3 ms·
Depends how much you quantize the model. For most general-purpose LLMs, the model completely dwarfs the size of the binary code.
by samus 3y ago
Depends how much you quantize the model. For most general-purpose LLMs, the model completely dwarfs the size of the binary code.