Joshua Goldberg

@yougetaquote @SC_Griffith Reduced row echelon. I think it’s computationally faster too.

@krasul huggingface.co/huggingface/ti…

@GoldbergData perhaps not the best medium to answer questions but let's see... if you have a look at the generate function it never uses the ground truth samples and thus you just need to use generate with the last context window together with the future time features

@krasul thanks for the great blog post on probabilistic time series. How do I generate out of sample predictions using the vanilla transformer for time series? I was able to run through the entire tutorial, but not sure how to make predictions beyond my validation set.

@krasul Thanks. I am wondering if I can reuse the test loader from the post for this, or if a modification is needed. I’ll check. I can also post on hugging face. Not sure if you’re on discord. Thanks.

I am using hrbrthemes and getting this error when I call theme_ipsum_ps with a ggplot

@hrbrmstr If I've installed this font on my system (macOS) and run import_*, is there a reason why R cannot locate the font? In grid.Call(C_stringMetric, as.graphicsAnnot(x$label)) : no font could be found for family "IBMPlexSans"

Am I the only one “forced” to use pyspark but tired of its verbosity? I wrote some helpers…inspired by you know what #python #rstats #DataScientist

@ChelseaParlett You don’t realize the state you’re in when you’re in it. So I think it’s normal to feel dumb doing difficult stuff, but then when you reflect later in life you’ll feel like “damn I was smart figuring that out.”

How on earth is everyone else doing difficult stuff and simultaneously NOT feeling dumb all the time 😂😅?

I read a quick brief on software versioning. It said versions of software go something like this: 12.6.33 where 12 stands for major releases. This 12 release has had 6 minor releases and 33 bug fixes. Are there any opinions if version ML models should follow something similar?

Rowwise is fun in dplyr! You can avoid loops and map! Although I have an accent to write map, this felt better. #R #rstats #programming

purrr::pluck() lets you deftly index into data structures 🦎 Fun to use, bonus you can say "purrrpluck purrrpluck purrrpluck" the whole time 👍 purrr.tidyverse.org/reference/pluc…

@dvaughan32 @topepos Why does this happen?

This one’s for you @topepos

@allison_horst I can’t find the lineup?

😍 There is an *amazing* lineup of workshops, speakers & panels shaping up for Observable: Insight 2022!! Virtual + free. Register at the link below, & come hang out with this inspiring community of creative data practitioners on Oct 26 & 27! ⬇️🎉

I get asked a lot about what does it take to land a job at DeepMind or any other world-class AI industry lab. For those of you that are unaware of it I wrote a detailed blog on that topic and shared my personal journey here: link.medium.com/dV0H7fay6rb If I could summarize.. 1/ 🧶

ggridges!

I am forgetting the name of the plots where you have numerous density plots lined up by a variable, say a date…does anyone remember the name? #tidyverse #ggplot2

@landsheapes @mtrw85 What’s being mastered here by using base?

@mtrw85 You want results — use the Tidyverse. You want mastery — base is the way.

I've updated my Data Wrangling in the Tidyverse course material and am uploading a 17-part video series. This assumes little previous R knowledge (although some). Covers tidying, manipulating variables, cleaning factors, dates, and strings. Enjoy! youtube.com/watch?v=EOpb3V…

Generalized Additive Models (GAMs) are fully interpretable ML models, unlike DNNs, but it's hard to make them efficient. We're sharing research on scaling GAMs to real world tasks w/o sacrificing accuracy or interpretability. arxiv.org/abs/2205.14120 arxiv.org/abs/2205.14108 [1/4]