At #SC22, Dr. Dirk Van Essendelft from the National Energy Technology Laboratory presented a Python API to build applications that run on Cerebras CS-2 systems with impressive results – 470x faster than the Joule supercomputer! Watch his lecture:
CODE
-
Anthropic Explores Automating Language Model Evaluation
By
–
We explored approaches with varying amounts of automation and human effort. In the simplest case, we generated thousands of yes-no questions for diverse behaviors just by instructing an LM (and filtering out bad examples with another LM). Random examples of LM-written evals:
-
LangChain v0.0.41 Release: Multiple Inputs, Bug Fixes, Agent Improvements
By
–
LangChain v0.0.41 Make `run` interface work for multiple inputs (
@ankush_gola11
) Bug fix for text splitter (h/t @prof_reed
) Lots of agent improvements (see ) -
PySpark Profilers: Identifying Hot Loops and Performance Improvements
By
–
Why are we kinda obsessed with #PySpark Profilers? They help identify hot loops, suggest potential improvements, and are easy to use for enhancing performance! Take a look https://
dbricks.co/3RLA09p -
Hugging Face surveys users for open-source roadmap directions
By
–
You know what's better than twitter polls? User surveys to inform the long-term directions of @huggingface
's open-source roadmap! Please take 5 mins to answer this or share it if you've used our open-source libraries: https://
docs.google.com/forms/d/e/1FAI
pQLSf4xFQKtpjr6I_l7OfNofqiR8s-WG6tcNbkchDJJf5gYD72zQ/viewform?usp=sf_link
… -
Fine-tuning AI Models on Successful Response Patterns
By
–
Yep but wait till it's finetuned/trained on successful replies etc.
-
IBM Launches Digital Self-Serve Co-Create Experience for AI Applications
By
–
We’re excited to introduce the new Digital Self-Serve Co-Create Experience (DSCE) from #IBM Build Lab: https://t.co/gGek0zvYht
— IBM Data, AI & Automation (@IBMData) 19 décembre 2022
^ a prescriptive, wizard-driven experience designed to onboard, scale and differentiate partner applications with embeddable #AI.
—#AIOps #MLOps pic.twitter.com/F81oBh9W7AWe’re excited to introduce the new Digital Self-Serve Co-Create Experience (DSCE) from #IBM Build Lab: https://
ibm.co/3Wpgbbn ^ a prescriptive, wizard-driven experience designed to onboard, scale and differentiate partner applications with embeddable #AI.
—
#AIOps #MLOps -
AI Data Scraping: Programmers Must Track Sources and Copyrights
By
–
You're looking for complexity where there is none. "AI" does not go out in the wild to scrape data, programmers implement code that does it, and thus can & should track the source and copyrights. If there is no clear & usable copyright information, then the code drops it…
-
FITM definition for Copilot’s cursor placeholder
By
–
FITM = fill-in-the-middle, models trained to generate text to fill in a placeholder within the prompt rather than complete it from the end. For Copilot, the placeholder is the cursor.