whisper.cpp/extra/quantize-all.sh
Georgi Gerganov b0502836b8
whisper : add full CUDA and Metal offloading (#1472)
* whisper : migrate to ggml-backend

* whisper : fix logit reading

* whisper : fix tensor allocation during load

* whisper : fix beam-search with CUDA

* whisper : free backends + fix compile warning

* whisper : print when CUDA is enabled

* whisper : fix CoreML

* make : clean-up

* talk : fix compile warning

* whisper : support ggml_conv with CUDA and Metal (#1473)

* ggml : add CUDA support for ggml_conv

* whisper : remove ggml_repeat for conv bias + single backend

* cuda : fix im2col kernel

* metal : add im2col support + mul mat-vec f16 x f16

* bench-all : add q4 models

* whisper : clean-up

* quantize-all : fix

* ggml : im2col opts

* whisper : avoid whisper_model_data wrapper

* whisper : add note that ggml_mul_mat_pad does not work with CUDA

* whisper : factor out graph compute in common function

* whisper : fixes

* whisper : fix UB with measure buffers

* whisper : try to fix the parallel whisper_state functionality (#1479)

* whisper : try to fix the parallel whisper_state functionality

* whisper : fix multi-state Metal

* whisper : free backend instances in whisper_state
2023-11-12 15:31:08 +02:00

38 lines
843 B
Bash
Executable File

#!/bin/bash
printf "Usage: $0 <upload>"
if [ $# -ne 1 ]; then
printf "\nError: Invalid number of arguments\n"
exit 1
fi
qtype0="q5_0"
qtype1="q5_1"
upload="$1"
declare -a filedex
cd `dirname $0`
cd ../
for i in `ls ./models | grep ^ggml-.*.bin | grep -v "\-q"`; do
m="models/$i"
if [ -f "$m" ]; then
if [ "${m##*.}" == "bin" ]; then
./quantize "${m}" "${m::${#m}-4}-${qtype1}.bin" ${qtype1};
./quantize "${m}" "${m::${#m}-4}-${qtype0}.bin" ${qtype0};
filedex+=( "${m::${#m}-4}-${qtype1}.bin" "${m::${#m}-4}-${qtype0}.bin" )
fi
fi
done
if [ "$upload" == "1" ]; then
for i in ${!filedex[@]}; do
if [ "${filedex[$i]:9:8}" != "for-test" ]; then
scp ${filedex[$i]} root@linode0:/mnt/Data/ggml/ggml-model-${filedex[$i]:9}
fi
done
fi