keep on compiling.
Brooklyn, NY
Joined May 2009
- Tweets3.8K
- Following242
- Followers155
- Likes327
Alex Barkan retweeted
Decision models in llama.cpp are now available
The `/v1/systemone` endpoint is available in the latest llama builds. Use it to do Jev-style inference locally, efficiently and privately. Multiple open models are supported with more to come.
huggingface.co/blog/ggml-org…
Alex Barkan retweeted
Super excited to announce the release of Stable Video Diffusion (SVD) -- the first set of video models in the Stable Diffusion series. To start with, we release 14-frame (SVD) and 25-frame image-to-video (SVD-XT) models. The code/weights are already out!
SVD: huggingface.co/stabilityai/s…
SVD-XT: huggingface.co/stabilityai/s…
Code: github.com/Stability-AI/gene…
Paper: stability.ai/research/stable…
Alex Barkan retweeted
Large text-generation AIs require massive resources (Anthropic is raising billions to beat OpenAI), but visual AI is advancing with tiny teams.
Example: Blockade Labs is under a dozen people & generates 3D scenes from text. Free & no registration needed: skybox.blockadelabs.com/e6dc…
Alex Barkan retweeted
Meme Legends 3 🔥 Photoshop AI.
I hope you cheer up and remember the legends. 🫡
What memes would you like to see in the next part?
👍🏼 #memelegends #ai #savvydone
Alex Barkan retweeted
A survey of LLMs with a practical guide and evolutionary tree.
Number of LLMs from Meta = 7
Number of open source LLMs from Meta = 7
The architecture nomenclature for LLMs is somewhat confusing and unfortunate.
What's called "encoder only" actually has an encoder and a decoder (just not an auto-regressive decoder).
What's called "encoder-decoder" really means "encoder with auto-regressive decoder"
What's called "decoder only" really means "auto-regressive encoder-decoder"
github.com/Mooler0410/LLMsPr…
Alex Barkan retweeted
Enough with LLMs - exciting things are happening in the world of atoms.
This is Stanford ALOHA, a low-cost and agile robot platform. The whole system is open-source (!!): hardware design, CAD models for 3D printing, simulator, and training code. Time to graft a physical arm onto GPTs 🦾
Led by my friend @tonyzzhao at Stanford, and advised by @chelseabfinn @svlevine @Vikashplus.
Project page: tonyzhaozh.github.io/aloha/
If you don't want to 3D print your own components, you can buy the setup at trossenrobotics.com/aloha.as…
Alex Barkan retweeted
Plugins for processing a video clip, no ffmpeg wizardry required. Actual use-case from today's launch.
Alex Barkan retweeted
Pastiche Master: Exemplar-Based High-Resolution Portrait Style Transfer
project page: mmlab-ntu.com/project/dualst…
github: github.com/williamyang1991/D…
Alex Barkan retweeted
I created a vid to showcase my vehicle damage system, but I may have inadvertently created the most violent car destruction video ever...
Full Video on YouTube: youtu.be/sI8ekGStpj0
#madeInUnity #indiegame #gamedev #indiegamedev
Alex Barkan retweeted
#MarketOnClose ➡️ @fuboTV Co-Founder & CEO @davidgandler discusses the $FUBO IPO debut & the overall company outlook: 🎥
pscp.tv/w/ck4d3DFsWkVwZ3pyUF…
La Liga is back with 7 games going in parallel all across Bein Sports! They all have these Southpark-like virtual audiences. I think I see Cartman in there..
Thank you @Use_Screen for making the product free during the coronavirus outbreak. My team is going to try it for pairing and code-reviews.
Alex Barkan retweeted
We’re on @ProductHunt! Check us out and share your feedback: producthunt.com/posts/screen…
A GPU History Wall at the Direct3D offices in these trying times? devblogs.microsoft.com/direc…
Alex Barkan retweeted
March 17, 2019: 5,568,464
March 17, 2020: 1,785,252
3.7 million of you chose not to ride with us yesterday because you want to #flattenthecurve.
We miss you, but for now, we’ll say: thank you for not riding with us. You’re keeping NYC safer.
istio.io/blog/2020/wasm-anno… Can't tell if this is service-mesh jumping the shark or what. It's like npm.js for your Kubernetes sidecars! I can't even.
It's bad enough that New York doesn't have open data for its coronavirus cases; having to scrape ever changing html with silly typos (like surprise ":" suffixes) is just putting salt on a wound.
Instead of everyone with a jupyter notebook making pretty graphs, I wish the focus was on building a reliable data-pipeline for github.com/CSSEGISandData/CO… so we could have consistent US data (backwards compatible, backfilled, id-driven, normalized)
John Hopkins' GitHub COVID-19 dataset has been blowing up with issues and PRs due to everyone trying to understand and chart the data
github.com/CSSEGISandData/CO…
Alex Barkan retweeted
SIGGRAPH reviewers didn't find your industry research or systems paper appropriate?
jcgt.org/ is a good place to publish. You can submit in the SIGGRAPH template with your reviews. Great work from Disney, NVIDIA, Unity, Pixar, Terathon has already appeared there.