tensorstudio 1.12.0__tar.gz → 1.13.0__tar.gz

{tensorstudio-1.12.0 → tensorstudio-1.13.0}/.github/workflows/ci.yml

@@ -102,3 +102,24 @@ jobs:
           python benchmark_all.py --help
       - name: Example
         run: python examples/basic_tensor_ops.py
+
+  artifact_smoke:
+    name: Artifact smoke (${{ matrix.os }})
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [windows-latest, ubuntu-latest, macos-latest]
+
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - name: Build wheel and sdist
+        run: |
+          python -m pip install -U pip build twine
+          python -m build
+          python -m twine check dist/*
+      - name: Clean wheel and sdist install smoke
+        run: python tools/verify_artifacts.py --wheel-dir dist --sdist-dir dist

tensorstudio-1.13.0/.github/workflows/docs.yml

@@ -0,0 +1,45 @@
+name: Docs
+
+on:
+  push:
+    branches: [main]
+  workflow_dispatch:
+
+permissions:
+  contents: read
+  pages: write
+  id-token: write
+
+concurrency:
+  group: pages
+  cancel-in-progress: false
+
+jobs:
+  build:
+    name: Build docs
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - name: Install docs dependencies
+        run: |
+          python -m pip install -U pip
+          python -m pip install -e ".[docs]"
+      - name: Build strict docs
+        run: python -m mkdocs build --strict
+      - uses: actions/upload-pages-artifact@v3
+        with:
+          path: site
+
+  deploy:
+    name: Deploy GitHub Pages
+    needs: build
+    runs-on: ubuntu-latest
+    environment:
+      name: github-pages
+      url: ${{ steps.deployment.outputs.page_url }}
+    steps:
+      - id: deployment
+        uses: actions/deploy-pages@v4

{tensorstudio-1.12.0 → tensorstudio-1.13.0}/.github/workflows/publish-testpypi.yml

@@ -27,6 +27,8 @@ jobs:
         uses: pypa/cibuildwheel@v2.23.3
         env:
           CIBW_ARCHS_WINDOWS: "AMD64"
+      - name: Clean wheel install smoke
+        run: python tools/verify_artifacts.py --wheel-dir wheelhouse --skip-sdist
       - uses: actions/upload-artifact@v4
         with:
           name: dist-windows
@@ -45,6 +47,8 @@ jobs:
         uses: pypa/cibuildwheel@v2.23.3
         env:
           CIBW_ARCHS_LINUX: "x86_64"
+      - name: Clean wheel install smoke
+        run: python tools/verify_artifacts.py --wheel-dir wheelhouse --skip-sdist
       - uses: actions/upload-artifact@v4
         with:
           name: dist-linux
@@ -62,7 +66,9 @@ jobs:
       - name: Build macOS wheels
         uses: pypa/cibuildwheel@v2.23.3
         env:
-          CIBW_ARCHS_MACOS: "auto"
+          CIBW_ARCHS_MACOS: "auto universal2"
+      - name: Clean wheel install smoke
+        run: python tools/verify_artifacts.py --wheel-dir wheelhouse --skip-sdist
       - uses: actions/upload-artifact@v4
         with:
           name: dist-macos
@@ -82,6 +88,7 @@ jobs:
           python -m pip install -U build twine
           python -m build --sdist
           python -m twine check dist/*
+          python tools/verify_artifacts.py --skip-wheel --sdist-dir dist
       - uses: actions/upload-artifact@v4
         with:
           name: dist-sdist

