@jacobmenick I also wanted to accomplish the same bad idea :). Came up with a different approach using the Jupyter's Tornado server (cell 11):
https://t.co/2r7UGULpDb
Errr ok wow, I am shook by the new ConvMixer architecture
https://t.co/crUMktQ0ig "the first model that achieves the elusive dual goals of 80%+ ImageNet top-1 accuracy while also fitting into a tweet" 😐
OH SNAP!!!!! #Python just hit #1 on TIOBE, displacing C and Java!
Congrats @gvanrossum and the many, many contributors who made this happen! @ThePSF@PyData
Tag a Pythonista in this thread!
https://t.co/8EPH7WzLdi
Hope everyone is enjoying their week-long stint as a data scientist, the most glamorous tech job of the last 10 years, supposed to be spent analyzing sophisticated models, but really spent mostly monitoring those few stray batch import jobs that haven't finished yet.
Unpopular view: Data scientists should be more end-to-end.
While this is frowned upon (too generalist!), I've seen it lead to more context, faster iteration, greater innovation—more value, faster.
More details and Stitch Fix & Netflix's experience 👇
https://t.co/aOBjuBSsSz
I've been reflecting a lot on why data science is so hard to do well. My current mental model: value is generated multiplicatively.
You must get each thing right or it's worth almost nothing:
- Formulate good questions
- Quality data
- Correct analysis
- Communicate findings
sktime - @scikit_learn like library for time series #machinelearning. Supports forecasting, time series classification, and time series regression.
https://t.co/b0hNsIIvoL
The trolley is barreling down the track towards five grandmothers. You can pull the switch and redirect it so that it kills no one. But you have to wear a mask to buy a toaster oven. What do you do?