short_description: Best in class human sensing SDK for far edge
license: unknown
extra_gated_heading: Acknowledge license to access the repository
extra_gated_button_content: Acknowledge license and request access
extra_gated_prompt: >
By requesting access to this model, you confirm that you have read and agree
to the Lattice Semiconductor Custom License and all associated usage
restrictions.
The full license terms are available here:
https://github.com/sensAI-Solution/sensAIReferenceDesign/blob/main/LICENSE.md
extra_gated_description: >-
You have to accept the conditions to access the files and content. Usage is
restricted according to the terms described in the README.
extra_gated_fields:
Full name: text
Organization: text
Industry:
type: select
options:
- Automotive
- Industrial
- Consumer electronics
- IoT / Embedded
- Healthcare
- Retail
- Robotics
- Aerospace / Defense
- Academic / Research
- Other
Targeted platform to run AI workloads:
type: select
options:
- FPGA
- MCU
- NPU
- CPU
- GPU
- Combination
- I don't know
What describes you the most:
type: select
options:
- I'm looking for AI models I can use as-is
- I'm looking for AI models I can fine-tune with my data
- I build and train my own models from scratch
Add me to your release notification distribution list for upcoming releases: checkbox
I agree to use this model for non-commercial purposes only and to comply with the Lattice Semiconductor Custom License as described in the README: checkbox
Lattice sensAI Edge Vision Engine SDK
Overview
The Lattice sensAI Edge Vision Engine SDK (or simply EVE!) provides an Extern-C interface and ready to use Python bindings to access retrieve the following from a camera image:
| Human Sensing Attribute | Description |
|---|---|
| 3D Head Pose | x, y, z, roll, pitch and yaw of user(s) |
| Face Landmarks | 23 or 68 face landmarks |
| Fatigue | Karolinska Sleepiness score |
| Face ID | Provides identification of detected user (if user was registered) |
| Gaze | Gaze vectors of main user |
| ROI Selection | Indicates if gaze (or projection of face) of main user intersects custom ROI(s) |
| Visual Speech Detection | Indicates if main user is speaking |
| Personal protective equipment detection | Indicates if protective eyewear, hat and gloves are worn for detected user(s) |
| Person Detection | Person bounding box(es) |
| Depth | Depth of user(s) up to 5m |
| Object Detection | Bounding box(es) and classification of 90 objects |
Our models have a low computation footprint and are ideal for 0.01-0.5 TOPS devices like FPGAs, small NPUs and SOCs. Packages will be provided for the configurations below. More are coming, including support for real time operating systems. If you have a specific need, please let us know at senseaiaccessrequest@latticesemi.com.
| Operating System | Archtecture |
|---|---|
| Windows 11 | x64 |
| Poky Linux | iMX 93 |
| Ubuntu 25.04 | x64 |
| Ubuntu 25.04 | ARM |
License and Terms of Use
This model is distributed under the Lattice Semiconductor Custom License. Usage is restricted and requires explicit acknowledgment of the license terms via the Hugging Face access request process.
Summary of Key Restrictions
This summary is provided for convenience and does not replace the full license.
- ✅ Non-commercial use only
- ✅ Internal research and evaluation permitted
- ❌ Redistribution of model weights prohibited
- ❌ Use in production or commercial systems prohibited
- ❌ Derivative model redistribution prohibited
Full License Text
The full license terms are available here: https://github.com/sensAI-Solution/sensAIReferenceDesign/blob/main/LICENSE.md By accessing or using this model, you agree to be bound by the applicable license terms.
Trial License Key
Once you have submitted the request form, you will automatically receive a 90-day trial license key via email (sent to the email associated with your Hugging Face account).
Activation
- Download the EVE SDK from this repository (your access is granted automatically).
- Install the SDK following the included documentation.
- Activate using the license key from your email.