sentence-acceptability

This model is a fine-tuned version of bert-base-uncased on the glue dataset. It achieves the following results on the evaluation set:

• Loss: 0.8257
• Accuracy: 0.8217

Model description

This model classifies English sentences according to two different labels: 1 if the sentence is grammatically acceptable and 0 if the sentence is grammatically unacceptable.

Training and evaluation data

The model was trained on the "cola" split of the glue dataset, using the 8551 instances of its "train" split. For the evaluation, the 1043 sentences of the "evaluation" split were used.

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.4868	1.0	1069	0.6279	0.7862
0.3037	2.0	2138	0.6184	0.8140
0.177	3.0	3207	0.8257	0.8217

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu116
• Datasets 2.9.0
• Tokenizers 0.13.2