Migrations
/
ggml
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from '4c1032f2d4'
${ noResults }
ggml/examples/whisper
History
Georgi Gerganov 4c1032f2d4
whisper : mem usage based on model format type 		2 years ago
..
CMakeLists.txt whisper : add whisper-qunatize tool 2 years ago
README.md Improve mul_mat performance for big matrices using Accelerate framework 2 years ago
common.cpp sync : latest whisper.cpp (scratch buffers in ggml) 2 years ago
common.h sync : latest whisper.cpp (scratch buffers in ggml) 2 years ago
convert-pt-to-ggml.py whisper : add whisper-qunatize tool 2 years ago
main.cpp sync : latest whisper.cpp 2 years ago
quantize.cpp whisper : add whisper-qunatize tool 2 years ago
whisper.cpp whisper : mem usage based on model format type 2 years ago
whisper.h sync : latest whisper.cpp 2 years ago

README.md

whisper

Port of OpenAI's Whisper ASR model in C/C++ using ggml

More info

Checkout https://github.com/ggerganov/whisper.cpp

Memory usage

Model Disk Mem
tiny 75 MB ~280 MB
base 142 MB ~430 MB
small 466 MB ~1.0 GB
medium 1.5 GB ~2.6 GB
large 2.9 GB ~4.7 GB

ggml format

The original models are converted to a custom binary format. This allows to pack everything needed into a single file:

  • model parameters
  • mel filters
  • vocabulary
  • weights

For more details, see the conversion script convert-pt-to-ggml.py