TIL Test-Time Scaling - It turns out there is a term for the concept that LLMs can produce better results by giving it more compute time - essentially the... #LLM
TIL Another thing I learned while diagnosing my warping 3D prints was VFAs or Vertical Fine Artifacts. I came across Prusa's post about this a few weeks a... #3D Printing
TIL I've been getting back into doing more 3D printing in order to build my robot arms (more on that in a future post). But, some of the larger prints hav... #3D Printing
TIL I just found out about KServe v2 Protocol. Late to the party I guess. It looks like ML/AI serving APIs are being standardized. I remember when Justin ... #MLOps
TIL Google published Primer: Searching for Efficient Transformers for Language Modeling in 2021 that describes the an variant of the ReLU activation funct... #deep learning
TIL optimize_for_inference isn't a function that is usually presented in beginner materials for pytorch. I just read about it in Aurélien Géron's new edit... #pytorch