Despite being 20x smaller than the top model, FlowState ranks #2 & #3 in zero-shot performance on the @huggingface GIFT-Eval leaderboard. A time-scale adjustable TSFM, FlowState sets a new benchmark in time series forecasting: https://t.co/dIVRwCJgr4

@ambitioninc @muradhem you should do a piece on TSFM and @internetvin and what’s changing in the tech ecosystem.

🆕 Time Series Foundation Models: What You Need to Know 1️⃣ Choosing the Right TSFM for Your Mission 2️⃣ The Problem with Adapting LLMs for Forecasting 3️⃣ Exploit enterprise relational graphs before chasing external text 👉

https://t.co/iT7nC2OvK2

時系列基盤モデル(TSFM: Time Series Foundation Model)ね、ふむふむ。

Wanna buy TSFM merch? https://t.co/OirPmrWTRm

A shoutout to all the people that helped make TSFM season 1 happen. Can’t overstate how grateful I am for the friends I’m surrounded by

TSFM lectures were recorded using Boom! I had enough things to think about during the series so it was a great relief that Boom made it so easy to record with great production quality

Kronos 是首个开源的金融 K 线（蜡烛图）基础模型，基于来自全球 45 多家交易所的数据训练而成。 它是一族仅解码器（decoder-only）的基础模型，专为金融市场的“语言”——K 线序列而预训练。与通用时间序列基础模型（TSFM）不同，Kronos

Shoutout to the man who helped me document the TSFM journey, the cohort, the speakers and everything else @cairox100v Also very grateful to @newsystems_ for giving me their space every Thursday for 3 months straight - no questions asked.

I was impressed with the quality of the question/discussion at TSFM. @sid_srk has got something special going. Made me really miss Toronto 🥺

শিউলি ফুল শীতের আবির্ভাব 🌸

@akaash @ambitioninc ### Canada's AI Edge: Why Programs Like TSFM Are Key to Capturing AI Value In the global AI race, Canada has long been a powerhouse, birthplace of pioneers like Geoffrey Hinton and home to hubs like Toronto's Vector Institute. Yet, translating research into economic wins remains

@goodalexander @const_reborn @DeanBuilds22 Yea I’m always training for no. Stationary distributions using TSFM and genetic algorithms. You should also use GARCH and multiscale fractal distributions for the tails. And don’t forget about rough path theory!

Excited to catch up with the season 1 cohort! you know where to find this

📌 #TSFM

Are Time-Indexed Foundation Models the Future of Time Series Imputation? Etienne Le Naour, Tahar Nabil, Adrien Petralia, Ghislain Agoua. Action editor: Jes Frellsen. https://t.co/W2WP1Drf9n #imputation #tsfm_imputation #datasets

The story of TSFM season 1 (and an easter egg about what I’ll be launching next) From my talk at @ctrlshift_ai last year.

i will be giving an automatic differentiation lecture at tsfm on nov 18