NOTE: This documentation is currently being updated for IV 5.0 release on June 16, 2021.

The Intelligent Voice system is a batch processing system designed to process hours of audio recordings as efficiently as possible.  We also offer systems designed for real-time transcription, low latency keyword spotting, IoT devices and embedded applications and more - for details on requirements for these please contact us.

Minimum Requirement: single Virtual Machine with no GPU acceleration

Intelligent Voice can be installed in a single Virtual machine without GPU acceleration.  This system will be functionally identical to GPU accelerated systems but lower performance.  Suitable applications include:

1. Low volume batch processing (0-500 hours of audio recordings per day)
2. Functional system evaluation and compatibility testing
3. Software development and integration testing
4. Test and QA systems

The minimum requirements to be able to process around 24 hours of audio recordings per day are:

version 4.13

• 2 or more x86 vCPUs, Intel "Ivy Bridge" or later / AMD "Bulldozer" or later.
• at least 32Gb RAM
• at least 60GB storage
• Ubuntu 16.04 or Red Hat 7.5

version 5.0

• 4 or more x86 vCPUs, Intel "Ivy Bridge" or later / AMD "Bulldozer" or later.
• at least 64Gb RAM
• at least 500GB storage
• Red Hat 7.9 (fully supported) or Ubuntu 20.04 (preview support)

Additional vCPUs and storage are required to increase performance, up to around 500 hours of audio recordings per day per VM. 

For help on sizing larger installations please see below or contact us.

High Performance GPU Single Server

To get the best performance from Intelligent Voice we recommend servers with NVIDIA GPU cards.  

An example system specification for processing up to 10,000 hours of audio recordings per day:

• 2 x NVIDIA Tesla T4
• 2 x 1TB SSD
• 128 GB RAM
• 1 x AMD EPYC 7302P (or equivalent)

An example server spec for processing up to 30,000 hours of audio recordings per day:

• 4 x NVIDIA Tesla V100 32GB
• 2 x 2TB NVMe SSD
• 512GB RAM
• 2 x AMD EPYC 7402 (or equivalent)

Additional storage will be required for storage of audio data and outputs, with the size depending on audio / video file formats, and the required retention period.

High Performance GPU Multiple Servers

Intelligent Voice can scale over any number of GPU servers.  An example system for processing up to 100,000 hours per day:

1 x application server:

• 2 x 2TB NVMe SSD
• 512GB RAM
• 2 x AMD EPYC 7402 (or equivalent)

5 x GPU processing node servers:

• 4 x NVIDIA Tesla T4
• 2 x 1TB SSD
• 256 GB RAM
• 1 x AMD EPYC 7302P (or equivalent)

Larger Clusters & Geographical Distribution

Intelligent Voice can scale to millions of audio hours per day and has options of synchronizing over multiple geographic regions - please contact us for information