3d attention is available in flash attn 2 now fyi
CODE
-
Professional struggles with malfunctioning web applications and system defeats
By
–
I swing into motion as Stand Back Honey, I Am A Professional At Coercing Bad Web Applications To Actually Work. I make two calls to the office during this. The system casually defeats me, too.
-
Electronic Medical Records System Fails on Common Pediatric Care Edge Case
By
–
So of course the electronic medical records system barfs on the edge case of a parent being responsible for the healthcare of a minor child, because when does that ever happen.
-
Text-to-Image Synthesis Models Now Available for Developers
By
–
For developers eager to run text-to-image synthesis on their own GPUs, the exciting news is that the method and pre-trained models are now accessible, inviting a much broader audience to experiment with these powerful tools.
-
YOLOv8 Models Comparison Using KerasCV Video
By
–
🎥New Video Alert🎥https://t.co/wkX8Cf1oaP
— Satya Mallick (@LearnOpenCV) 13 novembre 2023
Check out our latest video as we compare the YOLOv8 models using KerasCV. #computervision #ai #deeplearning #objectdetection #yolov8 #learnopencv pic.twitter.com/ged780xfyYNew Alert https://
youtube.com/watch?v=73gPZX
xUHO8
… Check out our latest video as we compare the YOLOv8 models using KerasCV. #computervision #ai #deeplearning #objectdetection #yolov8 #learnopencv -
Geospatial Data Science with Julia Programming Language
By
–
Geospatial Data Science with Julia https://
bit.ly/3rY5ClA
#AI #MachineLearning #DeepLearning #LLMs #DataScience -
Tokenization makes exact repetition of tokens harder than counting
By
–
A few things:
– “aaaa…” is tokenized differently from “ a a a a…” in a way that makes it much easier to say (many a’s = 1 token)
– Counting “out loud” is easy, including FizzBuzz variants. What’s hard is making it say exactly 100 repetitions of a token sequence unnumbered -
LLM Hallucinations: GPT-3, GPT4All, and Google get Super Bowl winner wrong
By
–
Oct 2022: I ask GPT-3, who won the Super Bowl the year Justin Bieber was born? GPT-3 says Packers. It’s the Cowboys. Example is added to LangChain. Other LLMs answer. Answers are added to GitHub. Now, citing GPT4All, Google says it’s the Buffalo Bills — who have never won:
-
GQA Architecture in 70B Language Models Explained
By
–
Oh wait this article is correct for a single layer of a 13b model — in fact it's only the 70b model that uses GQA AFAICT! 😀
