Skip to main content

TML120 Module EVK Kit - Details

The TML120 Module EVK Kit enables users to directly work with the TML120 module and rapidly evaluate machine learning use cases without requiring dedicated embedded software development. The module can be programmed using eFabric platform or PC Programming Tool via the programming serial UART available on the Type-C USB interface.

This EVK works with eFabric platform, a proprietary low-code / no-code AI development and deployment tool. The eFabric is used to generate, configure, and package optimized AI/ML models based on user datasets and application requirements. The generated model binaries are then flashed onto the TML120 EVK kit directly from eFabric.

EFabric Key Capabilities

  • Dataset ingestion, preprocessing, and data modeling
  • AI model generation optimized for TML architecture
  • Model validation and testing on real TML120 EVK hardware
  • Automated binary generation compatible with the PC Programming Tool
  • Rapid iteration and performance evaluation using live sensor and audio data

The TML120 EVK Kit, together with the eFabric, AI model development framework, is primarily intended for rapid Proof-of-Concept (PoC) development in IoT and AI/ML applications. This integrated hardware–software approach significantly reduces development time, simplifies AI model development, and enables fast validation of real-world edge AI use cases.

EFabric + TML120 EVK Value Proposition

The tight integration of EFabric with the TML120 EVK enables:

  • Faster PoC development and deployment
  • Elimination of custom embedded firmware development
  • Reduced AI development complexity
  • Seamless transition from data to deployed edge AI models This combined ecosystem positions the TML120 EVK as a complete, scalable, and production-ready edge AI development platform.

In addition to the programming UART, the EVK provides a host processor serial UART, through which the TML120 module reports ML inference and classification event messages. This enables easy integration of AI decision output messages into higher-level system applications.

This EVK kit has following features:

  1. TML120 Module

    • Integrated ML processor optimized for CNN-based inference
    • Supports up to 1 MB of 8-bit weight vectors for low-power AI execution

  2. Audio-0 Port

    • Two onboard stereo PDM microphones
    • Ready for immediate audio-based ML classification testing

  3. Audio-1 Port (Selectable via jumper configuration)

    • Option 1: Two onboard stereo PDM microphones for audio ML evaluation
    • Option 2: External I²S / TDM audio input via 3-pin connector

  4. Audio-2 Port (Selectable via jumper configuration)

    • Option 1: 3.5 mm TRRS audio jack with onboard DAC and audio amplifier
    • Supports headset output with digital volume control
    • Enables real-time monitoring and validation of DNN/DSP algorithms
    • Option 2: External I²S / TDM audio output via 3-pin connector

  5. I²C Interface (Configurable)

    • 3-axis accelerometer for motion-based ML data acquisition
    • 6-axis accelerometer + gyroscope for advanced motion analytics
    • External I²C sensor connectivity via 3-pin connector

  6. Programming UART Port

    • Used for flashing and configuration via the PC Programming Tool
    • Implemented using UART-to-USB converter with Type-C connector

  7. EFabric-Enabled AI Development Ecosystem

    • Native support for EFabric low-code / no-code AI SDK
    • Enables direct deployment and execution of EFabric-generated AI models on the TML120 module
    • Forms a tightly coupled hardware–software workflow for rapid AI model validation

  8. Host UART Port

    • Outputs ML classification and inference events to a host processor
    • Accessible via USB serial interface and 3-pin connector

  9. Power Supply

    • 5 V DC supplied through Type-C USB connector
    • Programming and Host UART share the USB data bus