Google's 200M-parameter time-series foundation model with 16k context

Re:Google's 200M-parameter time-series foundation model with... by EmilStenstrom on Tuesday March 31, 2026 @01:41AM (#47583161)

Here is the link to the blogpost, that actually describe what this is: https://github.com/google-research/timesfm?tab=readme-ov-fil...

[ Parent | on HN ]

Re:Google's 200M-parameter time-series foundation model with... by OliverGuy on Tuesday March 31, 2026 @03:14AM (#47583808)

Wish they gave some numbers for total GPU hours to train this model, seems comparatively tiny when compared to LLMs so interested to know how close this is to something trainable by your average hobbyist/university/small lab

[ Parent | on HN ]

Re:Google's 200M-parameter time-series foundation model with... by OliverGuy on Tuesday March 31, 2026 @03:27AM (#47583894)

Edit, it looks like the paper does

TPUv5e with 16 tensor cores for 2 days for the 200M param model.

Claude reckons this is 60 hours on a 8xA100 rig, so very accessibile compared to LLMs for smaller labs

[ Parent | on HN ]

Re:Google's 200M-parameter time-series foundation model with... by nels on Tuesday March 31, 2026 @01:56AM (#47583265)

I think you meant to link this page: https://research.google/blog/a-decoder-only-foundation-model...

[ Parent | on HN ]

Re:Google's 200M-parameter time-series foundation model with... by refulgentis on Tuesday March 31, 2026 @01:53AM (#47583244)

That takes me to the same content as the submission, a GitHub repo (Chrome on iOS)

[ Parent | on HN ]

Re:Google's 200M-parameter time-series foundation model with... by rockwotj on Tuesday March 31, 2026 @01:56AM (#47583267)

Probably the better link: https://research.google/blog/a-decoder-only-foundation-model...

[ Parent | on HN ]

Re:Google's 200M-parameter time-series foundation model with... by akshayshah on Tuesday March 31, 2026 @01:57AM (#47583278)

And https://arxiv.org/pdf/2310.10688 if you want the full paper.

[ Parent | on HN ]

Re:Google's 200M-parameter time-series foundation model with... by Cyuonut on Tuesday March 31, 2026 @01:57AM (#47583272)

I suppose they tried to link this: https://research.google/blog/a-decoder-only-foundation-model...

[ Parent | on HN ]