mirror/LocalAI
1
0
Fork 0
mirror of https://github.com/mudler/LocalAI.git synced 2026-05-24 16:51:44 -04:00
Code Issues Packages Projects Releases Wiki Activity

feat: add cuda13 images (#7404)

Browse Source 
* chore(ci): add cuda13 jobs

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* Add to pipelines and to capabilities. Start to work on the gallery

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* gallery

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* capabilities: try to detect by looking at /usr/local

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* neutts

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* backends.yaml

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* add cuda13 l4t requirements.txt

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* add cuda13 requirements.txt

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* Fixups

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* Fixups

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* Pin vllm

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* Not all backends are compatible

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* add vllm to requirements

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* vllm is not pre-compiled for cuda 13

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

---------

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
This commit is contained in:
Ettore Di Giacinto
2025-12-02 14:24:35 +01:00
committed by GitHub
parent 9872bdf455
commit cfd95745ed
22 changed files with 631 additions and 76 deletions
Download Patch File Download Diff File Expand all files Collapse all files

303
.github/workflows/backend.yml vendored
View File

@@ -245,7 +245,6 @@ jobs:
backend: "diffusers"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
# CUDA 12 additional backends
- build-type: 'cublas'
cuda-major-version: "12"
cuda-minor-version: "0"
@@ -306,6 +305,247 @@ jobs:
backend: "chatterbox"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "12"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-12-stablediffusion-ggml'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "stablediffusion-ggml"
dockerfile: "./backend/Dockerfile.golang"
context: "./"
- build-type: 'cublas'
cuda-major-version: "12"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-12-whisper'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "whisper"
dockerfile: "./backend/Dockerfile.golang"
context: "./"
- build-type: 'cublas'
cuda-major-version: "12"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-12-rfdetr'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "rfdetr"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "12"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-12-exllama2'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "exllama2"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "12"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-12-neutts'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "neutts"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
# cuda 13
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-13-rerankers'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "rerankers"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-13-llama-cpp'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "llama-cpp"
dockerfile: "./backend/Dockerfile.llama-cpp"
context: "./"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/arm64'
skip-drivers: 'true'
tag-latest: 'auto'
tag-suffix: '-nvidia-l4t-arm64-llama-cpp'
base-image: "nvcr.io/nvidia/l4t-jetpack:r36.4.0"
runs-on: 'ubuntu-24.04-arm'
backend: "llama-cpp"
dockerfile: "./backend/Dockerfile.llama-cpp"
context: "./"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-13-transformers'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "transformers"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-13-diffusers'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "diffusers"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'l4t'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/arm64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-l4t-diffusers'
runs-on: 'ubuntu-24.04-arm'
base-image: "nvcr.io/nvidia/l4t-jetpack:r36.4.0"
skip-drivers: 'true'
backend: "diffusers"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-13-kokoro'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "kokoro"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-13-faster-whisper'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "faster-whisper"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-13-bark'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "bark"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-13-chatterbox'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "chatterbox"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-13-stablediffusion-ggml'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "stablediffusion-ggml"
dockerfile: "./backend/Dockerfile.golang"
context: "./"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/arm64'
skip-drivers: 'true'
tag-latest: 'auto'
tag-suffix: '-nvidia-l4t-arm64-stablediffusion-ggml'
base-image: "nvcr.io/nvidia/l4t-jetpack:r36.4.0"
runs-on: 'ubuntu-24.04-arm'
backend: "stablediffusion-ggml"
dockerfile: "./backend/Dockerfile.golang"
context: "./"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-13-whisper'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "whisper"
dockerfile: "./backend/Dockerfile.golang"
context: "./"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/arm64'
skip-drivers: 'true'
tag-latest: 'auto'
tag-suffix: '-nvidia-l4t-arm64-whisper'
base-image: "nvcr.io/nvidia/l4t-jetpack:r36.4.0"
runs-on: 'ubuntu-24.04-arm'
backend: "whisper"
dockerfile: "./backend/Dockerfile.golang"
context: "./"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-13-rfdetr'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "rfdetr"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
# hipblas builds
- build-type: 'hipblas'
cuda-major-version: ""
@@ -625,18 +865,6 @@ jobs:
backend: "stablediffusion-ggml"
dockerfile: "./backend/Dockerfile.golang"
context: "./"
- build-type: 'cublas'
cuda-major-version: "12"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-12-stablediffusion-ggml'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "stablediffusion-ggml"
dockerfile: "./backend/Dockerfile.golang"
context: "./"
- build-type: 'cublas'
cuda-major-version: "11"
cuda-minor-version: "7"
@@ -710,18 +938,6 @@ jobs:
backend: "whisper"
dockerfile: "./backend/Dockerfile.golang"
context: "./"
- build-type: 'cublas'
cuda-major-version: "12"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-12-whisper'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "whisper"
dockerfile: "./backend/Dockerfile.golang"
context: "./"
- build-type: 'cublas'
cuda-major-version: "11"
cuda-minor-version: "7"
@@ -846,18 +1062,6 @@ jobs:
backend: "rfdetr"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "12"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-12-rfdetr'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "rfdetr"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "11"
cuda-minor-version: "7"
@@ -907,18 +1111,6 @@ jobs:
backend: "exllama2"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "12"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-12-exllama2'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "exllama2"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "11"
cuda-minor-version: "7"
@@ -1006,18 +1198,7 @@ jobs:
backend: "neutts"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'cublas'
cuda-major-version: "12"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-12-neutts'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
backend: "neutts"
dockerfile: "./backend/Dockerfile.python"
context: "./backend"
- build-type: 'hipblas'
cuda-major-version: ""
cuda-minor-version: ""

9
.github/workflows/image-pr.yml vendored
View File

@@ -43,6 +43,15 @@ jobs:
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
makeflags: "--jobs=3 --output-sync=target"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'false'
tag-suffix: '-gpu-nvidia-cuda-13'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
makeflags: "--jobs=3 --output-sync=target"
- build-type: 'hipblas'
platforms: 'linux/amd64'
tag-latest: 'false'

21
.github/workflows/image.yml vendored
View File

@@ -100,6 +100,17 @@ jobs:
skip-drivers: 'false'
makeflags: "--jobs=4 --output-sync=target"
aio: "-aio-gpu-nvidia-cuda-12"
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/amd64'
tag-latest: 'auto'
tag-suffix: '-gpu-nvidia-cuda-13'
runs-on: 'ubuntu-latest'
base-image: "ubuntu:22.04"
skip-drivers: 'false'
makeflags: "--jobs=4 --output-sync=target"
aio: "-aio-gpu-nvidia-cuda-13"
- build-type: 'vulkan'
platforms: 'linux/amd64'
tag-latest: 'auto'
@@ -152,3 +163,13 @@ jobs:
runs-on: 'ubuntu-24.04-arm'
makeflags: "--jobs=4 --output-sync=target"
skip-drivers: 'true'
- build-type: 'cublas'
cuda-major-version: "13"
cuda-minor-version: "0"
platforms: 'linux/arm64'
tag-latest: 'auto'
tag-suffix: '-nvidia-l4t-arm64-cuda-13'
base-image: "nvcr.io/nvidia/l4t-jetpack:r36.4.0"
runs-on: 'ubuntu-24.04-arm'
makeflags: "--jobs=4 --output-sync=target"
skip-drivers: 'true'