NyanNyanovich

Here you go: https://github.com/NyanNyanovich/nyan/releases/download/can_annot/cat_markup.tar.gz I used Lenta and gpt-4 annotations, here is the script to query gpt-4: https://github.com/NyanNyanovich/nyan/blob/master/scripts/annotate_categories.py And the training script: https://github.com/NyanNyanovich/nyan/blob/master/scripts/train_clf.py

In my case /data/mongo is 17GB for all the time the system is running. The container is indeed 35GB, which is unexpected, but it doesn't grow.

It is not a part of core functions. I have this command in `cron`, with three different launch times: ```python3 -m nyan.topics --mongo-config-path configs/mongo_config.json --client-config-path configs/client_config.json --renderer-config-path configs/renderer_config.json --prompt-path nyan/prompts/topics.txt...

Nothing. Archiving is manual using this [script](https://github.com/NyanNyanovich/nyan/blob/master/archive.sh). Nope, it was about [ratings.py](https://github.com/NyanNyanovich/nyan/blob/master/nyan/ratings.py), but having both topics and ratings seems too much for me.

Rarely. You can find all runs here: https://github.com/NyanNyanovich/nyan/releases

Нет таких моделей, по крайней мере сравнимых с gpt-4. Я сохраняю все выжимки, когда-нибудь на них можно будет дообучить модель. Для preview версии это выходит примерно в 50 сентов в...