lvwerra/gpt2-imdb-pos

A small GPT2 language model fine-tuned to produce positive movie reviews based the IMDB dataset . The full experiment setup can be found in the Jupyter notebook in the trl repo . The model is trained with rewards from a BERT sentiment classifier (lvwerra/gpt2-imdb) via PPO. The model was trained for 100 optimisation steps with a batch size of 256 which corresponds to 25600 training samples . It was trained with a set of 100 optimisations steps with the training sample of 256 training samples of the model . It is now being used to analyse movie reviews on IMDB and film reviews on the JUPyter repo,