ggml alpaca model don't work

[citizenofathens]

i download alpaca model into local "/models/" folder through this huggingface url

https://huggingface.co/Sosaka/Alpaca-native-4bit-ggml

and run examples/alpaca.sh

./main -m ./models/ggml-alpaca-7b-q4.bin
--color
-f ./prompts/alpaca.txt
--ctx_size 2048
-n -1
-ins -b 256
--top_k 10000
--temp 0.2
--repeat_penalty 1.1
-t 7 >> "$log_file"

program print this and it's over

main: build = 607 (ffb06a3)
main: seed = 1685605906

what should i do?

[SlyEcho]

Model file format is too old. The original model is linked in URL that you gave. It can be converted by the tools in llama.cpp already (convert.py, quantize)

[citizenofathens]

Model file format is too old. The original model is linked in URL that you gave. It can be converted by the tools in llama.cpp already (convert.py, quantize)

thank you! you godsend. i didn't even know origin model in my url
https://huggingface.co/chavinlo/alpaca-native

i download origin model and convert and quantize successfully

my try write down for other people
i convert and quantize the original model and run, but failed of same error

venv>python3 convert.py \models\{origin_huggingface_alpaca_reposity_files}
print Python.
it's over

venv>python convert.py \models\{origin_huggingface_alpaca_reposity_files}
this work.

and next,

first time my command was like README.md
venv>./main -m .\models\ggml-model-q4_0.bin -n 128
main: build = 607 (ffb06a3)
main: seed = 1685667571
it's over. don't work

venv>./main.exe -m .\models\ggml-model-q4_0.bin -n 128

now it's truely run

But still but there is a phenomenon that sometimes the same command failed Try twice, and it's run