6. Supported Devices

Tiny ML ModelZoo 1.5.0 can train and compile models for 33 target devices (common.target_device values) across nine TI device families. This section helps you find your device, see what it supports, and choose the right compilation preset.

Note

Device support in 1.5.0 is derived from the models: a device is offered for a task only if at least one model for that task lists the device in its target_devices. A device that has no model for your task is rejected by ModelMaker with a “task_type … is not supported for device …” error. The full per-device matrix is in Device Overview.

Contents

• 6.1. Device Overview
  • 6.1.1. Selecting the Device
  • 6.1.2. Complete Device List
  • 6.1.3. Compilation Presets
  • 6.1.4. NPU and Non-NPU Devices
  • 6.1.5. Feature-Extraction Memory Estimation
  • 6.1.6. Choosing a Device
• 6.2. NPU Guidelines
  • 6.2.1. NPU Devices
  • 6.2.2. How a Model Runs on the NPU
  • 6.2.3. Key Layer Rules
  • 6.2.4. Troubleshooting
• 6.3. NPU Configuration Guidelines for TI Neural Network Compiler
  • 6.3.1. Table of Contents
  • 6.3.2. Overview
  • 6.3.3. Supported Layer Types
  • 6.3.4. Terminology and Notation
  • 6.3.5. Layer Configuration Constraints
  • 6.3.6. Optimal Design Patterns
  • 6.3.7. Common Pitfalls to Avoid
  • 6.3.8. Model Design Checklist
  • 6.3.9. Quick Reference Card
  • 6.3.10. References
• 6.4. C2000 Family
  • 6.4.1. F28x Devices (C28x)
  • 6.4.2. F29x Devices (C29x)
  • 6.4.3. SDKs and Tools
  • 6.4.4. Typical Applications
  • 6.4.5. Next Steps
• 6.5. MSPM0 Family
  • 6.5.1. Supported Devices
  • 6.5.2. MSPM0G5187 (NPU)
  • 6.5.3. MSPM0G3507 and MSPM0G3519
  • 6.5.4. Running on the Board
  • 6.5.5. Next Steps
• 6.6. Connectivity, AM26x, MSPM33 and Radar Devices
  • 6.6.1. Connectivity Devices (CC)
  • 6.6.2. AM26x (AM263, AM263P, AM261)
  • 6.6.3. MSPM33C (MSPM33C32, MSPM33C34)
  • 6.6.4. IWRL6432 (Radar)
  • 6.6.5. Where to Go Next

6.7. Device Families at a Glance

Family	target_device values	CPU core	Hardware NPU
C2000 F28x	F28P55, F28P65, F28P551x, F2837, F2837xS, F2838x, F2807x, F28003, F28004, F28002x, F280013, F280015, F28E12	TI C28x DSP	F28P55
C2000 F29x	F29H85, F29P58, F29P32	TI C29x DSP	None
MSPM0	MSPM0G3507, MSPM0G3519, MSPM0G5187	Arm Cortex-M0+	MSPM0G5187
MSPM33C	MSPM33C32, MSPM33C34	Arm Cortex-M33	MSPM33C34 (no model support in 1.5.0)
AM13	AM13E2	Arm Cortex-M33	AM13E2
AM26x	AM263, AM263P, AM261	Arm Cortex-R5F	None
Connectivity	CC2755, CC2745, CC1312, CC1314, CC1352, CC1354, CC35X1	Arm Cortex-M33 / Cortex-M4	None (CC2755, CC2745, CC35X1 use the Arm Custom Datapath Extension)
Radar	IWRL6432	Arm Cortex-M4	None

6.8. What’s New for Devices in 1.5.0

• Gen3 F28x devices reconciled. F2838x, F28P551x, F28002x and the new F28E12 device profile are now listed by the regression, forecasting and anomaly-detection models. Over-broad device lists (for example arc fault and generic time series classification) were tightened to the devices each model was verified on.

• AM13E2 image classification. MobileNet-style image classifiers were compressed to fit the AM13 memory budget, and AM13E2 is now a target for image_classification.

• Radar. The new radar_classification task targets the IWRL6432 radar sensor.

• Feature-extraction memory estimation. For AM13E2, MSPM0G3507, MSPM0G5187, F29H85 and many C28x devices, ModelMaker compiles the SDK feature-extraction library and reports its Code/RO/RW memory next to the model memory. See Device Overview.

• Derived device lists. Task-level device lists for radar, image and audio classification are computed from each model’s real device support, so a task no longer advertises devices that no model supports.

6.9. Where to Go Next

Device Overview

All 33 devices with core, NPU, supported tasks, SDK example and memory estimation.

Device Overview

NPU Guidelines

Which devices have an NPU, the compilation presets, and the key layer rules.

NPU Guidelines

NPU Configuration Guidelines

The detailed layer-by-layer constraints for NPU-compatible models.

NPU Configuration Guidelines for TI Neural Network Compiler

Deploy to a Device

Copy the compiled model and golden vectors into an SDK example and run it.

Deployment