Merge pull request #33 from NexaAI/ethan/cd

Ethan/cd

.github/workflows/build-wheels-cpu-macos.yaml

@@ -0,0 +1,68 @@
+name: Build Wheels (CPU) (MacOS)
+
+on: workflow_dispatch
+
+permissions:
+  contents: write
+
+jobs:
+
+  build_wheels_macos:
+    name: Build wheels on  macos-${{ matrix.os }}
+    runs-on: macos-${{ matrix.os }}
+    strategy:
+      matrix:
+        os: [12, 13, 14]
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: "recursive"
+
+      # Used to host cibuildwheel
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.8"
+          cache: "pip"
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          # python -m pip install -e .
+          python -m pip install build wheel
+
+      - name: Build wheels
+        uses: pypa/[email protected]
+        env:
+          # disable repair
+          CIBW_REPAIR_WHEEL_COMMAND: ""
+          CIBW_ENVIRONMENT: CMAKE_ARGS="-DCMAKE_OSX_ARCHITECTURES=arm64;x86_64 "
+          CIBW_BUILD: "cp310-* cp311-* cp312-*"
+          MACOSX_DEPLOYMENT_TARGET: ${{ matrix.os }}
+          CMAKE_BUILD_PARALLEL_LEVEL: 4
+        with:
+          package-dir: .
+          output-dir: wheelhouse
+
+      - uses: actions/upload-artifact@v4
+        with:
+          name: wheels-macos-${{ matrix.os }}
+          path: ./wheelhouse/*.whl
+
+  release:
+    name: Release
+    needs: [build_wheels_macos]
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/download-artifact@v4
+        with:
+          merge-multiple: true
+          path: dist
+
+      - uses: softprops/action-gh-release@v2
+        with:
+          files: dist/*
+          tag_name: ${{ github.ref_name }}
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/build-wheels-cpu.yaml

@@ -88,48 +88,6 @@ jobs:
           name: wheels-${{ matrix.os }}
           path: ./wheelhouse/*.whl
 
-  build_wheels_macos:
-    name: Build wheels on  macos-${{ matrix.os }}
-    runs-on: macos-${{ matrix.os }}
-    strategy:
-      matrix:
-        os: [12, 13, 14]
-
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          submodules: "recursive"
-
-      # Used to host cibuildwheel
-      - uses: actions/setup-python@v5
-        with:
-          python-version: "3.8"
-          cache: "pip"
-
-      - name: Install dependencies
-        run: |
-          python -m pip install --upgrade pip
-          # python -m pip install -e .
-          python -m pip install build wheel
-
-      - name: Build wheels
-        uses: pypa/[email protected]
-        env:
-          # disable repair
-          CIBW_REPAIR_WHEEL_COMMAND: ""
-          CIBW_ENVIRONMENT: CMAKE_ARGS="-DCMAKE_OSX_ARCHITECTURES=arm64;x86_64 "
-          CIBW_BUILD: "cp310-* cp311-* cp312-*"
-          MACOSX_DEPLOYMENT_TARGET: ${{ matrix.os }}
-          CMAKE_BUILD_PARALLEL_LEVEL: 4
-        with:
-          package-dir: .
-          output-dir: wheelhouse
-
-      - uses: actions/upload-artifact@v4
-        with:
-          name: wheels-macos-${{ matrix.os }}
-          path: ./wheelhouse/*.whl
-
   build_wheels_arm64:
     name: Build arm64 wheels
     runs-on: ubuntu-20.04
@@ -185,7 +143,7 @@ jobs:
 
   release:
     name: Release
-    needs: [build_wheels, build_wheels_macos, build_wheels_arm64, build_sdist]
+    needs: [build_wheels_linux, build_wheels_win, build_wheels_arm64, build_sdist]
     runs-on: ubuntu-latest
 
     steps:

.github/workflows/build-wheels-cuda-linux.yaml

@@ -115,7 +115,8 @@ jobs:
           $env:VERBOSE = '1'
 
           $env:CMAKE_BUILD_PARALLEL_LEVEL = $(nproc)
-          cp tomls/pyproject_cuda.toml pyproject.toml
+          $env:CMAKE_ARGS = '-DSD_CUBLAS=ON -DGGML_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES=all'
+          $env:CMAKE_ARGS = $env:CMAKE_ARGS + ' -DGGML_CUDA_FORCE_MMQ=ON -DGGML_AVX2=OFF -DGGML_FMA=OFF -DGGML_F16C=OFF'
 
           python -m build --wheel
           # write the build tag to the output

.github/workflows/build-wheels-cuda-win.yaml

@@ -115,7 +115,8 @@ jobs:
           }
           $env:VERBOSE = '1'
 
