What's ggml-large-v3-q5_0.bin?

[playgithub]

ggml-large-v3-q5_0.bin is about 1.08GB,
ggml-large-v3.bin is about 3.1GB.
What's the difference?

[itsthisjustin]

The one with the Q is a quantized model

[playgithub]

What's a quantized model?

[jensdraht1999]

It's a model, where the data is "compressed" by removing redundant data. Since the data is removed, it is using less VRAM/RAM, takes less space on your device and works faster, because it's smaller. In the case of large3-v3, it is 3 times smaller.