Nuva Lab

[Nuva Lab @nuvalab]

WhatIf you put #FastH3 into the hands of anime lover.. Any #luckystar fans? 😜

Based on @MiniMax_AI H3 and open sourced FastH3 preview from @haoailab

[Nuva Lab @nuvalab]

@ostrisai @haoailab open source must win ;)

[Nuva Lab @nuvalab]

@kiyoshi_shin it can and acceleration applies to all GPU types too, just works better for blackwell as it has better kernel support. in subsequent versions we should take nvfp4 more seriously to make it easier to run on consumer GPUs :)

[Nuva Lab @nuvalab]

That was way better than what we've expected from the preview test with 1/10 of planned compute and data. WhatIf we scale it up properly 😜

[Hao AI Lab @haoailab]

45% preferred FastVideo FastH3 or are neutral compared with @MiniMax_AI H3.

Our open-weight FastH3 achieves 15s generation in 13s and 14x speedup. 🚀

Big shoutout to the folks at @fal and @isidentical for helping us run an eval against H3. As an open source project we always appreciate all the help we can get.

The entire team is quite excited by this result for a preview checkpoint trained on a fraction of our planned total compute and partial data. Excited for what our next versions can do!!

Weights: https://t.co/xMcHi0te0k
Blog: https://t.co/SBYKYq5tud

Thank you for your contribution to the open source H3 ecosystem @fal!

https://t.co/EHyNYZR5pG

[Nuva Lab @nuvalab]

A great start with many, many more to come! Omni ref, quality improvements, nvfp4, model customization and deployment, etc 🥳

[Hao AI Lab @haoailab]

(1/6) Open Weight @MiniMax_AI FastH3 v1: Generate 15s 768p video in 13s 🚀
- FastVideo collab w/ @nuvalab + FastGen - Up to 14x speedup on @NVIDIAAI Blackwell GPU
- Fully open so community can run and improve the acceleration recipe. The era of open weight video models just started, that calls for post training.

With great community effort such as Minimax H3 Max based on @MiniMax_AI , we have more to share about how post training can improve speed, quality and most importantly, how it’s done, with open weights and recipes.

[Nuva Lab @nuvalab]

@shuxiaoshushu Thanks for reading! Yes I'm aware of these and we even considered one GTM item to stream agents with real-time user interaction :)

[Nuva Lab @nuvalab]

@danielgross @JaneStreetGroup ChatGPT, take this training data! (pat)

[Nuva Lab @nuvalab]

@DynamicWebPaige @SingularMattrix Which app is this 😂

[Nuva Lab @nuvalab]

@Ning_Catsnail @saranormous Then I saw the `*good code` part and started questioning my commits 😂

[Nuva Lab @nuvalab]

@DynamicWebPaige You need stardew valley :)

[Nuva Lab @nuvalab]

@charliermarsh I just installed it for my vscode and feeling quite good for the first 5mins. Thanks for making this !

[Nuva Lab @nuvalab]

@TaliaRinger A hug from Femboy Emacs Dick Sucker is priceless

[Nuva Lab @nuvalab]

@LLCgame_ There is one interesting fact about videos we post every ~2 weeks or so ... all of them were new abilities that we built and tested during these 2 weeks, as a mechanism to keep us iterating :)

[Nuva Lab @nuvalab]

@LLCgame_ Thanks ! Still A LOT to improve 😉

[Nuva Lab @nuvalab]

@saranormous My wife has been telling me the same since her pregnancy

[Nuva Lab @nuvalab]

@LLCgame_ Just checkout the link!

[Nuva Lab @nuvalab]

@DynamicWebPaige We're building AI Agent native experiences and while building the e2e "game", a multi-agent framework emerged from it 😂

[Nuva Lab @nuvalab]

@LLCgame_ Thank you 😆

[Nuva Lab @nuvalab]

@mynameis_davis Please go to my competitor to negotiate with a discount on a $36 monthly. Keep their sales busy and less productive.

[Nuva Lab @nuvalab]

@jiayq I see what you did there 😉