Will a LLM trained with FP4 have competitive performance in 2 years time?

Plus

Ṁ1401

Jan 21

23%

chance

ALL

"Currently, the technology for 4-bit training does not exists, but research looks promising and I expect the first high performance FP4 Large Language Model (LLM) with competitive predictive performance to be trained in 1-2 years time." (see: https://timdettmers.com/2023/01/16/which-gpu-for-deep-learning/)

Granted, the model must be open source for us to know, so the market will resolve based on publicly available information.

This question is managed and resolved by Manifold.

Get

1,000

and

3.00

9 Comments

12 Holders

48 Trades

Sort by:

predictedNO

Exclusively in FP4? Or does partially in FP4 count. What if the model is on average 60% FP4 over the course of training?

I guess you covered this with "trained in 4-bit (to some extent)"

predictedNO

https://arxiv.org/pdf/2212.09720.pdf

predictedNO

@NoaNabeshima This is ab post-training precision adjustments

Competitive with what? SOTA with fp16?

predictedNO

This seems important @typedfemale
Will this resolve YES if scaling laws suggest a 4-bit model would be competitive if compute-matched to a SOTA 16-bit model?

predictedNO

(but there isn't a trained SOTA 4-bit model)

@NoaNabeshima Yes, you need to be better than everything else, but be trained in 4-bit (to some extent)

@typedfemale Finetuned w 4 bit would trigger Yes? 80% of parameters in 4 bit would trigger Yes?

Related questions

Related questions