New blog post: Collective Intelligence for Deep Learning
— hardmaru (@hardmaru) October 4, 2022
Recently, @yujin_tang and I published a paper about how ideas like swarm behavior, self-organization, emergence are gaining traction in deep learning.
I wrote a blog post summarizing the key ideas:https://t.co/S644KjM20e pic.twitter.com/1TYmuMPbKA
How diffusion models work: the math from scratch | AI Summer
— Jean de Nyandwi (@Jeande_d) October 3, 2022
A great article on diffusion models by AI summer(@theaisummer). Diffusion models have been the primary building block of first-class works in image generation and beyond. Worth reading about!!https://t.co/BzkZbczkc4 pic.twitter.com/uK9UGlrnCG
Spotted at Tesla AI day https://t.co/L2KUWhgr5Z pic.twitter.com/W1L2EWriXA
— Jeremy Howard (@jeremyphoward) October 1, 2022
The @Gradio Demo for CodeGeeX, a large-scale multilingual code generation model with 13 billion parameters, pre-trained on a large code corpus of more than 20 programming languages is now on @huggingface Spaces
— AK (@_akhaliq) September 30, 2022
demo: https://t.co/wzPItu8T40 pic.twitter.com/npOONYlsw5
To understand the foundations of NLP (pre-Transformers), where would you go?
— Radek Osmulski 🇺🇦 (@radekosmulski) September 28, 2022
This 48-page paper is the answer 🤩
✅ concise and clear explanations
✅ sklearn, spacy, and keras code snippets
✅ all the fundamentals of NLP in a single placehttps://t.co/DiZtrkdkZ3 pic.twitter.com/8LJ7v1DtCy
You can now finetune Dreambooth #stablediffusion on colab instances with bitsanbytes 8-bit Adam and xformers. https://t.co/1XxWaDwWH2
— Tim Dettmers (@Tim_Dettmers) September 28, 2022
Stay tuned. We are working on mem-efficient fine-tuning for LLMs. Should enable easy finetuning of OPT-175B/BLOOM on a single machine.
An attempt to summarize each modeling mindset in 3 words max. 👇
— Christoph Molnar (@ChristophMolnar) September 28, 2022
• Statistical modeling: Reason under uncertainty
• Bayesian inference: Update beliefs
• Frequentist inference: Estimate population parameters
• Likelihoodism: Weigh evidence
UniCLIP: Unified Framework for Contrastive Language–Image Pre-training
— AK (@_akhaliq) September 28, 2022
abs: https://t.co/7s5k4jYDIL pic.twitter.com/Ky3g54UFhj
You can run BLOOM or OPT-176B in Transformers without a 🖥️supercomputer thanks to Accelerate, but how does it work exactly?
— Sylvain Gugger (@GuggerSylvain) September 27, 2022
New blog post that dives deep into the awesome PyTorch features we use for this (also see summary in thread🧵)https://t.co/6UyNJ5I7U9
A new version of the {gt} package has been released! 🎉
— Posit PBC (@posit_pbc) September 26, 2022
Version 0.7.0 has many new things for #rstats tables: export {gt} tables to Word docs, format vectors with vec_fmt_*(), enhance HTML accessibility, and more!
Learn about them in the blog post: https://t.co/WW2Ub6g9fv
fast-stable-diffusion colabs, +25% speed increase + memory efficient
— AK (@_akhaliq) September 26, 2022
github: https://t.co/QO7UFwqXjf
colab: https://t.co/b8JiLOhYmJ pic.twitter.com/1hKZW7B3IL
Promptagator: Few-shot Dense Retrieval From 8 Examples
— AK (@_akhaliq) September 26, 2022
abs: https://t.co/YSYlD5IQHU
LLM prompting with no more than 8 examples allows dual encoders to outperform
heavily engineered models trained on MS MARCO like ColBERT v2 by more than 1.2 nDCG on average on 11 retrieval sets pic.twitter.com/X4sqN9RHtY
