INT4/8 Quantized Whisper CT2 - a jvh Collection

jvh 's Collections

updated Mar 19

Int4/8 Quantized Whisper Models by using the quanto package and the CTranslate2 package. Requires (much) less GPU resources while keeping performance.

Upvote

jvh/whisper-base-quant-ct2

Automatic Speech Recognition • Updated Mar 19 • 18 • 2
jvh/whisper-large-v2-quant-ct2

Automatic Speech Recognition • Updated Mar 19 • 4 • 3
jvh/whisper-medium-quant-ct2

Automatic Speech Recognition • Updated Mar 19 • 5 • 2
jvh/whisper-large-v3-quant-ct2

Automatic Speech Recognition • Updated Mar 19 • 19 • 1

Upvote

Collection guide
Browse collections