-          cp tomls/pyproject_cuda.toml pyproject.toml
+          $env:CMAKE_ARGS = '-DSD_CUBLAS=ON -DGGML_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES=all'
+          $env:CMAKE_ARGS = $env:CMAKE_ARGS + ' -DGGML_CUDA_FORCE_MMQ=ON -DGGML_AVX2=OFF -DGGML_FMA=OFF -DGGML_F16C=OFF'
           $env:CMAKE_BUILD_PARALLEL_LEVEL = $(nproc)
 
           python -m build --wheel

.github/workflows/build-wheels-metal.yaml

@@ -30,17 +30,14 @@ jobs:
           # python -m pip install -e .
           python -m pip install build wheel
 
-      - name: Copy pyproject.toml
-        run: |
-          cp tomls/pyproject_metal.toml pyproject.toml
-
       - name: Build wheels
         uses: pypa/[email protected]
         env:
           # disable repair
           CIBW_REPAIR_WHEEL_COMMAND: ""
           CIBW_ARCHS: "arm64"
           CIBW_BUILD: "cp310-* cp311-* cp312-*"
+          CIBW_ENVIRONMENT: CMAKE_ARGS="-DCMAKE_OSX_ARCHITECTURES=arm64 -DCMAKE_APPLE_SILICON_PROCESSOR=arm64 -DGGML_METAL=ON -DSD_METAL=ON"
           MACOSX_DEPLOYMENT_TARGET: ${{ matrix.os }}
           CMAKE_BUILD_PARALLEL_LEVEL: 4
         with:

.github/workflows/generate-index-from-release.yaml

@@ -37,9 +37,9 @@ jobs:
       - name: Build
         run: |
           ./scripts/releases-to-pep-503.sh index/whl/cpu '^[v]?[0-9]+\.[0-9]+\.[0-9]+$'
-          ./scripts/releases-to-pep-503.sh index/whl/cu121 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu121$'
-          ./scripts/releases-to-pep-503.sh index/whl/cu122 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu122$'
-          ./scripts/releases-to-pep-503.sh index/whl/cu123 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu123$'
+          # ./scripts/releases-to-pep-503.sh index/whl/cu121 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu121$'
+          # ./scripts/releases-to-pep-503.sh index/whl/cu122 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu122$'
+          # ./scripts/releases-to-pep-503.sh index/whl/cu123 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu123$'
           ./scripts/releases-to-pep-503.sh index/whl/cu124 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu124$'
           ./scripts/releases-to-pep-503.sh index/whl/metal '^[v]?[0-9]+\.[0-9]+\.[0-9]+-metal$'
       - name: Upload artifact

nexa/gguf/nexa_inference_image.py

@@ -144,7 +144,7 @@ def txt2img(self,
         )
         return images
 
-    def loop_txt2img(self):
+    def run_txt2img(self):
         while True:
             try:
                 prompt = nexa_prompt("Enter your prompt: ")
@@ -315,4 +315,4 @@ def run_streamlit(self, model_path: str):
         if args.img2img:
             inference.run_img2img()
         else:
-            inference.loop_txt2img()
+            inference.run_txt2img()

pyproject.toml

@@ -78,7 +78,7 @@ sdist.exclude = [".github", "build", "dist", "nexa.egg-info", "dependency/llama.
 build.verbose = true
 cmake.build-type = "Release"
 cmake.version = ">=3.16"
-cmake.args = ["-DCMAKE_CXX_FLAGS=-fopenmp"]
+# cmake.args = ["-DCMAKE_CXX_FLAGS=-fopenmp"]
 
 [tool.pytest.ini_options]
 testpaths = ["tests"]

requirements.txt

@@ -12,16 +12,13 @@ numpy>=1.20.0
 diskcache>=5.6.1
 jinja2>=2.11.3
 librosa>=0.8.0
-boto3>=1.34.148
-botocore>=1.34.148
 fastapi
 uvicorn
 pydantic
 pillow
-prompt_toolkit
 
 # For onnx
-optimum[onnxruntime]>=1.7.3  # for CPU version
+optimum[onnxruntime]  # for CPU version
 diffusers  # required for image generation
 optuna
 PyYAML
@@ -33,7 +30,7 @@ transformers
 ttstokenizer
 
 # Shared dependencies
+prompt_toolkit
 tqdm
-importlib-metadata>=1.4.0,<7.0
 tabulate
 streamlit

scripts/releases-to-pep-503.sh

@@ -1,7 +1,7 @@
 #!/bin/bash
 
 package_name="nexaai"
-repo_name="nexaai-sdk-cpp"
+repo_name="nexa-sdk"
 
 # Get output directory or default to index/whl/cpu
 output_dir=${1:-"index/whl/cpu"}