Latent Space

Building the next wave of generative models for businesses and creatives. Series A from Greylock and General Catalyst.

[Latent Space @LatentSpaceAI]

Catch @DarrylBarnhart speak @NVIDIAGTC today about our exciting work @LatentSpaceAI on scaling LLM with retrieval! 📚

There’ll be an open QA in 30 mins: https://t.co/C42yjpjEi1

[Ethan Perez @EthanJPerez]

Very cool to see Retrieval-Augmented Generation models being scaled up by @LatentSpaceAI and @awscloud. Excited to see what happens when RAG is pushed to GPT3 scale

[Ian Thompson @iantbd]

We @LatentSpaceAI have been making exciting progress on improved neural scaling laws, accelerated by a great collaboration with Amazon!

This post is the first of many on the nuts and bolts of how we’re doing this: https://t.co/FnkWifFgOE

[sarah hong ✨ @latentcodes]

Did a fun panel for #AWSInnovate conference where we talk about our triumphs and missteps of starting an ML startup🤭

It's live today, check it out (need registration tho -- hey, I don't make the rules!): https://t.co/PxT7RgsdNn

[sarah hong ✨ @latentcodes]

TL;DR of the discussion:

💥how well a model scales is better than chasing sota

🧑‍🎨 combining language + vision is good for generating images (and text!)

👾 the code don’t lie — check other resources first before going deep on a research paper

[sarah hong ✨ @latentcodes]

We 💕 @wandb!

The multi-user reports let’s us have a collaborative research process which is really a game-changer, even more so now that we @LatentSpaceAI are all remote!

[Latent Space @LatentSpaceAI]

Listen to this discussion with Sarah Jane Hong @latentcodes where she talks about the research process, scaling laws, and future of large scale models on the latest @ai_untitled podcast!

[Imbue @imbue_ai]

Our second episode of Generally Intelligent features Latent Space co-founder @latentcodes! https://t.co/uVJeBHkPSB

We cover why using natural language prompts to render a scene is much harder than you’d expect, Sarah's own learnings and mistakes as a researcher, & much more!

[Ian Thompson @iantbd]

Come by at @NeurIPSConf and chat about spatially varying representations!

[sarah hong ✨ @latentcodes]

We’ll be presenting our paper “Low Distortion Block-Resampling with Spatially Stochastic Networks” today at 9pm PST🧩🏰🎨

Joint work w/ @martinarjovsky @darrylbarnhart @iantbd @LatentSpaceAI!

➡️ paper: https://t.co/ZIUpj4iKNt
➡️ poster: https://t.co/SNbOCs9xqk

#NeurIPS2020

[Latent Space @LatentSpaceAI]

We are excited to present our work "Low Distortion Block-Resampling with Spatially Stochastic Networks" @NeurIPSConf 🎉🎉🎉

Come by today 9pm PST to discuss joint work by @latentcodes @martinarjovsky @DarrylBarnhart @iantbd

📝paper: https://t.co/Oqk77Bvdo2

#NeurIPS2020

[Arram @arram]

The Generative Age: AI will do for content production what the internet did for distribution.

What happens when the cost to produce a movie like The Avengers goes from $350M to 30 cents?

https://t.co/NksEF9IAjQ

[sarah hong ✨ @latentcodes]

If you're curious for more details on the above, check out this piece I wrote on how we're collaborating remotely @LatentSpaceAI!

Being able to log ~all the things~ with @wandb and share interactive artifacts with our team has been important for our research process.

[sarah hong ✨ @latentcodes]

Laplacian pyramid-like visualization of activations (as described in MSG-GAN by @AnimeshKarnewar) has been useful to observe during training in order to qualitatively identify new behaviors in our models (especially subtle bugs 🐛)

[Lukas Biewald @l2k]

.@latentcodes shares a little bit about how they collaborate on and debug GANs and other large ML models at a startup https://t.co/lpDvQcJJfs

[Ian Thompson @iantbd]

Here's a snippet of how we're collaborating remotely on debugging generative models @LatentSpaceAI ! @wandb has been an invaluable tool for understanding and testing new approaches.

[Lavanya @lavanyaai]

See how ML teams, like @LatentSpaceAI, working on cutting edge deep learning research use W&B Reports to collaborate remotely, debug models iteratively and to showcase results as mini research papers.

#machinelearning #deeplearning #100daysofmlcode
https://t.co/OkAbp5MEwz

[Ian Thompson @iantbd]

Looking forward to this remote discussion tomorrow on visual representation learning! #FlattenTheCurve

[Ian Thompson @iantbd]

The creativity workshop was a highlight of #neurips19 for me! Shoutout @ada_rob @jesseengel @the_asir @yaroslav_ganin and all the others who had great work here!

[sarah hong ✨ @latentcodes]

Hey optimization folks!

Catch @DarrylBarnhart and @yaroslavvb in between #NeurIPS2019 workshops today and tomorrow, esp if you’re interested in second order methods! 🌟

[Ian Thompson @iantbd]

@LatentSpaceAI @latentcodes @DarrylBarnhart and @yaroslavvb are here too! Catch them in between workshops :)