GLM-5.2: Chop off 84% of the volume from a 1.5TB model, still retain 82% power

Y	Hacker News new \| ask \| show \| jobs

	GLM-5.2: Chop off 84% of the volume from a 1.5TB model, still retain 82% power (twitter.com)
	6 points by vantareed 3 days ago

1 comments

aurenvale 3 days ago

238GB is the 4-bit quantized version, with an accuracy loss of about one tier compared to the original. For running complex logical reasoning, it's still recommended to use the original version~

link