While cool, quantization to FP4 is practically never lossless in actual use. A lot of providers are advertising high TPS on Kimi and GLM, but the models are functionally lobotomized and no longer close to frontier quality. Would love to see this not be true.
Kimi uses INT4 as its native format, there's no such thing as "better than 4-bit precision" for that model. This is in contrast with GLM for which 16-bit precision is native and 8-bit is in common use.
Performance per dollar is getting faster and cheaper
91–100 of 151 posts
Re: Performance per dollar is getting faster and cheaper
#92Re: Performance per dollar is getting faster and cheaper
#93Earlier quoted context omitted.
The fact that people have lived and worked near data centres for decades and didn't even know what the term meant - let alone be adversely impacted by them - probably indicates they're broadly an non issue. All of a sudden out of nowhere, AI and data centres got intermingled by the media and now people seem to have big issues with them.
Though, the new data centers are not entirely the same. Increasing use of onsite gas turbines to generate power instead of using grid power changes their noise+air pollution profile.
Re: Performance per dollar is getting faster and cheaper
#94then why is token per dollar getting more expensive?
Re: Performance per dollar is getting faster and cheaper
#95I think we should make it illegal to not specify the quantization in the headline for these types of posts.
Re: Performance per dollar is getting faster and cheaper
#96I think we should make it illegal to not specify the quantization in the headline for these types of posts.
Re: Performance per dollar is getting faster and cheaper
#97Re: Performance per dollar is getting faster and cheaper
#98Re: Performance per dollar is getting faster and cheaper
#99Earlier quoted context omitted.
Though, the new data centers are not entirely the same. Increasing use of onsite gas turbines to generate power instead of using grid power changes their noise+air pollution profile.
The problem these days is lack of nuance. It should seem entirely reasonable to be pro-datacenters-if-they're-done-right, but it feels like there are only two sides to any issue. Gas turbine whine noise isn't coming from the data center, it's being used to power the data center, but the camp is either pro data center or not, and fuck any nuance.
Re: Performance per dollar is getting faster and cheaper
#100Earlier quoted context omitted.
Though, the new data centers are not entirely the same. Increasing use of onsite gas turbines to generate power instead of using grid power changes their noise+air pollution profile.
The problem these days is lack of nuance. It should seem entirely reasonable to be pro-datacenters-if-they're-done-right, but it feels like there are only two sides to any issue. Gas turbine whine noise isn't coming from the data center, it's being used to power the data center, but the camp is either pro data center or not, and fuck any nuance.