From 9144c32ab2b7969186bd714567ec5fdc878d8e25 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Tue, 23 Sep 2025 13:39:28 -0400 Subject: [PATCH 01/23] Add build arg for ROCm version Signed-off-by: Simon Redman --- Dockerfile | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/Dockerfile b/Dockerfile index 30127f3b1da0..3b415b912c24 100644 --- a/Dockerfile +++ b/Dockerfile @@ -20,6 +20,7 @@ FROM requirements AS requirements-drivers ARG BUILD_TYPE ARG CUDA_MAJOR_VERSION=12 ARG CUDA_MINOR_VERSION=0 +ARG ROCM_VERSION=6.1.0 # ROCm version, in the format of their apt repo (https://repo.radeon.com/rocm/apt/). Like `latest` or `7.0_alpha` or `6.3.4`. ARG SKIP_DRIVERS=false ARG TARGETARCH ARG TARGETVARIANT @@ -146,6 +147,18 @@ RUN if [ "${BUILD_TYPE}" = "clblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then \ ; fi RUN if [ "${BUILD_TYPE}" = "hipblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then \ + # Setup for specific ROCm version as described here: https://rocm.docs.amd.com/projects/install-on-linux/en/latest/install/install-methods/package-manager/package-manager-ubuntu.html + apt-get update && \ + apt-get install -y --no-install-recommends \ + gpg wget && \ + mkdir --parents --mode=0755 /etc/apt/keyrings && \ + wget -qO - https://repo.radeon.com/rocm/rocm.gpg.key | gpg --yes --dearmor --output /etc/apt/keyrings/rocm.gpg && \ + echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] https://repo.radeon.com/rocm/apt/${ROCM_VERSION} jammy main" >> /etc/apt/sources.list.d/rocm.list && \ + echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] https://repo.radeon.com/graphics/${ROCM_VERSION}/ubuntu jammy main" >> /etc/apt/sources.list.d/rocm.list && \ + echo "Package: *" >> /etc/apt/preferences.d/rocm-pin-600 && \ + echo "Pin: release o=repo.radeon.com" >> /etc/apt/preferences.d/rocm-pin-600 && \ + echo "Pin-Priority: 600" >> /etc/apt/preferences.d/rocm-pin-600 && \ + # End setup steps for specific ROCm version apt-get update && \ apt-get install -y --no-install-recommends \ hipblas-dev \ From 98495129085f6e9f09ab5bcb5f40a8811ebbb2ce Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Tue, 23 Sep 2025 21:53:14 -0400 Subject: [PATCH 02/23] Break ROCM_VERSION into ROCM_{MAJOR,MINOR}_VERSION Signed-off-by: Simon Redman --- Dockerfile | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/Dockerfile b/Dockerfile index 3b415b912c24..48cd1a59b445 100644 --- a/Dockerfile +++ b/Dockerfile @@ -20,7 +20,8 @@ FROM requirements AS requirements-drivers ARG BUILD_TYPE ARG CUDA_MAJOR_VERSION=12 ARG CUDA_MINOR_VERSION=0 -ARG ROCM_VERSION=6.1.0 # ROCm version, in the format of their apt repo (https://repo.radeon.com/rocm/apt/). Like `latest` or `7.0_alpha` or `6.3.4`. +ARG ROCM_MAJOR_VERSION=6 +ARG ROCM_MINOR_VERSION=1 # ROCm version to append to the major version, in the format of their apt repo (https://repo.radeon.com/rocm/apt/). Like `0_alpha` or `3.4`. ARG SKIP_DRIVERS=false ARG TARGETARCH ARG TARGETVARIANT @@ -148,13 +149,16 @@ RUN if [ "${BUILD_TYPE}" = "clblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then \ RUN if [ "${BUILD_TYPE}" = "hipblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then \ # Setup for specific ROCm version as described here: https://rocm.docs.amd.com/projects/install-on-linux/en/latest/install/install-methods/package-manager/package-manager-ubuntu.html + ROCM_VERSION="${ROCM_MAJOR_VERSION}.${ROCM_MINOR_VERSION}" && \ apt-get update && \ apt-get install -y --no-install-recommends \ gpg wget && \ mkdir --parents --mode=0755 /etc/apt/keyrings && \ wget -qO - https://repo.radeon.com/rocm/rocm.gpg.key | gpg --yes --dearmor --output /etc/apt/keyrings/rocm.gpg && \ echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] https://repo.radeon.com/rocm/apt/${ROCM_VERSION} jammy main" >> /etc/apt/sources.list.d/rocm.list && \ - echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] https://repo.radeon.com/graphics/${ROCM_VERSION}/ubuntu jammy main" >> /etc/apt/sources.list.d/rocm.list && \ + if [ "${ROCM_MAJOR_VERSION}" -ge 7 ]; then \ + echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] https://repo.radeon.com/graphics/${ROCM_VERSION}/ubuntu jammy main" >> /etc/apt/sources.list.d/rocm.list \ + ; fi && \ echo "Package: *" >> /etc/apt/preferences.d/rocm-pin-600 && \ echo "Pin: release o=repo.radeon.com" >> /etc/apt/preferences.d/rocm-pin-600 && \ echo "Pin-Priority: 600" >> /etc/apt/preferences.d/rocm-pin-600 && \ From 22352b4916793c167eaa87e6ad0350e4ad99ded3 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Tue, 23 Sep 2025 21:53:26 -0400 Subject: [PATCH 03/23] Use correct ROCm package names Signed-off-by: Simon Redman --- Dockerfile | 3 +++ 1 file changed, 3 insertions(+) diff --git a/Dockerfile b/Dockerfile index 48cd1a59b445..c852772cae1a 100644 --- a/Dockerfile +++ b/Dockerfile @@ -165,6 +165,9 @@ RUN if [ "${BUILD_TYPE}" = "hipblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then # End setup steps for specific ROCm version apt-get update && \ apt-get install -y --no-install-recommends \ + rocm-developer-tools \ + rocm-hip-runtime-dev \ + rocm-hip-sdk \ hipblas-dev \ rocblas-dev && \ apt-get clean && \ From edd8c55daa1fe67ce16046e9d556dc26c73b20ab Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Thu, 25 Sep 2025 18:16:18 -0400 Subject: [PATCH 04/23] Add rocm package for runtime libs Signed-off-by: Simon Redman --- Dockerfile | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/Dockerfile b/Dockerfile index c852772cae1a..273096024e00 100644 --- a/Dockerfile +++ b/Dockerfile @@ -165,11 +165,14 @@ RUN if [ "${BUILD_TYPE}" = "hipblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then # End setup steps for specific ROCm version apt-get update && \ apt-get install -y --no-install-recommends \ + # Build dependencies rocm-developer-tools \ rocm-hip-runtime-dev \ rocm-hip-sdk \ hipblas-dev \ - rocblas-dev && \ + rocblas-dev \ + # Metapackage for the ROCm runtime + client tools + rocm && \ apt-get clean && \ rm -rf /var/lib/apt/lists/* && \ echo "amd" > /run/localai/capability && \ From e709a73ad58866a4104479fe0f290b1ce8994acb Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Thu, 25 Sep 2025 18:16:52 -0400 Subject: [PATCH 05/23] Remove hipblas-dev and rocblas-dev. I think they are not needed, and have been replaced by the libraries from AMD Signed-off-by: Simon Redman --- Dockerfile | 2 -- 1 file changed, 2 deletions(-) diff --git a/Dockerfile b/Dockerfile index 273096024e00..1905cc46a92a 100644 --- a/Dockerfile +++ b/Dockerfile @@ -169,8 +169,6 @@ RUN if [ "${BUILD_TYPE}" = "hipblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then rocm-developer-tools \ rocm-hip-runtime-dev \ rocm-hip-sdk \ - hipblas-dev \ - rocblas-dev \ # Metapackage for the ROCm runtime + client tools rocm && \ apt-get clean && \ From ab3bc75ef4169e2f7905b0af7274ec8d78130739 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Tue, 16 Dec 2025 14:12:04 -0500 Subject: [PATCH 06/23] Migrate ROCM build flags to llama-cpp docker file Signed-off-by: Simon Redman --- backend/Dockerfile.llama-cpp | 25 +++++++++++++++++++++++-- 1 file changed, 23 insertions(+), 2 deletions(-) diff --git a/backend/Dockerfile.llama-cpp b/backend/Dockerfile.llama-cpp index a93421ce37ed..8059a9c435fc 100644 --- a/backend/Dockerfile.llama-cpp +++ b/backend/Dockerfile.llama-cpp @@ -66,6 +66,8 @@ ARG CUDA_MINOR_VERSION ARG SKIP_DRIVERS=false ENV CUDA_MAJOR_VERSION=${CUDA_MAJOR_VERSION} ENV CUDA_MINOR_VERSION=${CUDA_MINOR_VERSION} +ARG ROCM_MAJOR_VERSION=7 +ARG ROCM_MINOR_VERSION=1.1 # ROCm version to append to the major version, in the format of their apt repo (https://repo.radeon.com/rocm/apt/). Like `0_alpha` or `3.4`. ENV DEBIAN_FRONTEND=noninteractive ARG TARGETARCH ARG TARGETVARIANT @@ -201,10 +203,29 @@ RUN if [ "${BUILD_TYPE}" = "clblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then \ ; fi RUN if [ "${BUILD_TYPE}" = "hipblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then \ + # Setup for specific ROCm version as described here: https://rocm.docs.amd.com/projects/install-on-linux/en/latest/install/install-methods/package-manager/package-manager-ubuntu.html + ROCM_VERSION="${ROCM_MAJOR_VERSION}.