Sabitlenmiş Tweet
enty
3.4K posts
enty
@chronurgist
no themes. only such model pretraining and gpu kernels
him Katılım Mayıs 2024
88 Takip Edilen112 Takipçiler
someone’s gonna put this on gpt 2 speedrun in like 10 minutes
Zhaofeng Wu@zhaofeng_wu
Introducing ><former Most transformers are rectangles◻️: every layer has the same width But is that optimal?🤔 We propose variable-width transformers that have different widths across layers, improving loss while cutting compute & KV cache size 🧵
English
these look really cool on big projects but when I plug (most of) my projects in into one of these that shit looks like a straight line and it makes me sad as hell
ruru@ruru_1x
Git graph with timeline. 📈
English
To continue the celebration, we have added GLM 5.2 support to ncode and the noumena platform and are making it free to use for the next week (or so) with your code.noumena.com account . please clone and rebuild the latest version of ncode from github.com/Noumena-Networ… and select GLM 5.2 from the /models slash command . hope y'all enjoy the tokens !!!
xjdr@_xjdr
Today marks the beginning of our launch calendar and to celebrate i am making ncode and our flagship kimi k2.7 model free to use for the next week (or until the traffic knocks us out). all you need to do is: 1) sign up for a noumena account at code.noumena.com 2) go to github.com/Noumena-Networ… and clone and build noumena code (ncode) 3) login to the platform with `ncode auth login` (or /login once you are in the app) 4) enjoy blazing fast tokens on the noumena platform with ncode
English
enty retweetledi
enty retweetledi
Incredible how Z. ai literally has their RL infrastructure open source.
The entire OPD post-training of GLM-5.2 took on this slime platform took ~2 days.
github.com/THUDM/slime
English
@jameshorvatt how much for a video of you calling me homophobic slurs in johnny bravo's voice serious inquiry
English
enty retweetledi
enty retweetledi
the most important thing to keep in mind about the PPO vs GRPO debate is that nobody agrees on what either algorithm even is
English
enty retweetledi
we're looking at sweet frontier secrets in there btw 🤗
Yacine Mahdid@yacinelearning
What Makes Good Synthetic Pretraining Data with Joël Niklaus from Hugginface x.com/i/broadcasts/1…
English
enty retweetledi
enty retweetledi
At OpenAI, we're continuing to bet on Rust as the future of systems programming.
I'm proud to announce that we're making a $600,000 commitment to the Rust Foundation, which combines our Platinum membership with additional support for maintainer efforts across the Rust ecosystem.
English
enty retweetledi
enty retweetledi
i made a map of everyone on twitter!
yes you're on there too ^w^
every account is placed next to the people they talk to, so you can find out where you are, which cluster claimed you, and exactly who you're stuck next to
atlas.tiago.zip/?ref=launch_tw…
English
so basically have fable 5 “fix” the code, then a human can look at the diff and reverse their way through the patches and create a script that targets to vulnerabilities in the original (insecure) code? Lmao
bling@blingdivinity
this is the "jailbreak" that got Fable shut down you’ve been prepared for the singularity for years. for AGI to change everything. for nation-states to go to war over GPUs. but were you prepared for it to be this retarded?
English