoremove -y \
&& apt-get clean -y \
&& rm -rf /tmp/* /var/tmp/* \
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
&& find /var/cache -type f -delete
COPY --from=build /app/lib/ /app
# Stage 3: Full
FROM base AS full
COPY --from=build /app/full /app
RUN mkdir -p /app/build/src
COPY --from=build /app/build/src /app/build/src
RUN apt-get update && \
apt-get install -yq --no-install-recommends libgomp1 curl ca-certificates && \
rm -rf /var/lib/apt/lists/*
WORKDIR /app
RUN apt-get update && apt-get install -yq \
git \
python3 \
python3-pip \
&& pip install --upgrade pip setuptools wheel \
&& pip install -r requirements.txt \
&& apt-get autoremove -y \
&& apt-get clean -y \
&& rm -rf /tmp/* /var/tmp/* \
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
&& find /var/cache -type f -delete
ENTRYPOINT ["/app/full/tools.sh"]
ENV LD_LIBRARY_PATH=/app/lib
COPY --from=build /app/dist/lib /app/lib
# Stage 3: Full (Python/Dev Tools)
FROM base AS full
COPY --from=build /app/dist/full /app
RUN apt-get update && \
apt-get install -yq --no-install-recommends git python3 python3-pip && \
pip install --break-system-packages -r requirements.txt && \
rm -rf /var/lib/apt/lists/*
ENTRYPOINT ["/app/tools.sh"]
# Stage 4: Server
FROM base AS server
ENV LLAMA_ARG_HOST=0.0.0.0
COPY --from=build /app/full/llama-server /app/llama-server
WORKDIR /app
HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ]
COPY --from=build /app/dist/bin/llama-server /app/llama-server
HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
CMD [ "curl", "-f", "http://localhost:8080/health" ]
ENTRYPOINT [ "/app/llama-server" ]
# Stage 5: Swap
FROM server AS swap
ARG LS_REPO=mostlygeek/llama-swap
ARG LS_VER=198
RUN curl -LO "https://github.com/${LS_REPO}/releases/download/v${LS_VER}/llama-swap_${LS_VER}_linux_amd64.tar.gz" \
&& tar -zxf "llama-swap_${LS_VER}_linux_amd64.tar.gz" \
&& rm "llama-swap_${LS_VER}_linux_amd64.tar.gz"
COPY ./ik_llama-cpu-swap.config.yaml /app/config.yaml
HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080"]
ENTRYPOINT [ "/app/llama-swap", "-config", "/app/config.yaml" ]
ARG LS_VER=199
RUN curl -sSL "https://github.com/${LS_REPO}/releases/download/v${LS_VER}/llama-swap_${LS_VER}_linux_amd64.tar.gz" \
| tar -xz
COPY --from=build /app/docker/ik_llama-cpu-swap.config.yaml /app/config.yaml
HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
CMD [ "curl", "-f", "http://localhost:8080"]
ENTRYPOINT [ "/app/llama-swap", "-config", "/app/config.yaml" ]

126
docker/ik_llama-cuda.Containerfile
View File

@ -5,72 +5,92 @@ ARG BASE_CUDA_RUN_CONTAINER=docker.io/nvidia/cuda:${CUDA_VERSION}-runtime-ubuntu
# Stage 1: Build
FROM ${BASE_CUDA_DEV_CONTAINER} AS build
ARG CUDA_DOCKER_ARCH=86 # CUDA architecture to build for
RUN apt-get update && apt-get install -yq build-essential git libcurl4-openssl-dev curl libgomp1 cmake
RUN git clone https://github.com/ikawrakow/ik_llama.cpp.git /app
# Build arguments
ARG CUDA_DOCKER_ARCH="86;90"
ARG GGML_NATIVE=ON
ARG USE_CCACHE=true
# Environment variables for portability and GitHub Actions
ENV CCACHE_DIR=/ccache
ENV CCACHE_UMASK=000
ENV CCACHE_MAXSIZE=5G
ENV CCACHE_COMPRESS=1
ENV CCACHE_BASEDIR=/app
RUN apt-get update && \
apt-get install -yq --no-install-recommends \
ca-certificates build-essential libcurl4-openssl-dev curl libgomp1 cmake ccache git && \
rm -rf /var/lib/apt/lists/*
# Copy non-hidden files first
COPY . /app
WORKDIR /app
RUN if [ "${CUDA_DOCKER_ARCH}" != "default" ]; then \
export CMAKE_ARGS="-DCMAKE_CUDA_ARCHITECTURES=${CUDA_DOCKER_ARCH}"; \
# Build using ccache and optional custom commit
RUN --mount=type=cache,target=/ccache \
--mount=type=bind,source=.git,target=.git \
if [ "${USE_CCACHE}" = "true" ]; then \
export PATH="/usr/lib/ccache:$PATH"; \
ccache -z; \
fi && \