${ROCM_MINOR_VERSION}" && \ apt-get update && \ apt-get install -y --no-install-recommends \ - hipblas-dev \ - rocblas-dev && \ + gpg wget && \ + mkdir --parents --mode=0755 /etc/apt/keyrings && \ + wget -qO - https://repo.radeon.com/rocm/rocm.gpg.key | gpg --yes --dearmor --output /etc/apt/keyrings/rocm.gpg && \ + echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] https://repo.radeon.com/rocm/apt/${ROCM_VERSION} jammy main" >> /etc/apt/sources.list.d/rocm.list && \ + if [ "${ROCM_MAJOR_VERSION}" -ge 7 ]; then \ + echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] https://repo.radeon.com/graphics/${ROCM_VERSION}/ubuntu jammy main" >> /etc/apt/sources.list.d/rocm.list \ + ; fi && \ + echo "Package: *" >> /etc/apt/preferences.d/rocm-pin-600 && \ + echo "Pin: release o=repo.radeon.com" >> /etc/apt/preferences.d/rocm-pin-600 && \ + echo "Pin-Priority: 600" >> /etc/apt/preferences.d/rocm-pin-600 && \ + # End setup steps for specific ROCm version + apt-get update && \ + apt-get install -y --no-install-recommends \ + # Build dependencies + rocm-developer-tools \ + rocm-hip-runtime-dev \ + rocm-hip-sdk \ + # Metapackage for the ROCm runtime + client tools + rocm && \ apt-get clean && \ rm -rf /var/lib/apt/lists/* && \ # I have no idea why, but the ROCM lib packages don't trigger ldconfig after they install, which results in local-ai and others not being able From 216ae36daf492014799a6b467ba835180636aafd Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Wed, 17 Dec 2025 22:51:42 -0500 Subject: [PATCH 07/23] Change base-level Dockerfile back to same runtime dependency packages as previously needed Signed-off-by: Simon Redman --- Dockerfile | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) diff --git a/Dockerfile b/Dockerfile index 1905cc46a92a..c92351b54b23 100644 --- a/Dockerfile +++ b/Dockerfile @@ -162,15 +162,11 @@ RUN if [ "${BUILD_TYPE}" = "hipblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then echo "Package: *" >> /etc/apt/preferences.d/rocm-pin-600 && \ echo "Pin: release o=repo.radeon.com" >> /etc/apt/preferences.d/rocm-pin-600 && \ echo "Pin-Priority: 600" >> /etc/apt/preferences.d/rocm-pin-600 && \ - # End setup steps for specific ROCm version + # End setup steps for specific ROCm version - the packages below will be installed from the configured repositories apt-get update && \ apt-get install -y --no-install-recommends \ - # Build dependencies - rocm-developer-tools \ - rocm-hip-runtime-dev \ - rocm-hip-sdk \ - # Metapackage for the ROCm runtime + client tools - rocm && \ + rocblas-dev \ + hipblas-dev && \ apt-get clean && \ rm -rf /var/lib/apt/lists/* && \ echo "amd" > /run/localai/capability && \ From a4ef6618d4ff13311c79e088fa87f96b532688e7 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Wed, 17 Dec 2025 22:52:03 -0500 Subject: [PATCH 08/23] Use default ROCm version of 5.5.1 to match Ubuntu package Signed-off-by: Simon Redman --- Dockerfile | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/Dockerfile b/Dockerfile index c92351b54b23..7441969781f9 100644 --- a/Dockerfile +++ b/Dockerfile @@ -20,8 +20,8 @@ FROM requirements AS requirements-drivers ARG BUILD_TYPE ARG CUDA_MAJOR_VERSION=12 ARG CUDA_MINOR_VERSION=0 -ARG ROCM_MAJOR_VERSION=6 -ARG ROCM_MINOR_VERSION=1 # ROCm version to append to the major version, in the format of their apt repo (https://repo.radeon.com/rocm/apt/). Like `0_alpha` or `3.4`. +ARG ROCM_MAJOR_VERSION=5 +ARG ROCM_MINOR_VERSION=5.1 # ROCm version to append to the major version, in the format of their apt repo (https://repo.radeon.com/rocm/apt/). Like `0_alpha` or `3.4`. ARG SKIP_DRIVERS=false ARG TARGETARCH ARG TARGETVARIANT From 54c5bb80561d549d6379083b7ab8c03a52f19acd Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Wed, 17 Dec 2025 22:55:14 -0500 Subject: [PATCH 09/23] Remove ROCm runtime package from llamacpp dockerfile Signed-off-by: Simon Redman --- backend/Dockerfile.llama-cpp | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/backend/Dockerfile.llama-cpp b/backend/Dockerfile.llama-cpp index 8059a9c435fc..009672224c3b 100644 --- a/backend/Dockerfile.llama-cpp +++ b/backend/Dockerfile.llama-cpp @@ -63,11 +63,13 @@ ARG BUILD_TYPE ENV BUILD_TYPE=${BUILD_TYPE} ARG CUDA_MAJOR_VERSION ARG CUDA_MINOR_VERSION +ARG ROCM_MAJOR_VERSION +ARG ROCM_MINOR_VERSION ARG SKIP_DRIVERS=false ENV CUDA_MAJOR_VERSION=${CUDA_MAJOR_VERSION} ENV CUDA_MINOR_VERSION=${CUDA_MINOR_VERSION} -ARG ROCM_MAJOR_VERSION=7 -ARG ROCM_MINOR_VERSION=1.1 # ROCm version to append to the major version, in the format of their apt repo (https://repo.radeon.com/rocm/apt/). Like `0_alpha` or `3.4`. +ENV ROCM_MAJOR_VERSION=${ROCM_MAJOR_VERSION} +ENV ROCM_MINOR_VERSION=${ROCM_MINOR_VERSION} ENV DEBIAN_FRONTEND=noninteractive ARG TARGETARCH ARG TARGETVARIANT @@ -223,9 +225,7 @@ RUN if [ "${BUILD_TYPE}" = "hipblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then # Build dependencies rocm-developer-tools \ rocm-hip-runtime-dev \ - rocm-hip-sdk \ - # Metapackage for the ROCm runtime + client tools - rocm && \ + rocm-hip-sdk && \ apt-get clean && \ rm -rf /var/lib/apt/lists/* && \ # I have no idea why, but the ROCM lib packages don't trigger ldconfig after they install, which results in local-ai and others not being able From 9a11a071540b60974350b85510ddfb0a5b011b21 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Thu, 18 Dec 2025 10:08:34 -0500 Subject: [PATCH 10/23] Annotate ROCm runtime dependencies Signed-off-by: Simon Redman --- Dockerfile | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/Dockerfile b/Dockerfile index 7441969781f9..3c1774c206f7 100644 --- a/Dockerfile +++ b/Dockerfile @@ -165,8 +165,9 @@ RUN if [ "${BUILD_TYPE}" = "hipblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then # End setup steps for specific ROCm version - the packages below will be installed from the configured repositories apt-get update && \ apt-get install -y --no-install-recommends \ - rocblas-dev \ - hipblas-dev && \ + rocm-hip-runtime # Gives libomp.so\ + rocblas-dev # Give librocblas.so \ + hipblas-dev # Gives libhipblas.so && \ apt-get clean && \ rm -rf /var/lib/apt/lists/* && \ echo "amd" > /run/localai/capability && \ From 7d95eabe40066a1eb9fe29c92640ff610dba64e8 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Thu, 18 Dec 2025 20:14:31 -0500 Subject: [PATCH 11/23] Remove comments in runtime dependencies install step Signed-off-by: Simon Redman --- Dockerfile | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/Dockerfile b/Dockerfile index 3c1774c206f7..f1a480de45bb 100644 --- a/Dockerfile +++ b/Dockerfile @@ -165,9 +165,9 @@ RUN if [ "${BUILD_TYPE}" = "hipblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then # End setup steps for specific ROCm version - the packages below will be installed from the configured repositories apt-get update && \ apt-get install -y --no-install-recommends \ - rocm-hip-runtime # Gives libomp.so\ - rocblas-dev # Give librocblas.so \ - hipblas-dev # Gives libhipblas.so && \ + rocm-hip-runtime \ + rocblas-dev \ + hipblas-dev && \ apt-get clean && \ rm -rf /var/lib/apt/lists/* && \ echo "amd" > /run/localai/capability && \ From 517c7e6f650422c727e8e66791c2f6570386fff1 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Fri, 19 Dec 2025 20:21:43 -0500 Subject: [PATCH 12/23] Change default ROCM version to 6.4.3 Signed-off-by: Simon Redman --- Dockerfile | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/Dockerfile b/Dockerfile index f1a480de45bb..6129bcb5771f 100644 --- a/Dockerfile +++ b/Dockerfile @@ -20,8 +20,8 @@ FROM requirements AS requirements-drivers ARG BUILD_TYPE ARG