Int4/8 Quantized Whisper Models by using the quanto package and the CTranslate2 package. Requires (much) less GPU resources while keeping performance.
-
jvh/whisper-base-quant-ct2
Automatic Speech Recognition • Updated • 11 • 2 -
jvh/whisper-large-v2-quant-ct2
Automatic Speech Recognition • Updated • 17 • 3 -
jvh/whisper-medium-quant-ct2
Automatic Speech Recognition • Updated • 9 • 2 -
jvh/whisper-large-v3-quant-ct2
Automatic Speech Recognition • Updated • 12 • 1