sparverius @vec

🤖 automagician 🔮 huggingface.co/typeof ¤ Joined July 2023

Tweets

615
Followers

121
Following

397
Likes

2K

sparverius @vec

a year ago

It's all smoke and mirrors ... until the smoke be the mirrors ... @alignment_lab go burr 👀🚀👀🐢

0 0 1 36 0

sparverius @vec

a year ago

Y' know ... just screwing around with linear attention ... me 3 years ago, wtf that exists!?

0 0 0 29 0

sparverius @vec

a year ago

Probs 50% of my followers are fake but thanks to the 100 peeps 🙏😵‍💫👽🛸

0 0 0 15 0

sparverius @vec

a year ago

When you know the truth, all of x seems garbage... lol @alignment_lab

0 0 1 32 0

Jeremy Howard @jeremyphoward

a year ago

@Euclaise_ @n0riskn0r3ward If fine tuning was fast and easy enough then general purpose assistants would be of far less utility.

0 1 1 270 0

Alignment Lab AI @alignment_lab

a year ago

ANNOUNCEMENT TIME UpvoteWeb-24-600M ! build out by Grass Foundation! a majorly useful research artifact! UpvoteWeb-24-600M contains: 600 million of the top posts and comments from Reddit in 2024 media links! reply lineage! The data has been totally anonymized to preserve user…

10 31 113 24K 53

sparverius @vec

a year ago

arxiv.org/abs/2401.09414

Andrej Karpathy @karpathy

a year ago

arxiv.org/abs/2401.09414

361 183 4K 388K 949

0 0 0 34 0

sparverius @vec

a year ago

arxiv.org/abs/1908.04319

0 0 1 24 0

Adam Hibble @Algomancer

a year ago

Most of my feed is about generative modeling and AI. So, most of yours probably is as well. Something different, you'll need 10 minutes. Some good headphones or loud speakers. If you haven't listened to Maggot Brain by Funkadelics - close your eyes and pop it on. You only get…

0 1 12 824 4

Download Image

Vaibhav (VB) Srivastav @reach_vb

a year ago

Honestly fuck cloude 3.5 and GPT 4o - I’m way too excited by Florence2 200M, Qwen2 500M, MSFT InstructLM 500M With little fine-tuning they unlock so many creative and on-device use cases! Slowly, then suddenly! Everything crucial will become on-device.

19 41 397 49K 233

Thomas Capelle @capetorch

a year ago

@Prince_Canuma You may want to check our work with @__vec__ and @alignment_lab on prune (downcycle) the Mistral 7B model to smaller sizes: - We took a 7B and removed layers to obtain a model between 2~3B params - Trained on around ~1B tokens to recover the damage on a high quality dataset like…

1 1 2 127 0

Thomas Capelle @capetorch

a year ago

I had a blast last Friday participating in @HamelHusain and @dan_s_becker's LLM fine-tuning course. I tried my best to showcase some of the fine-tuning work people shared on Discord. Some of the workspaces shown are from @TheZachMueller @augmxnt @sophiamyang @Vapi_AI, but…

3 15 68 7K 74

Download Video

Jeremy Howard @jeremyphoward

a year ago

This is important, so don't just read this thread, instead read the 6000+ word article I just published. In the article I explain how AI *actually* works, and why these details totally break legislation like SB 1047. Policy makers *need* to know this: answer.ai/posts/2024-06-…