t5-small-finetuned-contradiction

This model is a fine-tuned version of domenicrosati/t5-small-finetuned-contradiction on the snli dataset. It achieves the following results on the evaluation set:

• Loss: 2.0458
• Rouge1: 34.4237
• Rouge2: 14.5442
• Rougel: 32.5483
• Rougelsum: 32.5785

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5.6e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 8
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
1.8605	1.0	2863	2.0813	34.4597	14.5186	32.6909	32.7097
1.9209	2.0	5726	2.0721	34.3859	14.5733	32.5188	32.5524
1.9367	3.0	8589	2.0623	34.4192	14.455	32.581	32.5962
1.9539	4.0	11452	2.0565	34.5148	14.6131	32.6786	32.7174
1.9655	5.0	14315	2.0538	34.4393	14.6439	32.6344	32.6587
1.9683	6.0	17178	2.0493	34.7199	14.7763	32.8625	32.8782
1.9735	7.0	20041	2.0476	34.5366	14.6362	32.6939	32.7177
1.98	8.0	22904	2.0458	34.5	14.5695	32.6219	32.6478

Framework versions

• Transformers 4.18.0
• Pytorch 1.11.0+cu102
• Datasets 2.1.0
• Tokenizers 0.12.1