🍱 Visionbox
Minimum
- Windows, Mac, or Linux
- Docker 1.2+
- 6 GB of RAM
visionbox-gpu
requires a Nvidia GPU (with compute capability 5.2+) & Nvidia Driver- CPU that supports AVX instruction set
Recommended
- 8GB of RAM or better
- Intel I5 Core or better
- Nvidia GTX GPU or better
- Docker 19.03 +
💻 SDK
Minimum
- Linux, Mac, or Windows (Windows SDK only supports C++ API and has other limitations)
- 1 GB Ram (2 GB if using
FULL
orTFV5
face recognition models). - CPU that supports AVX instruction set
- 1 to N identification - 1250 Bytes RAM / template
- For GPU Support:
- The x86-64 GPU enabled SDK supports NVIDIA GPUs with GPU Compute Capability 5.2+, and currently supports CUDA 10.1 and CUDA 11.2. Both are only compatible with Linux.
- Determine your GPU compute capability here
Recommended
- Linux
- 8 GB of Ram or more
- Intel i5 CPU or better (4 cores or more) with support for AVX2
- Nvidia GPU (GTX or better)
⚙️ Embedded SDK
Minimum
- ARMv7-A
- For GPU Support:
- The AArch64 GPU enabled SDK supports NVIDIA GPUs with GPU Compute Capability 5.3+, and currently supports CUDA 10.2 (default on NVIDIA Jetson devices). Requires Ubuntu 18.04.
Recommended
- ARMv8-A
- AArch64 (64 bit execution state)
- Nvidia Jetson Xavier NX
Deprecated
🖥️ Trueface Aware
Minimum
- 16 GB of RAM
- i5 Intel CPU or Better
- 512 GB SSD
- Nvidia GPU with Comptheute Capability 6.0+ (determine your GPU compute capability here)
- Up to 4 1080-resolution cameras
Recommend
- 32 GB of RAM
- i7 Intel CPU or Better
- NVMe SSD
- 512 GB of storage
- Up to 6 1080-resolution cameras
🐍 Python SDK
Note, the deprecated Python SDK is not the same as the Python bindings for our C++ SDK. The deprecated SDK was written entirely in Python, whereas the new SDK is compiled from C++ for various targets and supports Python bindings.
Minimum
- Windows, Mac, or Linux
- Python 2.7 +
- 1 GB of RAM
- GPU version requires NVIDIA GPU
- GPU version requires CUDA 8.0 or higher
Recommend
- Windows, Mac, or Linux
- 2 GB of RAM
- Python 3.6 +
- CUDA 9.2 +
- Nvidia driver 440.33 or higher
- GPU Version requires mxnet-gpu:
pip install mxnet-cu100
How many cameras can a machine run?