CUDA_MAJOR_VERSION=12 ARG CUDA_MINOR_VERSION=0 -ARG ROCM_MAJOR_VERSION=5 -ARG ROCM_MINOR_VERSION=5.1 # ROCm version to append to the major version, in the format of their apt repo (https://repo.radeon.com/rocm/apt/). Like `0_alpha` or `3.4`. +ARG ROCM_MAJOR_VERSION=6 +ARG ROCM_MINOR_VERSION=4.3 # ROCm version to append to the major version, in the format of their apt repo (https://repo.radeon.com/rocm/apt/). Like `0_alpha` or `3.4`. ARG SKIP_DRIVERS=false ARG TARGETARCH ARG TARGETVARIANT From 54ed3e6fd69b41b8dade975ae8eb8f7dbbfbc8b3 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Fri, 19 Dec 2025 20:31:09 -0500 Subject: [PATCH 13/23] Rename ROCm6 build in image CI output to -gpu-amd-rocm-6 Signed-off-by: Simon Redman --- .github/workflows/backend.yml | 30 +++++++++++++++--------------- .github/workflows/image.yml | 15 ++++++++------- 2 files changed, 23 insertions(+), 22 deletions(-) diff --git a/.github/workflows/backend.yml b/.github/workflows/backend.yml index e47ca8be1ed5..32d3b078dd05 100644 --- a/.github/workflows/backend.yml +++ b/.github/workflows/backend.yml @@ -554,7 +554,7 @@ jobs: cuda-minor-version: "" platforms: 'linux/amd64' tag-latest: 'auto' - tag-suffix: '-gpu-rocm-hipblas-rerankers' + tag-suffix: '-gpu-amd-rocm-6-rerankers' runs-on: 'ubuntu-latest' base-image: "rocm/dev-ubuntu-24.04:6.4.4" skip-drivers: 'false' @@ -563,13 +563,13 @@ jobs: context: "./" ubuntu-version: '2404' - build-type: 'hipblas' - cuda-major-version: "" - cuda-minor-version: "" + rocm-major-version: "6" + rocm-minor-version: "4.3" platforms: 'linux/amd64' tag-latest: 'auto' - tag-suffix: '-gpu-rocm-hipblas-llama-cpp' + tag-suffix: '-gpu-amd-rocm-6-llama-cpp' runs-on: 'ubuntu-latest' - base-image: "rocm/dev-ubuntu-24.04:6.4.4" + base-image: "ubuntu:24.04" skip-drivers: 'false' backend: "llama-cpp" dockerfile: "./backend/Dockerfile.llama-cpp" @@ -580,7 +580,7 @@ jobs: cuda-minor-version: "" platforms: 'linux/amd64' tag-latest: 'auto' - tag-suffix: '-gpu-rocm-hipblas-vllm' + tag-suffix: '-gpu-amd-rocm-6-vllm' runs-on: 'arc-runner-set' base-image: "rocm/dev-ubuntu-24.04:6.4.4" skip-drivers: 'false' @@ -593,7 +593,7 @@ jobs: cuda-minor-version: "" platforms: 'linux/amd64' tag-latest: 'auto' - tag-suffix: '-gpu-rocm-hipblas-transformers' + tag-suffix: '-gpu-amd-rocm-6-transformers' runs-on: 'arc-runner-set' base-image: "rocm/dev-ubuntu-24.04:6.4.4" skip-drivers: 'false' @@ -606,7 +606,7 @@ jobs: cuda-minor-version: "" platforms: 'linux/amd64' tag-latest: 'auto' - tag-suffix: '-gpu-rocm-hipblas-diffusers' + tag-suffix: '-gpu-amd-rocm-6-diffusers' runs-on: 'arc-runner-set' base-image: "rocm/dev-ubuntu-24.04:6.4.4" skip-drivers: 'false' @@ -620,7 +620,7 @@ jobs: cuda-minor-version: "" platforms: 'linux/amd64' tag-latest: 'auto' - tag-suffix: '-gpu-rocm-hipblas-kokoro' + tag-suffix: '-gpu-amd-rocm-6-kokoro' runs-on: 'arc-runner-set' base-image: "rocm/dev-ubuntu-24.04:6.4.4" skip-drivers: 'false' @@ -633,7 +633,7 @@ jobs: cuda-minor-version: "" platforms: 'linux/amd64' tag-latest: 'auto' - tag-suffix: '-gpu-rocm-hipblas-vibevoice' + tag-suffix: '-gpu-amd-rocm-6-vibevoice' runs-on: 'arc-runner-set' base-image: "rocm/dev-ubuntu-24.04:6.4.4" skip-drivers: 'false' @@ -646,7 +646,7 @@ jobs: cuda-minor-version: "" platforms: 'linux/amd64' tag-latest: 'auto' - tag-suffix: '-gpu-rocm-hipblas-faster-whisper' + tag-suffix: '-gpu-amd-rocm-6-faster-whisper' runs-on: 'ubuntu-latest' base-image: "rocm/dev-ubuntu-24.04:6.4.4" skip-drivers: 'false' @@ -659,7 +659,7 @@ jobs: cuda-minor-version: "" platforms: 'linux/amd64' tag-latest: 'auto' - tag-suffix: '-gpu-rocm-hipblas-coqui' + tag-suffix: '-gpu-amd-rocm-6-coqui' runs-on: 'ubuntu-latest' base-image: "rocm/dev-ubuntu-24.04:6.4.4" skip-drivers: 'false' @@ -672,7 +672,7 @@ jobs: cuda-minor-version: "" platforms: 'linux/amd64' tag-latest: 'auto' - tag-suffix: '-gpu-rocm-hipblas-bark' + tag-suffix: '-gpu-amd-rocm-6-bark' runs-on: 'arc-runner-set' base-image: "rocm/dev-ubuntu-24.04:6.4.4" skip-drivers: 'false' @@ -1055,7 +1055,7 @@ jobs: cuda-minor-version: "" platforms: 'linux/amd64' tag-latest: 'auto' - tag-suffix: '-gpu-rocm-hipblas-whisper' + tag-suffix: '-gpu-amd-rocm-6-whisper' base-image: "rocm/dev-ubuntu-24.04:6.4.4" runs-on: 'ubuntu-latest' skip-drivers: 'false' @@ -1244,7 +1244,7 @@ jobs: cuda-minor-version: "" platforms: 'linux/amd64' tag-latest: 'auto' - tag-suffix: '-gpu-rocm-hipblas-neutts' + tag-suffix: '-gpu-amd-rocm-6-neutts' runs-on: 'arc-runner-set' base-image: "rocm/dev-ubuntu-24.04:6.4.4" skip-drivers: 'false' diff --git a/.github/workflows/image.yml b/.github/workflows/image.yml index ce571006e510..c0add9e68136 100644 --- a/.github/workflows/image.yml +++ b/.github/workflows/image.yml @@ -1,17 +1,17 @@ --- name: 'build container images' - + on: push: branches: - master tags: - '*' - + concurrency: group: ci-${{ github.head_ref || github.ref }}-${{ github.repository }} cancel-in-progress: true - + jobs: hipblas-jobs: uses: ./.github/workflows/image_build.yml @@ -38,9 +38,11 @@ matrix: include: - build-type: 'hipblas' + rocm-major-version: "6" + rocm-minor-version: "4.3" platforms: 'linux/amd64' tag-latest: 'auto' - tag-suffix: '-gpu-hipblas' + tag-suffix: '-gpu-amd-rocm-6' base-image: "rocm/dev-ubuntu-24.04:6.4.4" grpc-base-image: "ubuntu:24.04" runs-on: 'ubuntu-latest' @@ -48,7 +50,7 @@ aio: "-aio-gpu-hipblas" ubuntu-version: '2404' ubuntu-codename: 'noble' - + core-image-build: uses: ./.github/workflows/image_build.yml with: @@ -134,7 +136,7 @@ aio: "-aio-gpu-intel" ubuntu-version: '2404' ubuntu-codename: 'noble' - + gh-runner: uses: ./.github/workflows/image_build.yml with: @@ -184,4 +186,3 @@ skip-drivers: 'false' ubuntu-version: '2404' ubuntu-codename: 'noble' - \ No newline at end of file From 20117c4536495911881b53538d7c082615066e7c Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Fri, 19 Dec 2025 20:32:13 -0500 Subject: [PATCH 14/23] Add ROCm 7 image builds Signed-off-by: Simon Redman --- .github/workflows/backend.yml | 13 +++++++++++++ .github/workflows/image.yml | 11 +++++++++++ 2 files changed, 24 insertions(+) diff --git a/.github/workflows/backend.yml b/.github/workflows/backend.yml index 32d3b078dd05..53567d21aba2 100644 --- a/.github/workflows/backend.yml +++ b/.github/workflows/backend.yml @@ -575,6 +575,19 @@ jobs: dockerfile: "./backend/Dockerfile.llama-cpp" context: "./" ubuntu-version: '2404' + - build-type: 'hipblas' + rocm-major-version: "7" + rocm-minor-version: "1.1" + platforms: 'linux/amd64' + tag-latest: 'auto' + tag-suffix: '-gpu-amd-rocm-7-llama-cpp' + runs-on: 'ubuntu-latest' + base-image: "ubuntu:22.04" + skip-drivers: 'false' + backend: "llama-cpp" + dockerfile: "./backend/Dockerfile.llama-cpp" + context: "./" + ubuntu-version: '2204' - build-type: 'hipblas' cuda-major-version: "" cuda-minor-version: "" diff --git a/.github/workflows/image.yml b/.github/workflows/image.yml index c0add9e68136..7413207034c1 100644 --- a/.github/workflows/image.yml +++ b/.github/workflows/image.yml @@ -50,6 +50,17 @@ aio: "-aio-gpu-hipblas" ubuntu-version: '2404' ubuntu-codename: 'noble' + - build-type: 'hipblas' + rocm-major-version: "7" + rocm-minor-version: "1.1" + platforms: 'linux/amd64' + tag-latest: 'auto' + tag-suffix: '-gpu-amd-rocm-7' + base-image: "ubuntu:24.04" + runs-on: 'ubuntu-latest' + makeflags: "--jobs=3 --output-sync=target" + aio: "-aio-gpu-amd-rocm-7" + ubuntu-version: '2204' core-image-build: uses: ./.github/workflows/image_build.yml From 6e7fd301dcbfa43fbd3470e53d33f829b12bfbc8 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Sun, 21 Dec 2025 16:03:01 -0500 Subject: [PATCH 15/23] Add rocm-*-version arguments to backend.yml header Signed-off-by: Simon Redman --- .github/workflows/backend.yml | 2 ++ 1 file changed, 2 insertions(+) diff --git a/.github/workflows/backend.yml b/.github/workflows/backend.yml index 53567d21aba2..9ee7ff0fe48f 100644 --- a/.github/workflows/backend.yml +++ b/.github/workflows/backend.yml @@ -21,6 +21,8 @@ jobs: build-type: ${{ matrix.build-type }} cuda-major-version: ${{ matrix.cuda-major-version }} cuda-minor-version: ${{ matrix.cuda-minor-version }} + rocm-major-version: ${{ matrix.rocm-major-version }} + rocm-minor-version: ${{ matrix.rocm-minor-version }} platforms: ${{ matrix.platforms }} runs-on: ${{ matrix.runs-on }} base-image: ${{ matrix.base-image }} From 94a16206241905167fb33b63a5ff382ccb2c7576 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Sun, 21 Dec 2025 16:03:23 -0500 Subject: [PATCH 16/23] Change 'amd' capability to 'amd-rocm-' Signed-off-by: Simon Redman --- Dockerfile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Dockerfile b/Dockerfile index 6129bcb5771f..8d6545516634 100644 --- a/Dockerfile +++ b/Dockerfile @@ -170,7 +170,7 @@ RUN if [ "${BUILD_TYPE}" = "hipblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then hipblas-dev && \ apt-get clean && \ rm -rf /var/lib/apt/lists/* && \ - echo "amd" > /run/localai/capability && \ + echo "amd-rocm-${ROCM_MAJOR_VERSION}" > /run/localai/capability && \ # I have no idea why, but the ROCM lib packages don't trigger ldconfig after they install, which results in local-ai and others not being able # to locate the libraries. We run ldconfig ourselves to work around this packaging deficiency ldconfig \ From 88a3ced386de0f033652db86ac2e5060fa717b78 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Sun, 21 Dec 2025 16:07:08 -0500 Subject: [PATCH 17/23] Translate all backend index.yaml entries to use 'amd-rocm-6' instead of 'amd' Signed-off-by: Simon Redman --- backend/index.yaml | 188 ++++++++++++++++++++++----------------------- 1 file changed, 94 insertions(+), 94 deletions(-) diff --git a/backend/index.yaml b/backend/index.yaml index 41befc625a44..70acc3dfa710 100644 --- a/backend/index.yaml +++ b/backend/index.yaml @@ -21,7 +21,7 @@ default: "cpu-llama-cpp" nvidia: "cuda12-llama-cpp" intel: "intel-sycl-f16-llama-cpp" - amd: "rocm-llama-cpp" + amd-rocm-6: "rocm6-llama-cpp" metal: "metal-llama-cpp" vulkan: "vulkan-llama-cpp" nvidia-l4t: "nvidia-l4t-arm64-llama-cpp" @@ -49,7 +49,7 @@ nvidia: "cuda12-whisper" intel: "intel-sycl-f16-whisper" metal: "metal-whisper" - amd: "rocm-whisper" + amd-rocm-6: "rocm6-whisper" vulkan: "vulkan-whisper" nvidia-l4t: "nvidia-l4t-arm64-whisper" nvidia-cuda-13: "cuda13-whisper" @@ -76,7 +76,7 @@ default: "cpu-stablediffusion-ggml" nvidia: "cuda12-stablediffusion-ggml" intel: "intel-sycl-f16-stablediffusion-ggml" - # amd: "rocm-stablediffusion-ggml" + # amd-rocm-6: "rocm6-stablediffusion-ggml" vulkan: "vulkan-stablediffusion-ggml" nvidia-l4t: "nvidia-l4t-arm64-stablediffusion-ggml" metal: "metal-stablediffusion-ggml" @@ -103,7 +103,7 @@ capabilities: nvidia: "cuda12-rfdetr" intel: "intel-rfdetr" - #amd: "rocm-rfdetr" + #amd-rocm-6: "rocm6-rfdetr" nvidia-l4t: "nvidia-l4t-arm64-rfdetr" default: "cpu-rfdetr" nvidia-cuda-13: "cuda13-rfdetr" @@ -139,7 +139,7 @@ alias: "vllm" capabilities: nvidia: "cuda12-vllm" - amd: "rocm-vllm" + amd-rocm-6: "rocm6-vllm" intel: "intel-vllm" nvidia-cuda-12: "cuda12-vllm" - &mlx @@ -197,7 +197,7 @@ capabilities: nvidia: "cuda12-rerankers" intel: "intel-rerankers" - amd: "rocm-rerankers" + amd-rocm-6: "rocm6-rerankers" - &transformers name: "transformers" icon: https://camo.githubusercontent.com/26569a27b8a30a488dd345024b71dbc05da7ff1b2ba97bb6080c9f1ee0f26cc7/68747470733a2f2f68756767696e67666163652e636f2f64617461736574732f68756767696e67666163652f646f63756d656e746174696f6e2d696d616765732f7265736f6c76652f6d61696e2f7472616e73666f726d6572732f7472616e73666f726d6572735f61735f615f6d6f64656c5f646566696e6974696f6e2e706e67 @@ -214,7 +214,7 @@ capabilities: nvidia: "cuda12-transformers" intel: "intel-transformers" - amd: "rocm-transformers" + amd-rocm-6: "rocm6-transformers" nvidia-cuda-13: "cuda13-transformers" nvidia-cuda-12: "cuda12-transformers" - &diffusers @@ -233,7 +233,7 @@ capabilities: nvidia: "cuda12-diffusers" intel: "intel-diffusers" - amd: "rocm-diffusers" + amd-rocm-6: "rocm6-diffusers" nvidia-l4t: "nvidia-l4t-diffusers" metal: "metal-diffusers" default: "cpu-diffusers" @@ -272,7 +272,7 @@ capabilities: nvidia: "cuda12-faster-whisper" intel: "intel-faster-whisper" - amd: "rocm-faster-whisper" + amd-rocm-6: "rocm6-faster-whisper" nvidia-cuda-13: "cuda13-faster-whisper" nvidia-cuda-12: "cuda12-faster-whisper" - &moonshine @@ -310,7 +310,7 @@ capabilities: nvidia: "cuda12-kokoro" intel: "intel-kokoro" - amd: "rocm-kokoro" + amd-rocm-6: "rocm6-kokoro" nvidia-l4t: "nvidia-l4t-kokoro" nvidia-cuda-13: "cuda13-kokoro" nvidia-cuda-12: "cuda12-kokoro" @@ -335,7 +335,7 @@ capabilities: nvidia: "cuda12-coqui" intel: "intel-coqui" - amd: "rocm-coqui" + amd-rocm-6: "rocm6-coqui" nvidia-cuda-13: "cuda13-coqui" nvidia-cuda-12: "cuda12-coqui" icon: https://avatars.githubusercontent.com/u/1338804?s=200&v=4 @@ -420,7 +420,7 @@ capabilities: nvidia: "cuda12-vibevoice" intel: "intel-vibevoice" - amd: "rocm-vibevoice" + amd-rocm-6: "rocm6-vibevoice" nvidia-l4t: "nvidia-l4t-vibevoice" default: "cpu-vibevoice" nvidia-cuda-13: "cuda13-vibevoice" @@ -514,7 +514,7 @@ capabilities: default: "cpu-neutts" nvidia: "cuda12-neutts" - amd: "rocm-neutts" + amd-rocm-6: "rocm6-neutts" nvidia-l4t: "nvidia-l4t-neutts" nvidia-cuda-12: "cuda12-neutts" nvidia-l4t-cuda-12: "nvidia-l4t-arm64-neutts" @@ -523,7 +523,7 @@ capabilities: default: "cpu-neutts-development" nvidia: "cuda12-neutts-development" - amd: "rocm-neutts-development" + amd-rocm-6: "rocm6-neutts-development" nvidia-l4t: "nvidia-l4t-neutts-development" nvidia-cuda-12: "cuda12-neutts-development" nvidia-l4t-cuda-12: "nvidia-l4t-arm64-neutts-development" @@ -533,7 +533,7 @@ default: "cpu-llama-cpp-development" nvidia: "cuda12-llama-cpp-development" intel: "intel-sycl-f16-llama-cpp-development" - amd: "rocm-llama-cpp-development" + amd-rocm-6: "rocm6-llama-cpp-development" metal: "metal-llama-cpp-development" vulkan: "vulkan-llama-cpp-development" nvidia-l4t: "nvidia-l4t-arm64-llama-cpp-development" @@ -552,10 +552,10 @@ mirrors: - localai/localai-backends:latest-gpu-nvidia-cuda-12-neutts - !!merge <<: *neutts - name: "rocm-neutts" - uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-rocm-hipblas-neutts" + name: "rocm6-neutts" + uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-6-neutts" mirrors: - - localai/localai-backends:latest-gpu-rocm-hipblas-neutts + - localai/localai-backends:latest-gpu-amd-rocm-6-neutts - !!merge <<: *neutts name: "nvidia-l4t-arm64-neutts" uri: "quay.io/go-skynet/local-ai-backends:latest-nvidia-l4t-arm64-neutts" @@ -572,10 +572,10 @@ mirrors: - localai/localai-backends:master-gpu-nvidia-cuda-12-neutts - !!merge <<: *neutts - name: "rocm-neutts-development" - uri: "quay.io/go-skynet/local-ai-backends:master-gpu-rocm-hipblas-neutts" + name: "rocm6-neutts-development" + uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-neutts" mirrors: - - localai/localai-backends:master-gpu-rocm-hipblas-neutts + - localai/localai-backends:master-gpu-amd-rocm-6-neutts - !!merge <<: *neutts name: "nvidia-l4t-arm64-neutts-development" uri: "quay.io/go-skynet/local-ai-backends:master-nvidia-l4t-arm64-neutts" @@ -658,10 +658,10 @@ mirrors: - localai/localai-backends:latest-gpu-nvidia-cuda-12-llama-cpp - !!merge <<: *llamacpp - name: "rocm-llama-cpp" - uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-rocm-hipblas-llama-cpp" + name: "rocm6-llama-cpp" + uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-6-llama-cpp" mirrors: - - localai/localai-backends:latest-gpu-rocm-hipblas-llama-cpp + - localai/localai-backends:latest-gpu-amd-rocm-6-llama-cpp - !!merge <<: *llamacpp name: "intel-sycl-f32-llama-cpp" uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-intel-sycl-f32-llama-cpp" @@ -698,10 +698,10 @@ mirrors: - localai/localai-backends:master-gpu-nvidia-cuda-12-llama-cpp - !!merge <<: *llamacpp - name: "rocm-llama-cpp-development" - uri: "quay.io/go-skynet/local-ai-backends:master-gpu-rocm-hipblas-llama-cpp" + name: "rocm6-llama-cpp-development" + uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-llama-cpp" mirrors: - - localai/localai-backends:master-gpu-rocm-hipblas-llama-cpp + - localai/localai-backends:master-gpu-amd-rocm-6-llama-cpp - !!merge <<: *llamacpp name: "intel-sycl-f32-llama-cpp-development" uri: "quay.io/go-skynet/local-ai-backends:master-gpu-intel-sycl-f32-llama-cpp" @@ -769,10 +769,10 @@ mirrors: - localai/localai-backends:latest-gpu-nvidia-cuda-12-whisper - !!merge <<: *whispercpp - name: "rocm-whisper" - uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-rocm-hipblas-whisper" + name: "rocm6-whisper" + uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-6-whisper" mirrors: - - localai/localai-backends:latest-gpu-rocm-hipblas-whisper + - localai/localai-backends:latest-gpu-amd-rocm-6-whisper - !!merge <<: *whispercpp name: "intel-sycl-f32-whisper" uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-intel-sycl-f32-whisper" @@ -809,10 +809,10 @@ mirrors: - localai/localai-backends:master-gpu-nvidia-cuda-12-whisper - !!merge <<: *whispercpp - name: "rocm-whisper-development" - uri: "quay.io/go-skynet/local-ai-backends:master-gpu-rocm-hipblas-whisper" + name: "rocm6-whisper-development" + uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-whisper" mirrors: - - localai/localai-backends:master-gpu-rocm-hipblas-whisper + - localai/localai-backends:master-gpu-amd-rocm-6-whisper - !!merge <<: *whispercpp name: "intel-sycl-f32-whisper-development" uri: "quay.io/go-skynet/local-ai-backends:master-gpu-intel-sycl-f32-whisper" @@ -927,7 +927,7 @@ name: "vllm-development" capabilities: nvidia: "cuda12-vllm-development" - amd: "rocm-vllm-development" + amd-rocm-6: "rocm6-vllm-development" intel: "intel-vllm-development" - !!merge <<: *vllm name: "cuda12-vllm" @@ -935,10 +935,10 @@ mirrors: - localai/localai-backends:latest-gpu-nvidia-cuda-12-vllm - !!merge <<: *vllm - name: "rocm-vllm" - uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-rocm-hipblas-vllm" + name: "rocm6-vllm" + uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-6-vllm" mirrors: - - localai/localai-backends:latest-gpu-rocm-hipblas-vllm + - localai/localai-backends:latest-gpu-amd-rocm-6-vllm - !!merge <<: *vllm name: "intel-vllm" uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-intel-vllm" @@ -950,10 +950,10 @@ mirrors: - localai/localai-backends:master-gpu-nvidia-cuda-12-vllm - !!merge <<: *vllm - name: "rocm-vllm-development" - uri: "quay.io/go-skynet/local-ai-backends:master-gpu-rocm-hipblas-vllm" + name: "rocm6-vllm-development" + uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-vllm" mirrors: - - localai/localai-backends:master-gpu-rocm-hipblas-vllm + - localai/localai-backends:master-gpu-amd-rocm-6-vllm - !!merge <<: *vllm name: "intel-vllm-development" uri: "quay.io/go-skynet/local-ai-backends:master-gpu-intel-vllm" @@ -965,7 +965,7 @@ capabilities: nvidia: "cuda12-rfdetr-development" intel: "intel-rfdetr-development" - #amd: "rocm-rfdetr-development" + #amd-rocm-6: "rocm6-rfdetr-development" nvidia-l4t: "nvidia-l4t-arm64-rfdetr-development" default: "cpu-rfdetr-development" nvidia-cuda-13: "cuda13-rfdetr-development" @@ -980,7 +980,7 @@ mirrors: - localai/localai-backends:latest-gpu-intel-rfdetr # - !!merge <<: *rfdetr -# name: "rocm-rfdetr" +# name: "rocm6-rfdetr" # uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-hipblas-rfdetr" # mirrors: # - localai/localai-backends:latest-gpu-hipblas-rfdetr @@ -1010,7 +1010,7 @@ mirrors: - localai/localai-backends:master-gpu-intel-rfdetr # - !!merge <<: *rfdetr -# name: "rocm-rfdetr-development" +# name: "rocm6-rfdetr-development" # uri: "quay.io/go-skynet/local-ai-backends:master-gpu-hipblas-rfdetr" # mirrors: # - localai/localai-backends:master-gpu-hipblas-rfdetr @@ -1040,7 +1040,7 @@ capabilities: nvidia: "cuda12-rerankers-development" intel: "intel-rerankers-development" - amd: "rocm-rerankers-development" + amd-rocm-6: "rocm6-rerankers-development" nvidia-cuda-13: "cuda13-rerankers-development" - !!merge <<: *rerankers name: "cuda12-rerankers" @@ -1053,8 +1053,8 @@ mirrors: - localai/localai-backends:latest-gpu-intel-rerankers - !!merge <<: *rerankers - name: "rocm-rerankers" - uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-rocm-hipblas-rerankers" + name: "rocm6-rerankers" + uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-6-rerankers" mirrors: - localai/localai-backends:latest-gpu-rocm-hipblas-rerankers - !!merge <<: *rerankers @@ -1063,10 +1063,10 @@ mirrors: - localai/localai-backends:master-gpu-nvidia-cuda-12-rerankers - !!merge <<: *rerankers - name: "rocm-rerankers-development" - uri: "quay.io/go-skynet/local-ai-backends:master-gpu-rocm-hipblas-rerankers" + name: "rocm6-rerankers-development" + uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-rerankers" mirrors: - - localai/localai-backends:master-gpu-rocm-hipblas-rerankers + - localai/localai-backends:master-gpu-amd-rocm-6-rerankers - !!merge <<: *rerankers name: "intel-rerankers-development" uri: "quay.io/go-skynet/local-ai-backends:master-gpu-intel-rerankers" @@ -1088,7 +1088,7 @@ capabilities: nvidia: "cuda12-transformers-development" intel: "intel-transformers-development" - amd: "rocm-transformers-development" + amd-rocm-6: "rocm6-transformers-development" nvidia-cuda-13: "cuda13-transformers-development" - !!merge <<: *transformers name: "cuda12-transformers" @@ -1096,10 +1096,10 @@ mirrors: - localai/localai-backends:latest-gpu-nvidia-cuda-12-transformers - !!merge <<: *transformers - name: "rocm-transformers" - uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-rocm-hipblas-transformers" + name: "rocm6-transformers" + uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-6-transformers" mirrors: - - localai/localai-backends:latest-gpu-rocm-hipblas-transformers + - localai/localai-backends:latest-gpu-amd-rocm-6-transformers - !!merge <<: *transformers name: "intel-transformers" uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-intel-transformers" @@ -1111,10 +1111,10 @@ mirrors: - localai/localai-backends:master-gpu-nvidia-cuda-12-transformers - !!merge <<: *transformers - name: "rocm-transformers-development" - uri: "quay.io/go-skynet/local-ai-backends:master-gpu-rocm-hipblas-transformers" + name: "rocm6-transformers-development" + uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-transformers" mirrors: - - localai/localai-backends:master-gpu-rocm-hipblas-transformers + - localai/localai-backends:master-gpu-amd-rocm-6-transformers - !!merge <<: *transformers name: "intel-transformers-development" uri: "quay.io/go-skynet/local-ai-backends:master-gpu-intel-transformers" @@ -1136,7 +1136,7 @@ capabilities: nvidia: "cuda12-diffusers-development" intel: "intel-diffusers-development" - amd: "rocm-diffusers-development" + amd-rocm-6: "rocm6-diffusers-development" nvidia-l4t: "nvidia-l4t-diffusers-development" metal: "metal-diffusers-development" default: "cpu-diffusers-development" @@ -1177,8 +1177,8 @@ mirrors: - localai/localai-backends:latest-gpu-nvidia-cuda-12-diffusers - !!merge <<: *diffusers - name: "rocm-diffusers" - uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-rocm-hipblas-diffusers" + name: "rocm6-diffusers" + uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-6-diffusers" mirrors: - localai/localai-backends:latest-gpu-rocm-hipblas-diffusers - !!merge <<: *diffusers @@ -1192,10 +1192,10 @@ mirrors: - localai/localai-backends:master-gpu-nvidia-cuda-12-diffusers - !!merge <<: *diffusers - name: "rocm-diffusers-development" - uri: "quay.io/go-skynet/local-ai-backends:master-gpu-rocm-hipblas-diffusers" + name: "rocm6-diffusers-development" + uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-diffusers" mirrors: - - localai/localai-backends:master-gpu-rocm-hipblas-diffusers + - localai/localai-backends:master-gpu-amd-rocm-6-diffusers - !!merge <<: *diffusers name: "intel-diffusers-development" uri: "quay.io/go-skynet/local-ai-backends:master-gpu-intel-diffusers" @@ -1243,7 +1243,7 @@ capabilities: nvidia: "cuda12-kokoro-development" intel: "intel-kokoro-development" - amd: "rocm-kokoro-development" + amd-rocm-6: "rocm6-kokoro-development" nvidia-l4t: "nvidia-l4t-kokoro-development" - !!merge <<: *kokoro name: "cuda12-kokoro-development" @@ -1251,10 +1251,10 @@ mirrors: - localai/localai-backends:master-gpu-nvidia-cuda-12-kokoro - !!merge <<: *kokoro - name: "rocm-kokoro-development" - uri: "quay.io/go-skynet/local-ai-backends:master-gpu-rocm-hipblas-kokoro" + name: "rocm6-kokoro-development" + uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-kokoro" mirrors: - - localai/localai-backends:master-gpu-rocm-hipblas-kokoro + - localai/localai-backends:master-gpu-amd-rocm-6-kokoro - !!merge <<: *kokoro name: "intel-kokoro" uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-intel-kokoro" @@ -1281,10 +1281,10 @@ mirrors: - localai/localai-backends:latest-gpu-nvidia-cuda-12-kokoro - !!merge <<: *kokoro - name: "rocm-kokoro" - uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-rocm-hipblas-kokoro" + name: "rocm6-kokoro" + uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-6-kokoro" mirrors: - - localai/localai-backends:latest-gpu-rocm-hipblas-kokoro + - localai/localai-backends:latest-gpu-amd-rocm-6-kokoro - !!merge <<: *kokoro name: "cuda13-kokoro" uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-nvidia-cuda-13-kokoro" @@ -1301,7 +1301,7 @@ capabilities: nvidia: "cuda12-faster-whisper-development" intel: "intel-faster-whisper-development" - amd: "rocm-faster-whisper-development" + amd-rocm-6: "rocm6-faster-whisper-development" nvidia-cuda-13: "cuda13-faster-whisper-development" - !!merge <<: *faster-whisper name: "cuda12-faster-whisper-development" @@ -1309,10 +1309,10 @@ mirrors: - localai/localai-backends:master-gpu-nvidia-cuda-12-faster-whisper - !!merge <<: *faster-whisper - name: "rocm-faster-whisper-development" - uri: "quay.io/go-skynet/local-ai-backends:master-gpu-rocm-hipblas-faster-whisper" + name: "rocm6-faster-whisper-development" + uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-faster-whisper" mirrors: - - localai/localai-backends:master-gpu-rocm-hipblas-faster-whisper + - localai/localai-backends:master-gpu-amd-rocm-6-faster-whisper - !!merge <<: *faster-whisper name: "intel-faster-whisper" uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-intel-faster-whisper" @@ -1378,7 +1378,7 @@ capabilities: nvidia: "cuda12-coqui-development" intel: "intel-coqui-development" - amd: "rocm-coqui-development" + amd-rocm-6: "rocm6-coqui-development" - !!merge <<: *coqui name: "cuda12-coqui" uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-nvidia-cuda-12-coqui" @@ -1390,10 +1390,10 @@ mirrors: - localai/localai-backends:master-gpu-nvidia-cuda-12-coqui - !!merge <<: *coqui - name: "rocm-coqui-development" - uri: "quay.io/go-skynet/local-ai-backends:master-gpu-rocm-hipblas-coqui" + name: "rocm6-coqui-development" + uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-coqui" mirrors: - - localai/localai-backends:master-gpu-rocm-hipblas-coqui + - localai/localai-backends:master-gpu-amd-rocm-6-coqui - !!merge <<: *coqui name: "intel-coqui" uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-intel-coqui" @@ -1405,22 +1405,22 @@ mirrors: - localai/localai-backends:master-gpu-intel-coqui - !!merge <<: *coqui - name: "rocm-coqui" - uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-rocm-hipblas-coqui" + name: "rocm6-coqui" + uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-6-coqui" mirrors: - - localai/localai-backends:latest-gpu-rocm-hipblas-coqui + - localai/localai-backends:latest-gpu-amd-rocm-6-coqui ## bark - !!merge <<: *bark name: "bark-development" capabilities: nvidia: "cuda12-bark-development" intel: "intel-bark-development" - amd: "rocm-bark-development" + amd-rocm-6: "rocm6-bark-development" - !!merge <<: *bark - name: "rocm-bark-development" - uri: "quay.io/go-skynet/local-ai-backends:master-gpu-rocm-hipblas-bark" + name: "rocm6-bark-development" + uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-bark" mirrors: - - localai/localai-backends:master-gpu-rocm-hipblas-bark + - localai/localai-backends:master-gpu-amd-rocm-6-bark - !!merge <<: *bark name: "intel-bark" uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-intel-bark" @@ -1437,10 +1437,10 @@ mirrors: - localai/localai-backends:latest-gpu-nvidia-cuda-12-bark - !!merge <<: *bark - name: "rocm-bark" - uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-rocm-hipblas-bark" + name: "rocm6-bark" + uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-6-bark" mirrors: - - localai/localai-backends:latest-gpu-rocm-hipblas-bark + - localai/localai-backends:latest-gpu-amd-rocm-6-bark - !!merge <<: *bark name: "cuda12-bark-development" uri: "quay.io/go-skynet/local-ai-backends:master-gpu-nvidia-cuda-12-bark" @@ -1528,7 +1528,7 @@ capabilities: nvidia: "cuda12-vibevoice-development" intel: "intel-vibevoice-development" - amd: "rocm-vibevoice-development" + amd-rocm-6: "rocm6-vibevoice-development" nvidia-l4t: "nvidia-l4t-vibevoice-development" default: "cpu-vibevoice-development" nvidia-cuda-13: "cuda13-vibevoice-development" @@ -1576,15 +1576,15 @@ mirrors: - localai/localai-backends:master-gpu-intel-vibevoice - !!merge <<: *vibevoice - name: "rocm-vibevoice" - uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-rocm-hipblas-vibevoice" + name: "rocm6-vibevoice" + uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-6-vibevoice" mirrors: - - localai/localai-backends:latest-gpu-rocm-hipblas-vibevoice + - localai/localai-backends:latest-gpu-amd-rocm-6-vibevoice - !!merge <<: *vibevoice - name: "rocm-vibevoice-development" - uri: "quay.io/go-skynet/local-ai-backends:master-gpu-rocm-hipblas-vibevoice" + name: "rocm6-vibevoice-development" + uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-vibevoice" mirrors: - - localai/localai-backends:master-gpu-rocm-hipblas-vibevoice + - localai/localai-backends:master-gpu-amd-rocm-6-vibevoice - !!merge <<: *vibevoice name: "nvidia-l4t-vibevoice" uri: "quay.io/go-skynet/local-ai-backends:latest-nvidia-l4t-vibevoice" From 9297568deb011ac0316a0262342dfd46c05246d2 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Sun, 21 Dec 2025 16:07:39 -0500 Subject: [PATCH 18/23] Add backend/index.yaml entries for llama-cpp on rocm7 Signed-off-by: Simon Redman --- backend/index.yaml | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/backend/index.yaml b/backend/index.yaml index 70acc3dfa710..df6f2da1e70a 100644 --- a/backend/index.yaml +++ b/backend/index.yaml @@ -22,6 +22,7 @@ nvidia: "cuda12-llama-cpp" intel: "intel-sycl-f16-llama-cpp" amd-rocm-6: "rocm6-llama-cpp" + amd-rocm-7: "rocm7-llama-cpp" metal: "metal-llama-cpp" vulkan: "vulkan-llama-cpp" nvidia-l4t: "nvidia-l4t-arm64-llama-cpp" @@ -534,6 +535,7 @@ nvidia: "cuda12-llama-cpp-development" intel: "intel-sycl-f16-llama-cpp-development" amd-rocm-6: "rocm6-llama-cpp-development" + amd-rocm-7: "rocm7-llama-cpp-development" metal: "metal-llama-cpp-development" vulkan: "vulkan-llama-cpp-development" nvidia-l4t: "nvidia-l4t-arm64-llama-cpp-development" @@ -662,6 +664,11 @@ uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-6-llama-cpp" mirrors: - localai/localai-backends:latest-gpu-amd-rocm-6-llama-cpp +- !!merge <<: *llamacpp + name: "rocm7-llama-cpp" + uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-7-llama-cpp" + mirrors: + - localai/localai-backends:latest-gpu-amd-rocm-7-llama-cpp - !!merge <<: *llamacpp name: "intel-sycl-f32-llama-cpp" uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-intel-sycl-f32-llama-cpp" @@ -702,6 +709,11 @@ uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-llama-cpp" mirrors: - localai/localai-backends:master-gpu-amd-rocm-6-llama-cpp +- !!merge <<: *llamacpp + name: "rocm7-llama-cpp-development" + uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-7-llama-cpp" + mirrors: + - localai/localai-backends:master-gpu-amd-rocm-7-llama-cpp - !!merge <<: *llamacpp name: "intel-sycl-f32-llama-cpp-development" uri: "quay.io/go-skynet/local-ai-backends:master-gpu-intel-sycl-f32-llama-cpp" From 1a1af079a927c25daf399aa0cbe627c623b47339 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Sun, 21 Dec 2025 16:20:29 -0500 Subject: [PATCH 19/23] Update docker tags for previously mis-named backends Signed-off-by: Simon Redman --- .github/workflows/backend.yml | 4 ++-- backend/index.yaml | 8 ++++---- 2 files changed, 6 insertions(+), 6 deletions(-) diff --git a/.github/workflows/backend.yml b/.github/workflows/backend.yml index 9ee7ff0fe48f..2a4bab963a46 100644 --- a/.github/workflows/backend.yml +++ b/.github/workflows/backend.yml @@ -1193,7 +1193,7 @@ jobs: platforms: 'linux/amd64' skip-drivers: 'true' tag-latest: 'auto' - tag-suffix: '-gpu-hipblas-exllama2' + tag-suffix: '-gpu-amd-rocm-6-exllama2' base-image: "rocm/dev-ubuntu-24.04:6.4.4" runs-on: 'ubuntu-latest' backend: "exllama2" @@ -1219,7 +1219,7 @@ jobs: # cuda-minor-version: "" # platforms: 'linux/amd64' # tag-latest: 'auto' - # tag-suffix: '-gpu-hipblas-rfdetr' + # tag-suffix: '-gpu-amd-rocm-6-rfdetr' # base-image: "rocm/dev-ubuntu-24.04:6.4.4" # runs-on: 'ubuntu-latest' # skip-drivers: 'false' diff --git a/backend/index.yaml b/backend/index.yaml index df6f2da1e70a..70da8dbb0a8e 100644 --- a/backend/index.yaml +++ b/backend/index.yaml @@ -993,9 +993,9 @@ - localai/localai-backends:latest-gpu-intel-rfdetr # - !!merge <<: *rfdetr # name: "rocm6-rfdetr" -# uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-hipblas-rfdetr" +# uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-amd-rocm-6-rfdetr" # mirrors: -# - localai/localai-backends:latest-gpu-hipblas-rfdetr +# - localai/localai-backends:latest-gpu-amd-rocm-6-rfdetr - !!merge <<: *rfdetr name: "nvidia-l4t-arm64-rfdetr" uri: "quay.io/go-skynet/local-ai-backends:latest-nvidia-l4t-arm64-rfdetr" @@ -1023,9 +1023,9 @@ - localai/localai-backends:master-gpu-intel-rfdetr # - !!merge <<: *rfdetr # name: "rocm6-rfdetr-development" -# uri: "quay.io/go-skynet/local-ai-backends:master-gpu-hipblas-rfdetr" +# uri: "quay.io/go-skynet/local-ai-backends:master-gpu-amd-rocm-6-rfdetr" # mirrors: -# - localai/localai-backends:master-gpu-hipblas-rfdetr +# - localai/localai-backends:master-gpu-amd-rocm-6-rfdetr - !!merge <<: *rfdetr name: "cpu-rfdetr-development" uri: "quay.io/go-skynet/local-ai-backends:master-cpu-rfdetr" From 7ea4b737737ebb86225322e04a3ccd409ee2b7d0 Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Sun, 21 Dec 2025 16:21:12 -0500 Subject: [PATCH 20/23] Bulk update documentation with new image tag names Signed-off-by: Simon Redman --- README.md | 14 +++++++------- docs/content/features/GPU-acceleration.md | 4 ++-- docs/content/getting-started/container-images.md | 10 +++++----- docs/content/installation/docker.md | 6 +++--- docs/static/install.sh | 4 ++-- 5 files changed, 19 insertions(+), 19 deletions(-) diff --git a/README.md b/README.md index 092432241837..abb90194fbd7 100644 --- a/README.md +++ b/README.md @@ -43,7 +43,7 @@ > :bulb: Get help - [❓FAQ](https://localai.io/faq/) [πŸ’­Discussions](https://github.com/go-skynet/LocalAI/discussions) [:speech_balloon: Discord](https://discord.gg/uJAeKSAGDy) [:book: Documentation website](https://localai.io/) > -> [πŸ’» Quickstart](https://localai.io/basics/getting_started/) [πŸ–ΌοΈ Models](https://models.localai.io/) [πŸš€ Roadmap](https://github.com/mudler/LocalAI/issues?q=is%3Aissue+is%3Aopen+label%3Aroadmap) [πŸ›« Examples](https://github.com/mudler/LocalAI-examples) Try on +> [πŸ’» Quickstart](https://localai.io/basics/getting_started/) [πŸ–ΌοΈ Models](https://models.localai.io/) [πŸš€ Roadmap](https://github.com/mudler/LocalAI/issues?q=is%3Aissue+is%3Aopen+label%3Aroadmap) [πŸ›« Examples](https://github.com/mudler/LocalAI-examples) Try on [![Telegram](https://img.shields.io/badge/Telegram-2CA5E0?style=for-the-badge&logo=telegram&logoColor=white)](https://t.me/localaiofficial_bot) [![tests](https://github.com/go-skynet/LocalAI/actions/workflows/test.yml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/test.yml)[![Build and Release](https://github.com/go-skynet/LocalAI/actions/workflows/release.yaml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/release.yaml)[![build container images](https://github.com/go-skynet/LocalAI/actions/workflows/image.yml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/image.yml)[![Bump dependencies](https://github.com/go-skynet/LocalAI/actions/workflows/bump_deps.yaml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/bump_deps.yaml)[![Artifact Hub](https://img.shields.io/endpoint?url=https://artifacthub.io/badge/repository/localai)](https://artifacthub.io/packages/search?repo=localai) @@ -131,10 +131,10 @@ For more installation options, see [Installer Options](https://localai.io/instal Or run with docker: > **πŸ’‘ Docker Run vs Docker Start** -> +> > - `docker run` creates and starts a new container. If a container with the same name already exists, this command will fail. > - `docker start` starts an existing container that was previously created with `docker run`. -> +> > If you've already run LocalAI before and want to start it again, use: `docker start -i local-ai` ### CPU only image: @@ -163,7 +163,7 @@ docker run -ti --name local-ai -p 8080:8080 --gpus all localai/localai:latest-nv ### AMD GPU Images (ROCm): ```bash -docker run -ti --name local-ai -p 8080:8080 --device=/dev/kfd --device=/dev/dri --group-add=video localai/localai:latest-gpu-hipblas +docker run -ti --name local-ai -p 8080:8080 --device=/dev/kfd --device=/dev/dri --group-add=video localai/localai:latest-gpu-amd-rocm-6 ``` ### Intel GPU Images (oneAPI): @@ -194,7 +194,7 @@ docker run -ti --name local-ai -p 8080:8080 --gpus all localai/localai:latest-ai docker run -ti --name local-ai -p 8080:8080 localai/localai:latest-aio-gpu-intel # AMD GPU version -docker run -ti --name local-ai -p 8080:8080 --device=/dev/kfd --device=/dev/dri --group-add=video localai/localai:latest-aio-gpu-hipblas +docker run -ti --name local-ai -p 8080:8080 --device=/dev/kfd --device=/dev/dri --group-add=video localai/localai:latest-aio-gpu-amd-rocm-6 ``` For more information about the AIO images and pre-downloaded models, see [Container Documentation](https://localai.io/basics/container/). @@ -254,7 +254,7 @@ Roadmap items: [List of issues](https://github.com/mudler/LocalAI/issues?q=is%3A - πŸ—£ [Text to Audio](https://localai.io/features/text-to-audio/) - πŸ”ˆ [Audio to Text](https://localai.io/features/audio-to-text/) (Audio transcription with `whisper.cpp`) - 🎨 [Image generation](https://localai.io/features/image-generation) -- πŸ”₯ [OpenAI-alike tools API](https://localai.io/features/openai-functions/) +- πŸ”₯ [OpenAI-alike tools API](https://localai.io/features/openai-functions/) - 🧠 [Embeddings generation for vector databases](https://localai.io/features/embeddings/) - ✍️ [Constrained grammars](https://localai.io/features/constrained_grammars/) - πŸ–ΌοΈ [Download Models directly from Huggingface ](https://localai.io/models/) @@ -362,7 +362,7 @@ Other: - Github bot which answer on issues, with code and documentation as context https://github.com/JackBekket/GitHelper - Github Actions: https://github.com/marketplace/actions/start-localai - Examples: https://github.com/mudler/LocalAI/tree/master/examples/ - + ### πŸ”— Resources diff --git a/docs/content/features/GPU-acceleration.md b/docs/content/features/GPU-acceleration.md index 2f10054d3639..d2a661ebcad0 100644 --- a/docs/content/features/GPU-acceleration.md +++ b/docs/content/features/GPU-acceleration.md @@ -182,7 +182,7 @@ The following are examples of the ROCm specific configuration elements required. ```yaml # For full functionality select a non-'core' image, version locking the image is recommended for debug purposes. - image: quay.io/go-skynet/local-ai:master-aio-gpu-hipblas + image: quay.io/go-skynet/local-ai:master-aio-gpu-amd-rocm-6 environment: - DEBUG=true # If your gpu is not already included in the current list of default targets the following build details are required. @@ -205,7 +205,7 @@ docker run \ -e GPU_TARGETS=gfx906 \ --device /dev/dri \ --device /dev/kfd \ - quay.io/go-skynet/local-ai:master-aio-gpu-hipblas + quay.io/go-skynet/local-ai:master-aio-gpu-amd-rocm-6 ``` Please ensure to add all other required environment variables, port forwardings, etc to your `compose` file or `run` command. diff --git a/docs/content/getting-started/container-images.md b/docs/content/getting-started/container-images.md index 7ea98965de4f..3d28d9ee7e78 100644 --- a/docs/content/getting-started/container-images.md +++ b/docs/content/getting-started/container-images.md @@ -84,9 +84,9 @@ Standard container images do not have pre-installed models. Use these if you wan | Description | Quay | Docker Hub | | --- | --- |-------------------------------------------------------------| -| Latest images from the branch (development) | `quay.io/go-skynet/local-ai:master-gpu-hipblas` | `localai/localai:master-gpu-hipblas` | -| Latest tag | `quay.io/go-skynet/local-ai:latest-gpu-hipblas` | `localai/localai:latest-gpu-hipblas` | -| Versioned image | `quay.io/go-skynet/local-ai:{{< version >}}-gpu-hipblas` | `localai/localai:{{< version >}}-gpu-hipblas` | +| Latest images from the branch (development) | `quay.io/go-skynet/local-ai:master-gpu-amd-rocm-6` | `localai/localai:master-gpu-amd-rocm-6` | +| Latest tag | `quay.io/go-skynet/local-ai:latest-gpu-amd-rocm-6` | `localai/localai:latest-gpu-amd-rocm-6` | +| Versioned image | `quay.io/go-skynet/local-ai:{{< version >}}-gpu-amd-rocm-6` | `localai/localai:{{< version >}}-gpu-amd-rocm-6` | {{% /tab %}} @@ -190,7 +190,7 @@ services: **Models caching**: The **AIO** image will download the needed models on the first run if not already present and store those in `/models` inside the container. The AIO models will be automatically updated with new versions of AIO images. -You can change the directory inside the container by specifying a `MODELS_PATH` environment variable (or `--models-path`). +You can change the directory inside the container by specifying a `MODELS_PATH` environment variable (or `--models-path`). If you want to use a named model or a local directory, you can mount it as a volume to `/models`: @@ -215,7 +215,7 @@ docker run -p 8080:8080 --name local-ai -ti -v localai-models:/models localai/lo | Versioned image (e.g. for CPU) | `quay.io/go-skynet/local-ai:{{< version >}}-aio-cpu` | `localai/localai:{{< version >}}-aio-cpu` | | Latest images for Nvidia GPU (CUDA12) | `quay.io/go-skynet/local-ai:latest-aio-gpu-nvidia-cuda-12` | `localai/localai:latest-aio-gpu-nvidia-cuda-12` | | Latest images for Nvidia GPU (CUDA13) | `quay.io/go-skynet/local-ai:latest-aio-gpu-nvidia-cuda-13` | `localai/localai:latest-aio-gpu-nvidia-cuda-13` | -| Latest images for AMD GPU | `quay.io/go-skynet/local-ai:latest-aio-gpu-hipblas` | `localai/localai:latest-aio-gpu-hipblas` | +| Latest images for AMD GPU | `quay.io/go-skynet/local-ai:latest-aio-gpu-amd-rocm-6` | `localai/localai:latest-aio-gpu-amd-rocm-6` | | Latest images for Intel GPU | `quay.io/go-skynet/local-ai:latest-aio-gpu-intel` | `localai/localai:latest-aio-gpu-intel` | ### Available environment variables diff --git a/docs/content/installation/docker.md b/docs/content/installation/docker.md index 7cb354f98a8a..c5b7fa574828 100644 --- a/docs/content/installation/docker.md +++ b/docs/content/installation/docker.md @@ -70,7 +70,7 @@ docker run -ti --name local-ai -p 8080:8080 --gpus all localai/localai:latest-gp **AMD GPU (ROCm):** ```bash -docker run -ti --name local-ai -p 8080:8080 --device=/dev/kfd --device=/dev/dri --group-add=video localai/localai:latest-gpu-hipblas +docker run -ti --name local-ai -p 8080:8080 --device=/dev/kfd --device=/dev/dri --group-add=video localai/localai:latest-gpu-amd-rocm-6 ``` **Intel GPU:** @@ -119,7 +119,7 @@ docker run -ti --name local-ai -p 8080:8080 --gpus all localai/localai:latest-ai **AMD GPU (ROCm):** ```bash -docker run -ti --name local-ai -p 8080:8080 --device=/dev/kfd --device=/dev/dri --group-add=video localai/localai:latest-aio-gpu-hipblas +docker run -ti --name local-ai -p 8080:8080 --device=/dev/kfd --device=/dev/dri --group-add=video localai/localai:latest-aio-gpu-amd-rocm-6 ``` **Intel GPU:** @@ -140,7 +140,7 @@ services: # image: localai/localai:latest-aio-gpu-nvidia-cuda-13 # image: localai/localai:latest-aio-gpu-nvidia-cuda-12 # image: localai/localai:latest-aio-gpu-nvidia-cuda-11 - # image: localai/localai:latest-aio-gpu-hipblas + # image: localai/localai:latest-aio-gpu-amd-rocm-6 # image: localai/localai:latest-aio-gpu-intel healthcheck: test: ["CMD", "curl", "-f", "http://localhost:8080/readyz"] diff --git a/docs/static/install.sh b/docs/static/install.sh index 8dfe7e4a1111..00fbb8253592 100755 --- a/docs/static/install.sh +++ b/docs/static/install.sh @@ -702,10 +702,10 @@ install_docker() { $envs \ -d -p $PORT:8080 --name local-ai localai/localai:$IMAGE_TAG $STARTCOMMAND elif [ "$HAS_AMD" ]; then - IMAGE_TAG=${LOCALAI_VERSION}-gpu-hipblas + IMAGE_TAG=${LOCALAI_VERSION}-gpu-amd-rocm-6 # AIO if [ "$USE_AIO" = true ]; then - IMAGE_TAG=${LOCALAI_VERSION}-aio-gpu-hipblas + IMAGE_TAG=${LOCALAI_VERSION}-aio-gpu-amd-rocm-6 fi info "Starting LocalAI Docker container..." From 2043272ddc729c4429989cfd76a833d2230c41dd Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Sat, 10 Jan 2026 22:25:08 -0500 Subject: [PATCH 21/23] Add rocm version inputs to backend_build.yml --- .github/workflows/backend_build.yml | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/.github/workflows/backend_build.yml b/.github/workflows/backend_build.yml index e458dc3cb6b2..88081ef3eee9 100644 --- a/.github/workflows/backend_build.yml +++ b/.github/workflows/backend_build.yml @@ -20,6 +20,14 @@ on: description: 'CUDA minor version' default: "1" type: string + rocm-major-version: + description: 'ROCm major version' + default: "6" + type: string + rocm-minor-version: + description: 'ROCm minor version' + default: "3.4" + type: string platforms: description: 'Platforms' default: '' From 4aad63fff212a289fb8cdca3ea8886e46ca43f0a Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Sat, 10 Jan 2026 22:26:39 -0500 Subject: [PATCH 22/23] Align the default version in backend_build with the rest of the universe --- .github/workflows/backend_build.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/workflows/backend_build.yml b/.github/workflows/backend_build.yml index 88081ef3eee9..fa1a87e0b0a5 100644 --- a/.github/workflows/backend_build.yml +++ b/.github/workflows/backend_build.yml @@ -26,7 +26,7 @@ on: type: string rocm-minor-version: description: 'ROCm minor version' - default: "3.4" + default: "4.3" type: string platforms: description: 'Platforms' From 63449eb5a749031e6839eef9d921c987c1696bff Mon Sep 17 00:00:00 2001 From: Simon Redman Date: Sat, 10 Jan 2026 22:28:33 -0500 Subject: [PATCH 23/23] Fix merge conflict'd base-image --- .github/workflows/image.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/workflows/image.yml b/.github/workflows/image.yml index 7413207034c1..7b7af2e8ff7b 100644 --- a/.github/workflows/image.yml +++ b/.github/workflows/image.yml @@ -43,7 +43,7 @@ platforms: 'linux/amd64' tag-latest: 'auto' tag-suffix: '-gpu-amd-rocm-6' - base-image: "rocm/dev-ubuntu-24.04:6.4.4" + base-image: "ubuntu:24.04" grpc-base-image: "ubuntu:24.04" runs-on: 'ubuntu-latest' makeflags: "--jobs=3 --output-sync=target"