{tensorstudio-1.12.0 → tensorstudio-1.13.0}/.github/workflows/publish.yml

@@ -29,6 +29,8 @@ jobs:
         uses: pypa/cibuildwheel@v2.23.3
         env:
           CIBW_ARCHS_WINDOWS: "AMD64"
+      - name: Clean wheel install smoke
+        run: python tools/verify_artifacts.py --wheel-dir wheelhouse --skip-sdist
       - uses: actions/upload-artifact@v4
         with:
           name: dist-windows
@@ -47,6 +49,8 @@ jobs:
         uses: pypa/cibuildwheel@v2.23.3
         env:
           CIBW_ARCHS_LINUX: "x86_64"
+      - name: Clean wheel install smoke
+        run: python tools/verify_artifacts.py --wheel-dir wheelhouse --skip-sdist
       - uses: actions/upload-artifact@v4
         with:
           name: dist-linux
@@ -64,7 +68,9 @@ jobs:
       - name: Build macOS wheels
         uses: pypa/cibuildwheel@v2.23.3
         env:
-          CIBW_ARCHS_MACOS: "auto"
+          CIBW_ARCHS_MACOS: "auto universal2"
+      - name: Clean wheel install smoke
+        run: python tools/verify_artifacts.py --wheel-dir wheelhouse --skip-sdist
       - uses: actions/upload-artifact@v4
         with:
           name: dist-macos
@@ -84,14 +90,37 @@ jobs:
           python -m pip install -U build twine
           python -m build --sdist
           python -m twine check dist/*
+          python tools/verify_artifacts.py --skip-wheel --sdist-dir dist
       - uses: actions/upload-artifact@v4
         with:
           name: dist-sdist
           path: dist/*.tar.gz
 
+  benchmark_artifacts:
+    name: Benchmark artifacts
+    runs-on: ubuntu-latest
+    needs: build_sdist
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - name: Run benchmark reports
+        run: |
+          python -m pip install -U pip
+          python -m pip install -e ".[dev]"
+          python benchmark_all.py --check-thresholds
+          python benchmarks/bench_matmul.py
+      - uses: actions/upload-artifact@v4
+        with:
+          name: dist-benchmark-reports
+          path: |
+            benchmarks/results.md
+            benchmarks/results_matmul.md
+
   publish:
     name: Publish to PyPI
-    needs: [windows_wheels, linux_wheels, macos_wheels, build_sdist]
+    needs: [windows_wheels, linux_wheels, macos_wheels, build_sdist, benchmark_artifacts]
     runs-on: ubuntu-latest
     environment: pypi
     steps:

{tensorstudio-1.12.0 → tensorstudio-1.13.0}/.github/workflows/wheels.yml

@@ -29,6 +29,8 @@ jobs:
         uses: pypa/cibuildwheel@v2.23.3
         env:
           CIBW_ARCHS_WINDOWS: "AMD64"
+      - name: Clean wheel install smoke
+        run: python tools/verify_artifacts.py --wheel-dir wheelhouse --skip-sdist
       - uses: actions/upload-artifact@v4
         with:
           name: wheels-windows
@@ -47,6 +49,8 @@ jobs:
         uses: pypa/cibuildwheel@v2.23.3
         env:
           CIBW_ARCHS_LINUX: "x86_64"
+      - name: Clean wheel install smoke
+        run: python tools/verify_artifacts.py --wheel-dir wheelhouse --skip-sdist
       - uses: actions/upload-artifact@v4
         with:
           name: wheels-linux
@@ -64,7 +68,9 @@ jobs:
       - name: Build macOS wheels
         uses: pypa/cibuildwheel@v2.23.3
         env:
-          CIBW_ARCHS_MACOS: "auto"
+          CIBW_ARCHS_MACOS: "auto universal2"
+      - name: Clean wheel install smoke
+        run: python tools/verify_artifacts.py --wheel-dir wheelhouse --skip-sdist
       - uses: actions/upload-artifact@v4
         with:
           name: wheels-macos
@@ -84,7 +90,30 @@ jobs:
           python -m pip install -U build twine
           python -m build --sdist
           python -m twine check dist/*
+          python tools/verify_artifacts.py --skip-wheel --sdist-dir dist
       - uses: actions/upload-artifact@v4
         with:
           name: sdist
           path: dist/*.tar.gz
+
+  benchmark_artifacts:
+    name: Benchmark artifacts
+    runs-on: ubuntu-latest
+    needs: build_sdist
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - name: Run benchmark reports
+        run: |
+          python -m pip install -U pip
+          python -m pip install -e ".[dev]"
+          python benchmark_all.py --check-thresholds
+          python benchmarks/bench_matmul.py
+      - uses: actions/upload-artifact@v4
+        with:
+          name: benchmark-reports
+          path: |
+            benchmarks/results.md
+            benchmarks/results_matmul.md

