` | Custom ONNX file | `-Dmodel_path=models/custom.onnx` |
| `-Dlog=true` | Enable detailed logging | `-Dlog=true` |
| `-Dcomm=true` | Add comments to generated code | `-Dcomm=true` |
| `-Dstm32n6_accel=true` | Enable STM32 N6 accelerator dispatch layer | `-Dstm32n6_accel=true` |
| `-Dstm32n6_cmsis_path=/abs/path` | Optional CMSIS include root used when the accelerator flag is set | `-Dstm32n6_cmsis_path="/opt/CMSIS_6/Source"` |
| `-Dstm32n6_use_cmsis=true` | Use CMSIS Helium helpers (requires CMSIS-DSP headers or `third_party/CMSIS-NN`) | `zig build test -Dstm32n6_accel=true -Dstm32n6_use_cmsis=true` |
| `-Dstm32n6_use_ethos=true` | Enable Ethos-U execution path (requires Ethos-U driver headers) | `zig build test -Dstm32n6_accel=true -Dstm32n6_use_ethos=true` |
| `-Dstm32n6_force_native=true` | Force the STM32 N6 accelerator shim to run on the host (useful for smoke tests) | `zig build test -Dstm32n6_accel=true -Dstm32n6_force_native=true` |
### Optional SDK downloads
```bash
# Fetch CMSIS-NN into third_party/CMSIS-NN
./scripts/fetch_cmsis_nn.sh
# (set CMSIS_NN_REPO or CMSIS_NN_REF to use a mirror/specific release)
# (set CMSIS_NN_ARCHIVE=/absolute/path/to/CMSIS-NN-main.zip to install from a local archive without network access)
# Fetch the Arm Ethos-U core driver
./scripts/fetch_ethos_u_driver.sh
# (set ETHOS_U_REPO / ETHOS_U_REF to use a fork or pinned revision)
# (set ETHOS_U_ARCHIVE=/absolute/path/to/ethos-u-driver.zip for offline installs)
# Install qemu-system-arm for the STM32 N6 QEMU regression harness
./scripts/install_qemu.sh
# (requires administrator privileges; set QEMU_SKIP_APT_UPDATE=1 to skip `apt-get update` on Debian/Ubuntu)
```
### STM32 N6 accelerator testing
```bash
# Host smoke test for the C shim (builds reference/CMSIS/Ethos shared objects)
./scripts/test_stm32n6_conv.py
# Bare-metal regression harness executed inside QEMU
# Automatically discovers CMSIS-DSP / CMSIS-NN in third_party; pass --cmsis-include/--cmsis-nn-include to override.
./scripts/test_stm32n6_qemu.py --arm-prefix arm-none-eabi --repeat 3
# Sample output (PASS markers are emitted by the firmware, the harness exits immediately after the first PASS):
# [run] reference
# stm32n6 reference PASS
# โ
reference completed in 40.57 ms
```
The script terminates QEMU as soon as the PASS banner appears, so the reported time reflects the actual firmware runtime instead of the former 3โฏs watchdog timeout. A non-zero exit status accompanied by `fatal: unexpected exception` indicates a crash inside the firmware before the PASS message is printed.
## ๐ผ Integration Examples
### CMake Integration
```cmake
target_link_libraries(your_project PUBLIC path/to/libzant.a)
```
### Arduino/Embedded C
```c
#include "lib_my_model.h"
// Optional: Set custom logging
extern void setLogFunction(void (*log_function)(uint8_t *string));
// Your inference code here
```
## ๐ฏ Real-World Examples
### Image Classification on Cortex-M33
```bash
# Generate optimized library for image classifier
zig build codegen -Dmodel=mobilenet_v2 -Dmodel_path=models/mobilenet.onnx
zig build lib -Dmodel=mobilenet_v2 -Dtarget=thumb-freestanding -Dcpu=cortex_m33 -Doutput_path=deployment/
```
### Multi-Platform Testing
```bash
# Test on different architectures
zig build test-generated-lib -Dmodel=my_model -Dtarget=native
zig build test-generated-lib -Dmodel=my_model -Dtarget=thumb-freestanding -Dcpu=cortex_m4
```
## ๐ ๏ธ Development
### For Contributors
```bash
# Run full test suite
zig build test --summary all
# Test heavy computational operations
zig build test -Dheavy=true
# Test specific operator implementations
zig build op-codegen-test -Dop=Conv
# Generate and test single operations
zig build op-codegen-gen -Dop=Add
```
### Project Structure
```
Z-Ant/
โโโ src/ # Core source code
โ โโโ Core/ # Neural network core functionality
โ โโโ CodeGen/ # Code generation engine
โ โโโ ImageToTensor/ # Image preprocessing pipeline
โ โโโ onnx/ # ONNX model parsing
โ โโโ Utils/ # Utilities and helpers
โโโ tests/ # Comprehensive test suite
โโโ datasets/ # Sample models and test data
โโโ generated/ # Generated code output
โโโ examples/ # Arduino and microcontroller examples
โโโ docs/ # Documentation and guides
```
---
## ๐ค Contributing
We welcome contributions from developers of all skill levels! Here's how to get involved:
### Getting Started
1. **Fork the repository** on GitHub
2. **Clone your fork** locally
3. **Create a feature branch** for your work
4. **Make your changes** following our coding standards
5. **Run tests** to ensure everything works
6. **Submit a pull request** for review
### Ways to Contribute
- **๐ Bug Reports**: Found an issue? Let us know!
- **โจ Feature Requests**: Have an idea? Share it with us!
- **๐ป Code Contributions**: Improve the codebase or add new features
- **๐ Documentation**: Help make the project easier to understand
- **๐งช Testing**: Write tests or improve test coverage
### Community Guidelines
- Follow our [Code of Conduct](.github/CODE_OF_CONDUCT.md)
- Check out the [Contributing Guide](docs/CONTRIBUTING.md) for detailed guidelines
- Join discussions on GitHub Issues and Discussions
### Recognition
All contributors are recognized in our [Contributors list](https://github.com/ZIGTinyBook/Z-Ant/contributors). Thank you for helping shape the future of tinyML!
---
## ๐ License
This project is licensed under the [LICENSE](LICENSE) file in the repository.
---
**Join us in revolutionizing AI on edge devices! ๐**
[GitHub](https://github.com/ZIGTinyBook/Z-Ant) โข [Documentation](docs/) โข [Examples](examples/) โข [Community](https://github.com/ZIGTinyBook/Z-Ant/discussions)
