Working on the site - and have a better idea about what I can optimize:
- Finished the bulletin item form - which took ridiculously long time for some reason (or I have been dithering on it)
- Started inputting bulletin items - and I saw what Petra meant by some automating options
- e.g. importing from EEB events
Currently trying out google/flan-t5-small but it's kind of crappy? Will try the previous version, google-t5/t5-small
Playing around with models available for two Hugging Face tasks:
- Question-Answering - apparently this is no longer available in Hugging Face interface > v5
- Summarization
Some models were too large to be loaded into memory - I upgraded my Linode instance to include more memory and I think the result is better?