Picking up on this. Two things I realised:
1. I should develop this with a focus on localisation. I will use it for different languages than English only.
2. There are many different standards to specify a language. Even within nltk, modules are not coherent with language codes.
Working on a side project to improve how I learn new words by reading books...
Something I found out on the way: my code says there are ~5300 different words in Harry Potter vol.1, excluding names and numbers!
Also learned about lemmatisation. All pretty interesting so far:)
Working on a side project to improve how I learn new words by reading books...
Something I found out on the way: my code says there are ~5300 different words in Harry Potter vol.1, excluding names and numbers!
Also learned about lemmatisation. All pretty interesting so far:)
Thinking about char-level events. Would certainly be more of an extravagant than useful behaviour but I can imagine a few interesting creative use-cases.
How I imagine an ideal summer retreat:
3 months of
- Surfing for sunrise and sunset for that endorphin boost
- Good food in-between to stay happy
- A bit of piano and reading during the day to stay intellectually stimulated
No computer.
Currently off to a Hackathon in the Swiss alpsโฆ still unsure whether I actually like this type of event, but taking the train to get up the mountains already made the trip worth it!
Just finished my submission for the @ArcInternet hackathon; I essentially built an integration of #chatGPT for @Wikipedia.
While I think this could be an incredibly useful addition, I'm worried about LLMs and factual inaccuracy.
Similarly, when I memorise a piano piece, I prompt my brain to recall a specific section of the piece. Often, I can derive that section if I first think of the bits before and after the section.