{tensorstudio-1.12.0 → tensorstudio-1.13.0}/CHANGELOG.md

@@ -2,6 +2,23 @@
 
 ## Unreleased
 
+## 1.13.0 - 2026-07-07
+
+- Completed the ordered Packaging, CI, And Release Quality roadmap section as
+  one release batch.
+- Added `tools/verify_artifacts.py` for clean wheel and sdist install smoke
+  tests in isolated virtual environments.
+- Expanded CI with cross-platform wheel/sdist artifact smoke jobs on Windows,
+  Linux, and macOS.
+- Hardened wheel, TestPyPI, and PyPI workflows with clean artifact verification
+  before upload.
+- Added benchmark report artifacts to release workflows.
+- Added GitHub Pages docs publishing automation with strict MkDocs builds.
+- Expanded macOS wheel coverage to include universal2 builds where supported by
+  cibuildwheel.
+- Added native ABI, wheel-tag, source-build, BLAS, and platform compatibility
+  documentation.
+
 ## 1.12.0 - 2026-07-07
 
 - Completed the ordered Serialization And Interchange roadmap section as one

{tensorstudio-1.12.0 → tensorstudio-1.13.0}/CMakeLists.txt

@@ -1,6 +1,6 @@
 cmake_minimum_required(VERSION 3.18)
 
-project(tensorstudio VERSION 1.12.0 LANGUAGES CXX)
+project(tensorstudio VERSION 1.13.0 LANGUAGES CXX)
 
 find_package(Python COMPONENTS Interpreter Development.Module REQUIRED)
 set(PYBIND11_FINDPYTHON ON)

{tensorstudio-1.12.0 → tensorstudio-1.13.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: tensorstudio
-Version: 1.12.0
+Version: 1.13.0
 Summary: TensorStudio is a compact C++ tensor and autograd engine with a Python API for learning, experimentation, and lightweight ML workloads.
 Keywords: tensor,autograd,machine-learning,cpp,pybind11
 Author: TensorStudio contributors
@@ -76,7 +76,7 @@ Description-Content-Type: text/markdown
 TensorStudio is a compact C++ tensor and autograd engine with a Python API for
 learning, experimentation, and lightweight ML workloads.
 
