What we could borrow from gguf to improve model-spec?

[caozhuozi]

I recently noticed the gguf.md: https://github.com/ggml-org/ggml/blob/master/docs/gguf.md#general-metadata
It appears that the GGUF format is more precise and strictly defined in certain areas. For instance, it provides a more accurate way to describe model size:

SizeLabel: Parameter weight class (useful for leader boards) represented as <expertCount>x<count><scale-prefix>
This can be derived from gguf metadata general.size_label if available or calculated if missing.
Rounded decimal point is supported in count with a single letter scale prefix to assist in floating point exponent shown below
Q: Quadrillion parameters.
T: Trillion parameters.
B: Billion parameters.
M: Million parameters.
K: Thousand parameters.

There might be elements from GGUF that we could adopt to improve the clarity and precision of model-spec.

[aftersnow]

Yes, can you make a PR to improve it?

[caozhuozi]

Sure, currently I can prepare a PR to get a more precise size definition. but maybe there are more elements we can learn from it.