Add Python 3.13 support

coremltools versions prior to 9.0 did not support Python 3.13. Now
executorch uses 9.0, we can finally enable building wheels with
Python 3.13.

Author: PetarTerziev-UL

.github/workflows/build-wheels-aarch64-linux.yml

@@ -32,7 +32,7 @@ jobs:
       test-infra-ref: main
       with-cuda: disabled
       with-rocm: disabled
-      python-versions: '["3.10", "3.11", "3.12"]'
+      python-versions: '["3.10", "3.11", "3.12", "3.13"]'
 
   build:
     needs: generate-matrix

.github/workflows/build-wheels-linux.yml

@@ -32,7 +32,7 @@ jobs:
       test-infra-ref: main
       with-cuda: disabled
       with-rocm: disabled
-      python-versions: '["3.10", "3.11", "3.12"]'
+      python-versions: '["3.10", "3.11", "3.12", "3.13"]'
 
   build:
     needs: generate-matrix

.github/workflows/build-wheels-macos.yml

@@ -32,7 +32,7 @@ jobs:
       test-infra-ref: main
       with-cuda: disabled
       with-rocm: disabled
-      python-versions: '["3.10", "3.11", "3.12"]'
+      python-versions: '["3.10", "3.11", "3.12", "3.13"]'
 
   build:
     needs: generate-matrix

.github/workflows/build-wheels-windows.yml

@@ -27,7 +27,7 @@ jobs:
       test-infra-ref: main
       with-cuda: disabled
       with-rocm: disabled
-      python-versions: '["3.10", "3.11", "3.12"]'
+      python-versions: '["3.10", "3.11", "3.12", "3.13"]'
 
   build:
     needs: generate-matrix

.github/workflows/pull.yml

@@ -22,7 +22,7 @@ jobs:
   #   strategy:
   #     fail-fast: false
   #     matrix:
-  #       python-version: [ "3.10", "3.11", "3.12" ]
+  #       python-version: [ "3.10", "3.11", "3.12", "3.13" ]
   #   with:
   #     runner: linux.2xlarge
   #     docker-image: ci-image:executorch-ubuntu-22.04-qnn-sdk

README-wheel.md

@@ -5,7 +5,7 @@ ExecuTorch is to enable wider customization and deployment capabilities of the
 PyTorch programs.
 
 The `executorch` pip package is in beta.
-* Supported python versions: 3.10, 3.11, 3.12
+* Supported python versions: 3.10, 3.11, 3.12, 3.13
 * Compatible systems: Linux x86_64, macOS aarch64
 
 The prebuilt `executorch.runtime` module included in this package provides a way

docs/source/backends/coreml/coreml-troubleshooting.md

@@ -7,11 +7,6 @@ This page describes common issues that you may encounter when using the Core ML
 
 This happens because the model is in FP16, but Core ML interprets some of the arguments as FP32, which leads to a type mismatch.  The solution is to keep the PyTorch model in FP32.  Note that the model will be still be converted to FP16 during lowering to Core ML unless specified otherwise in the compute_precision [Core ML `CompileSpec`](coreml-partitioner.md#coreml-compilespec).  Also see the [related issue in coremltools](https://github.com/apple/coremltools/issues/2480).
 
-2. coremltools/converters/mil/backend/mil/load.py", line 499, in export
-    raise RuntimeError("BlobWriter not loaded")
-
-If you're using Python 3.13, try reducing your python version to Python 3.12.  coremltools does not support Python 3.13 per [coremltools issue #2487](https://github.com/apple/coremltools/issues/2487).
-
 ### Issues during runtime
 1. [ETCoreMLModelCompiler.mm:55] [Core ML]  Failed to compile model, error = Error Domain=com.apple.mlassetio Code=1 "Failed to parse the model specification. Error: Unable to parse ML Program: at unknown location: Unknown opset 'CoreML7'." UserInfo={NSLocalizedDescription=Failed to par$
 

docs/source/getting-started.md

@@ -8,7 +8,7 @@ This section is intended to describe the necessary steps to take a PyTorch model
 ## System Requirements
 The following are required to install the ExecuTorch host libraries, needed to export models and run from Python. Requirements for target end-user devices are backend dependent. See the appropriate backend documentation for more information.
 
-- Python 3.10 - 3.12
+- Python 3.10 - 3.13
 - g++ version 7 or higher, clang++ version 5 or higher, or another C++17-compatible toolchain.
 - Linux (x86_64 or ARM64), macOS (ARM64), or Windows (x86_64).
     - Intel-based macOS systems require building PyTorch from source (see [Building From Source](using-executorch-building-from-source.md) for instructions).

docs/source/quick-start-section.md

@@ -17,7 +17,7 @@ Follow these guides in order to get started with ExecuTorch:
 
 ## Prerequisites
 
-- Python 3.10-3.12
+- Python 3.10-3.13
 - PyTorch 2.9+
 - Basic familiarity with PyTorch model development
 

docs/source/raspberry_pi_llama_tutorial.md

@@ -4,7 +4,7 @@
 
 This tutorial demonstrates how to deploy **Llama models on Raspberry Pi 4/5 devices** using ExecuTorch:
 
-- **Prerequisites**: Linux host machine, Python 3.10-3.12, conda environment, Raspberry Pi 4/5
+- **Prerequisites**: Linux host machine, Python 3.10-3.13, conda environment, Raspberry Pi 4/5
 - **Setup**: Automated cross-compilation using `setup.sh` script for ARM toolchain installation
 - **Export**: Convert Llama models to optimized `.pte` format with quantization options
 - **Deploy**: Transfer binaries to Raspberry Pi and configure runtime libraries
@@ -19,7 +19,7 @@ This tutorial demonstrates how to deploy **Llama models on Raspberry Pi 4/5 devi
 
 **Software Dependencies**:
 
-- **Python 3.10-3.12** (ExecuTorch requirement)
+- **Python 3.10-3.13** (ExecuTorch requirement)
 - **conda** or **venv** for environment management
 - **CMake 3.29.6+**
 - **Git** for repository cloning
@@ -42,7 +42,7 @@ uname -s  # Should output: Linux
 uname -m  # Should output: x86_64
 
 # Check Python version
-python3 --version  # Should be 3.10-3.12
+python3 --version  # Should be 3.10-3.13
 
 # Check required tools
 hash cmake git md5sum 2>/dev/null || echo "Missing required tools"