cmake -B build -DGGML_NATIVE=ON -DGGML_CUDA=ON -DLLAMA_CURL=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \
cmake --build build --config Release -j$(nproc)
RUN mkdir -p /app/lib && \
find build -name "*.so" -exec cp {} /app/lib \;
RUN mkdir -p /app/build/src && \
find build -name "*.so" -exec cp {} /app/build/src \;
RUN mkdir -p /app/full \
&& cp build/bin/* /app/full \
&& cp *.py /app/full \
&& cp -r gguf-py /app/full \
&& cp -r requirements /app/full \
&& cp requirements.txt /app/full \
&& cp .devops/tools.sh /app/full/tools.sh
cmake -B build \
-DGGML_NATIVE=${GGML_NATIVE} \
-DGGML_CUDA=ON \
-DCMAKE_CUDA_ARCHITECTURES="${CUDA_DOCKER_ARCH}" \
-DLLAMA_CURL=ON \
-DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined && \
cmake --build build --config Release -j$(nproc) && \
if [ "${USE_CCACHE}" = "true" ]; then \
ccache -s; \
fi
# Stage 2: base
# Collect build artifacts
RUN mkdir -p /app/dist/lib /app/dist/full /app/dist/bin && \
find build -name "*.so" -exec cp {} /app/dist/lib \; && \
cp build/bin/* /app/dist/bin/ && \
cp build/bin/* /app/dist/full/ && \
cp *.py /app/dist/full/ && \
cp -r gguf-py /app/dist/full/ && \
cp -r requirements /app/dist/full/ && \
cp requirements.txt /app/dist/full/ && \
cp .devops/tools.sh /app/dist/full/
# Stage 2: Base (Shared Runtime)
FROM ${BASE_CUDA_RUN_CONTAINER} AS base
RUN apt-get update && apt-get install -yq libgomp1 curl \
&& update-ca-certificates \
&& apt-get autoremove -y \
&& apt-get clean -y \
&& rm -rf /tmp/* /var/tmp/* \
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
&& find /var/cache -type f -delete
COPY --from=build /app/lib/ /app
# Stage 3: full
FROM base AS full
COPY --from=build /app/full /app
RUN mkdir -p /app/build/src
COPY --from=build /app/build/src /app/build/src
RUN apt-get update && \
apt-get install -yq --no-install-recommends libgomp1 curl ca-certificates && \
rm -rf /var/lib/apt/lists/*
WORKDIR /app
RUN apt-get update && apt-get install -yq \
git \
python3 \
python3-pip \
&& pip3 install --break-system-packages -r requirements.txt \
&& apt-get autoremove -y \
&& apt-get clean -y \
&& rm -rf /tmp/* /var/tmp/* \
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
&& find /var/cache -type f -delete
ENV LD_LIBRARY_PATH=/app/lib
COPY --from=build /app/dist/lib /app/lib
# Stage 3: Full (Python/Dev Tools)
FROM base AS full
COPY --from=build /app/dist/full /app
RUN apt-get update && \
apt-get install -yq --no-install-recommends git python3 python3-pip && \
pip install --break-system-packages -r requirements.txt && \
rm -rf /var/lib/apt/lists/*
ENTRYPOINT ["/app/tools.sh"]
# Stage 4: Server
FROM base AS server
ENV LLAMA_ARG_HOST=0.0.0.0
COPY --from=build /app/full/llama-server /app/llama-server
WORKDIR /app
HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ]
COPY --from=build /app/dist/bin/llama-server /app/llama-server
HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
CMD [ "curl", "-f", "http://localhost:8080/health" ]
ENTRYPOINT [ "/app/llama-server" ]
# Stage 5: Swap
FROM server AS swap
ARG LS_REPO=mostlygeek/llama-swap
ARG LS_VER=198
RUN curl -LO "https://github.com/${LS_REPO}/releases/download/v${LS_VER}/llama-swap_${LS_VER}_linux_amd64.tar.gz" \
&& tar -zxf "llama-swap_${LS_VER}_linux_amd64.tar.gz" \
&& rm "llama-swap_${LS_VER}_linux_amd64.tar.gz"
COPY ./ik_llama-cuda-swap.config.yaml /app/config.yaml
HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080"]
ENTRYPOINT [ "/app/llama-swap", "-config", "/app/config.yaml" ]
ARG LS_VER=199
RUN curl -sSL "https://github.com/${LS_REPO}/releases/download/v${LS_VER}/llama-swap_${LS_VER}_linux_amd64.tar.gz" \
| tar -xz
COPY --from=build /app/docker/ik_llama-cuda-swap.config.yaml /app/config.yaml
HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
CMD [ "curl", "-f", "http://localhost:8080"]
ENTRYPOINT [ "/app/llama-swap", "-config", "/app/config.yaml" ]