-TensorStudio `1.12.0` is a CPU-only stable API foundation with native C++
+TensorStudio `1.13.0` is a CPU-only stable API foundation with native C++
 threading, storage reuse, SIMD-friendly typed kernels, and optional
 CBLAS/Accelerate matrix multiplication when available. It adds native stable
 softmax/logsumexp, batched matrix multiplication, statistical reductions,
@@ -324,7 +324,7 @@ classification workflows: Pillow-backed image IO, transform pipelines,
 deterministic augmentations, `ImageFolder` datasets, metrics, image grids,
 bounding-box drawing, and compact CNN classifiers running through native
 Conv2d/pooling kernels.
-The `1.12.0` vision surface includes batch-aware transforms, color jitter, random
+The `1.13.0` vision surface includes batch-aware transforms, color jitter, random
 resized crop, rotation, affine transforms, cutout, mixup, CutMix, detection
 helpers, segmentation mask helpers, detection/segmentation folder datasets,
 ResNet-style blocks, MobileNet-style depthwise blocks, a compact UNet, and
@@ -451,11 +451,11 @@ Run the loose local regression thresholds with:
 python benchmark_all.py --check-thresholds
 ```
 
-On one Windows CPython 3.10 development run reporting `1.12.0`, with
+On one Windows CPython 3.10 development run reporting `1.13.0`, with
 TensorStudio threads enabled, storage pooling enabled, SSE2 autovectorization
-reported, and no BLAS provider found, TensorStudio beat NumPy on 6 local
-benchmark cases and lost on 97 NumPy-comparable cases. JAX CPU dispatch was
-available on that machine; TensorStudio won 39 local cases and lost 59. The
+reported, and no BLAS provider found, TensorStudio beat NumPy on 8 local
+benchmark cases and lost on 95 NumPy-comparable cases. JAX CPU dispatch was
+available on that machine; TensorStudio won 51 local cases and lost 47. The
 strongest local wins were the simple NumPy convolution/pooling reference loops
 and some small JAX-dispatch-heavy eager cases. NumPy and JAX were faster for
 many elementwise, reduction, matrix multiplication, larger activation, and
@@ -467,15 +467,15 @@ Snapshot from that local run:
 
 | operation | shape | TensorStudio | NumPy | JAX CPU dispatch | TS vs NumPy | TS vs JAX |
 |---|---:|---:|---:|---:|---:|---:|
-| `sigmoid` | `(32,)` | 0.0149 ms | 0.0044 ms | 0.0753 ms | 0.2933x | 5.0713x |
-| `mean` | `(32,)` | 0.0154 ms | 0.0080 ms | 0.0122 ms | 0.5173x | 0.7927x |
-| `sum_axis1` | `(16, 16)` | 0.0160 ms | 0.0030 ms | 0.0131 ms | 0.1891x | 0.8180x |
-| `chain_relu` | `(128,)` | 0.0849 ms | 0.0055 ms | 0.0877 ms | 0.0651x | 1.0335x |
-| `matmul` | `(256, 256)` | 2.3427 ms | 0.4332 ms | 0.2482 ms | 0.1849x | 0.1060x |
-| `conv2d_3x3_padding1` | `(1, 1, 8, 8)` | 0.1983 ms | 1.2565 ms | 0.0997 ms | 6.3378x | 0.5029x |
-| `max_pool2d_2x2` | `(1, 1, 16, 16)` | 0.0275 ms | 0.1678 ms | n/a | 6.0918x | n/a |
-| `avg_pool2d_2x2` | `(1, 1, 16, 16)` | 0.0275 ms | 0.5499 ms | n/a | 20.0031x | n/a |
-| `elementwise_backward` | `(1024,)` | 2.7782 ms | n/a | n/a | n/a | n/a |
+| `sigmoid` | `(32,)` | 0.0156 ms | 0.0048 ms | 0.0803 ms | 0.3067x | 5.1476x |
+| `mean` | `(32,)` | 0.0158 ms | 0.0091 ms | 0.0124 ms | 0.5738x | 0.7857x |
+| `sum_axis1` | `(16, 16)` | 0.0162 ms | 0.0028 ms | 0.0125 ms | 0.1725x | 0.7724x |
+| `chain_relu` | `(128,)` | 0.0873 ms | 0.0058 ms | 0.0976 ms | 0.0663x | 1.1173x |
+| `matmul` | `(256, 256)` | 2.5256 ms | 0.4482 ms | 0.2533 ms | 0.1775x | 0.1003x |
+| `conv2d_3x3_padding1` | `(1, 1, 8, 8)` | 0.1924 ms | 1.5425 ms | 0.1282 ms | 8.0171x | 0.6665x |
+| `max_pool2d_2x2` | `(1, 1, 16, 16)` | 0.0284 ms | 0.1697 ms | n/a | 5.9745x | n/a |
+| `avg_pool2d_2x2` | `(1, 1, 16, 16)` | 0.0270 ms | 0.5957 ms | n/a | 22.0918x | n/a |
+| `elementwise_backward` | `(1024,)` | 2.8153 ms | n/a | n/a | n/a | n/a |
 
 Speedup is `competitor median / TensorStudio median`, so values above `1.0x`
 favor TensorStudio.

{tensorstudio-1.12.0 → tensorstudio-1.13.0}/README.md

@@ -7,7 +7,7 @@
 TensorStudio is a compact C++ tensor and autograd engine with a Python API for
 learning, experimentation, and lightweight ML workloads.
 
-TensorStudio `1.12.0` is a CPU-only stable API foundation with native C++
+TensorStudio `1.13.0` is a CPU-only stable API foundation with native C++
 threading, storage reuse, SIMD-friendly typed kernels, and optional
 CBLAS/Accelerate matrix multiplication when available. It adds native stable
 softmax/logsumexp, batched matrix multiplication, statistical reductions,
@@ -255,7 +255,7 @@ classification workflows: Pillow-backed image IO, transform pipelines,
 deterministic augmentations, `ImageFolder` datasets, metrics, image grids,
 bounding-box drawing, and compact CNN classifiers running through native
 Conv2d/pooling kernels.
-The `1.12.0` vision surface includes batch-aware transforms, color jitter, random
+The `1.13.0` vision surface includes batch-aware transforms, color jitter, random
 resized crop, rotation, affine transforms, cutout, mixup, CutMix, detection
 helpers, segmentation mask helpers, detection/segmentation folder datasets,
 ResNet-style blocks, MobileNet-style depthwise blocks, a compact UNet, and
@@ -382,11 +382,11 @@ Run the loose local regression thresholds with:
 python benchmark_all.py --check-thresholds
 ```
 
