lightningRalf / whisper.cpp

Port of OpenAI's Whisper model in C/C++
MIT License
0 stars 0 forks source link

runs perfectly with the regular models, but not the quantized ones #89

Open lightningRalf opened 1 year ago

lightningRalf commented 1 year ago

I have run quantization on ggml-small.en.bin to produce ggml-small.en.bin-q5_0.bin Quantization proceeded without any errors.

When I run the model, it attempts to load, but throws the following: GGML_ASSERT: ggml.c:4288: wtype != GGML_TYPE_COUNT Abort trap: 6

Same issue occurs when I try 8-bit quantization.

Running on a 2015 Macbook Air, if this is any use.