search

EdoardoPona / predicting-inductive-biases-RL

fork of https://openreview.net/forum?id=mNtmhaDkAr - extending for inductive bias in RL
1 stars 0 forks source link

Implementing and running different LLM tasks #19

Open diogo-cruz opened 1 year ago

diogo-cruz commented 1 year ago

Right now we have three tasks available (labeled 1, 2, and 3 in the toy parameter of sentiment_run.py). Their associated datasets can be generated by running

python imdb_dataset_preprocessing.py --true_property=1 --max_tokens=10
python imdb_dataset_preprocessing.py --true_property=2 --max_tokens=10
python imdb_dataset_preprocessing.py --true_property=3 --max_tokens=20 --train_size=20_000
diogo-cruz commented 1 year ago

Implemented task 4. Run it with

python imdb_dataset_preprocessing.py --true_property=4 --max_tokens=20 --train_size=20_000