bug: Custom model not loading.

[NeuralAIM]

Describe the bug

Code: !ls /content Out: ggml-model.bin sample_data whisper whisper.CPP

Code:

from whispercpp import Whisper
w = Whisper.from_pretrained("/content/ggml-model.bin")

Out:

[<ipython-input-16-daf3870535af>](https://localhost:8080/#) in <cell line: 4>()
from whispercpp import Whisper
w = Whisper.from_pretrained("/content/ggml-model.bin")

/usr/local/lib/python3.10/dist-packages/whispercpp/__init__.py in from_pretrained(model_name, basedir, no_state)

RuntimeError: '/content/ggml-model.bin' is not a valid preconverted model. Choose one of ['tiny.en', 'tiny', 'base.en', 'base', 'small.en', 'small', 'medium.en', 'medium', 'large-v1', 'large']

To reproduce

I downloaded the model and converted it using a script:

!python whisper.cpp/models/convert-h5-to-ggml.py /content/whisper-small whisper ./

Expected behavior

No response

Environment

python 3.10.6 Google Colab

[kornpow]

I have a similar problem: In the main whisper.cpp repo I did:

bash ./models/download-ggml-model.sh tiny.en

which downloads the model as ./models/ggml-tiny.en.bin