CPU vs. GPU: Availability of Accelerated Speech Technologies

Our platform leverages a range of technologies optimized for different processing units. Some components may run efficiently on CPUs, while others take advantage of GPU acceleration for intensive computations. The table below summarizes which technologies are associated with each type of hardware

Feature	CPU	GPU
Speaker Identification	✅	✅
Deepfake Detection	✅	✅
Speech to Text Phonexia	✅	✅**
Enhanced Speech to Text Built on Whisper	✅	✅*
Enhanced Speech Translation Built on Whisper	✅	✅*
Language Identification	✅	✅
Speaker Diarization	✅	✅
Keyword Spotting	✅	✅
Authenticity Verification	✅	✅
Gender Identification	✅	✅
Age Estimation	✅	✅
Emotion Recognition	✅	✅
Denoiser	✅	❌
Voice Activity Detection	✅	✅
Audio Quality Estimation	✅	✅
Time Analysis of Speech	✅	❌

* Highly recommended to be run on GPU
** Not significant difference in speed if running on a GPU