feat(transformers): add support for CPU and MPS

Signed-off-by: Ettore Di Giacinto <[email protected]>

Author: mudler

.github/workflows/backend.yml

@@ -63,6 +63,18 @@ jobs:
             backend: "llama-cpp"
             dockerfile: "./backend/Dockerfile.llama-cpp"
             context: "./"
+          - build-type: ''
+            cuda-major-version: ""
+            cuda-minor-version: ""
+            platforms: 'linux/amd64'
+            tag-latest: 'auto'
+            tag-suffix: '-cpu-transformers'
+            runs-on: 'ubuntu-latest'
+            base-image: "ubuntu:22.04"
+            skip-drivers: 'true'
+            backend: "transformers"
+            dockerfile: "./backend/Dockerfile.python"
+            context: "./backend"
           - build-type: 'cublas'
             cuda-major-version: "11"
             cuda-minor-version: "7"
@@ -957,6 +969,20 @@ jobs:
             backend: "kitten-tts"
             dockerfile: "./backend/Dockerfile.python"
             context: "./backend"
+  transformers-darwin:
+    uses: ./.github/workflows/backend_build_darwin.yml
+    with:
+      backend: "transformers"
+      build-type: "mps"
+      go-version: "1.24.x"
+      tag-suffix: "-metal-darwin-arm64-transformers"
+      use-pip: true
+      runs-on: "macOS-14"
+    secrets:
+      dockerUsername: ${{ secrets.DOCKERHUB_USERNAME }}
+      dockerPassword: ${{ secrets.DOCKERHUB_PASSWORD }}
+      quayUsername: ${{ secrets.LOCALAI_REGISTRY_USERNAME }}
+      quayPassword: ${{ secrets.LOCALAI_REGISTRY_PASSWORD }}
   diffusers-darwin:
     uses: ./.github/workflows/backend_build_darwin.yml
     with:

backend/index.yaml

@@ -183,6 +183,8 @@
     nvidia: "cuda12-transformers"
     intel: "intel-transformers"
     amd: "rocm-transformers"
+    metal: "metal-transformers"
+    default: "cpu-transformers"
 - &diffusers
   name: "diffusers"
   icon: https://github.com/raw/huggingface/diffusers/main/docs/source/en/imgs/diffusers_library.jpg
@@ -851,6 +853,28 @@
     nvidia: "cuda12-transformers-development"
     intel: "intel-transformers-development"
     amd: "rocm-transformers-development"
+    default: "cpu-transformers-development"
+    metal: "metal-transformers-development"
+- !!merge <<: *transformers
+  name: "cpu-transformers"
+  uri: "quay.io/go-skynet/local-ai-backends:latest-cpu-transformers"
+  mirrors:
+    - localai/localai-backends:latest-cpu-transformers
+- !!merge <<: *transformers
+  name: "cpu-transformers-development"
+  uri: "quay.io/go-skynet/local-ai-backends:master-cpu-transformers"
+  mirrors:
+    - localai/localai-backends:master-cpu-transformers
+- !!merge <<: *transformers
+  name: "metal-transformers"
+  uri: "quay.io/go-skynet/local-ai-backends:latest-metal-darwin-arm64-transformers"
+  mirrors:
+    - localai/localai-backends:latest-metal-darwin-arm64-transformers
+- !!merge <<: *transformers
+  name: "metal-transformers-development"
+  uri: "quay.io/go-skynet/local-ai-backends:master-metal-darwin-arm64-transformers"
+  mirrors:
+    - localai/localai-backends:master-metal-darwin-arm64-transformers
 - !!merge <<: *transformers
   name: "cuda12-transformers"
   uri: "quay.io/go-skynet/local-ai-backends:latest-gpu-nvidia-cuda-12-transformers"

backend/python/transformers/requirements-cpu.txt

@@ -1,3 +1,4 @@
+--extra-index-url https://download.pytorch.org/whl/cpu
 torch==2.7.1
 llvmlite==0.43.0
 numba==0.60.0

backend/python/transformers/requirements-mps.txt

@@ -0,0 +1,9 @@
+torch==2.7.1
+accelerate
+llvmlite==0.43.0
+numba==0.60.0
+transformers
+bitsandbytes
+outetts
+sentence-transformers==5.1.0
+protobuf==6.32.0