-On one Windows CPython 3.10 development run reporting `1.12.0`, with
+On one Windows CPython 3.10 development run reporting `1.13.0`, with
 TensorStudio threads enabled, storage pooling enabled, SSE2 autovectorization
-reported, and no BLAS provider found, TensorStudio beat NumPy on 6 local
-benchmark cases and lost on 97 NumPy-comparable cases. JAX CPU dispatch was
-available on that machine; TensorStudio won 39 local cases and lost 59. The
+reported, and no BLAS provider found, TensorStudio beat NumPy on 8 local
+benchmark cases and lost on 95 NumPy-comparable cases. JAX CPU dispatch was
+available on that machine; TensorStudio won 51 local cases and lost 47. The
 strongest local wins were the simple NumPy convolution/pooling reference loops
 and some small JAX-dispatch-heavy eager cases. NumPy and JAX were faster for
 many elementwise, reduction, matrix multiplication, larger activation, and
@@ -398,15 +398,15 @@ Snapshot from that local run:
 
 | operation | shape | TensorStudio | NumPy | JAX CPU dispatch | TS vs NumPy | TS vs JAX |
 |---|---:|---:|---:|---:|---:|---:|
-| `sigmoid` | `(32,)` | 0.0149 ms | 0.0044 ms | 0.0753 ms | 0.2933x | 5.0713x |
-| `mean` | `(32,)` | 0.0154 ms | 0.0080 ms | 0.0122 ms | 0.5173x | 0.7927x |
-| `sum_axis1` | `(16, 16)` | 0.0160 ms | 0.0030 ms | 0.0131 ms | 0.1891x | 0.8180x |
-| `chain_relu` | `(128,)` | 0.0849 ms | 0.0055 ms | 0.0877 ms | 0.0651x | 1.0335x |
-| `matmul` | `(256, 256)` | 2.3427 ms | 0.4332 ms | 0.2482 ms | 0.1849x | 0.1060x |
-| `conv2d_3x3_padding1` | `(1, 1, 8, 8)` | 0.1983 ms | 1.2565 ms | 0.0997 ms | 6.3378x | 0.5029x |
-| `max_pool2d_2x2` | `(1, 1, 16, 16)` | 0.0275 ms | 0.1678 ms | n/a | 6.0918x | n/a |
-| `avg_pool2d_2x2` | `(1, 1, 16, 16)` | 0.0275 ms | 0.5499 ms | n/a | 20.0031x | n/a |
-| `elementwise_backward` | `(1024,)` | 2.7782 ms | n/a | n/a | n/a | n/a |
+| `sigmoid` | `(32,)` | 0.0156 ms | 0.0048 ms | 0.0803 ms | 0.3067x | 5.1476x |
+| `mean` | `(32,)` | 0.0158 ms | 0.0091 ms | 0.0124 ms | 0.5738x | 0.7857x |
+| `sum_axis1` | `(16, 16)` | 0.0162 ms | 0.0028 ms | 0.0125 ms | 0.1725x | 0.7724x |
+| `chain_relu` | `(128,)` | 0.0873 ms | 0.0058 ms | 0.0976 ms | 0.0663x | 1.1173x |
+| `matmul` | `(256, 256)` | 2.5256 ms | 0.4482 ms | 0.2533 ms | 0.1775x | 0.1003x |
+| `conv2d_3x3_padding1` | `(1, 1, 8, 8)` | 0.1924 ms | 1.5425 ms | 0.1282 ms | 8.0171x | 0.6665x |
+| `max_pool2d_2x2` | `(1, 1, 16, 16)` | 0.0284 ms | 0.1697 ms | n/a | 5.9745x | n/a |
+| `avg_pool2d_2x2` | `(1, 1, 16, 16)` | 0.0270 ms | 0.5957 ms | n/a | 22.0918x | n/a |
+| `elementwise_backward` | `(1024,)` | 2.8153 ms | n/a | n/a | n/a | n/a |
 
 Speedup is `competitor median / TensorStudio median`, so values above `1.0x`
 favor TensorStudio.