For OPT-IML, we boosted performance of our OPT-175B work using instruction tuning, allowing us to adapt it for more diverse language applications (Ex. Answering Q’s, summarizing text & translation). A 30B parameter model is available today with a 175B model available soon.
INNOVATION
-
MoleculeSTM: ChatGPT Technology Applied to Molecular Analysis
By
–
MoleculeSTM paving the way for #chatGPT for molecules https://t.co/EtPho1oj0b
— Prof. Anima Anandkumar (@AnimaAnandkumar) 22 décembre 2022MoleculeSTM paving the way for #chatGPT for molecules
-
Largest Text-Molecule Model Enables ChatGPT-like Molecule Retrieval and Editing
By
–
We build the largest Text-molecule model that does not rely only on aligned training pairs. Now you can retrieve and edit molecules based on text prompts. This will pave the way for #ChatGPT for #molecules https://
chao1224.github.io/MoleculeSTM @nvidia @Mila_Quebec @Caltech -
MIT researchers link brain language systems to code ML representations
By
–
The brain has multiple demand and language systems responsible for various cognitive tasks. MIT researchers found that these systems align w/ML representations of code, potentially helping us understand how our brains read programs: http://
bit.ly/3vb5ehn -
OODA Loop 2022: ChatGPT GPT-3 OpenAI Future Analysis
By
–
OODA Loop 2022: The Past, Present, and Future of ChatGPT, GPT-3, OpenAI, NLMs, and NLP https://
oodaloop.com/ooda-original/
disruptive-technology/2022/12/22/ooda-loop-2022-the-past-present-and-future-of-chatgpt-gpt-3-openai-nlms-and-nlp/
… via @ooda -
spaCy Extension for Zero-Shot Labeling with Prodigy Integration
By
–
Following up with the theme of zero and few-shot labelling, this spaCy extension by @IBMResearch is super cool, and could be used in combination with our recent Prodigy recipe: https://
github.com/explosion/prod
igy-openai-recipes
… -
Utilities Automate Asset Performance Management with Digitalization
By
–
By digitalizing and automating asset performance management, utilities can detect, react and resolve issues faster, even preemptively. Find out more: http://
ow.ly/l38U50Ljqir @CyrilCoste @HaroldSinnott via @fogoros #hitachivantara_iiot #sponsored #IIoT #digitaltransformation -
SantaCoder: First LLM with Opted-Out Training Dataset
By
–
So cool to see @BigCodeProject release SantaCoder, afaik the very first LLM trained on an "opted-out" dataset, aka allowing people to opt-out from the training dataset. 1.1B parameters that outperforms larger models on both generation and infilling! https://
huggingface.co/bigcode/santac
oder
… -
Building a World-Class Analytics Center of Excellence
By
–
For a mature analytics Center of Excellence, look no farther than at @Allstate
. In our latest blog post, Meg Walters, Senior Manager and Data Science Leader of its Analytics Center of Excellence offers real-world advice on building a world-class CoE: https://
domino.buzz/3v7EM8z
