VictorSanh/roberta-base-finetuned-yelp-polarity

This is a RoBERTa-base checkpoint fine-tuned on binary sentiment classifcation from Yelp polarity . It gets 98.08% accuracy on the test set . We used the following hyper-parameters to train the model on one GPU:num_train_epochs, weight_decay, gradient_accumulation_steps, seed seed, seed and seed . The model is trained on a single GPU with a 32GB device with 32GB memory . It is now being used to train a model that gets 98% accuracy for accuracy on a test set of Yelp content: [RoberTa-Base-base-finetuned-yelp-,