t5-small-finetuned-FairytaleQA-AnswerExtraction

This model is a fine-tuned version of t5-small on the fairytale_qa dataset. It achieves the following results on the evaluation set:

• Loss: 0.0695
• Rouge1: 10.7124
• Rouge2: 3.2292
• Rougel: 10.375
• Rougelsum: 10.3824
• F1: 0.2626
• Exact Match: 0.4878
• Gen Len: 11.9668

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: 2e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	F1	Exact Match	Gen Len
0.0777	1.0	2137	0.0727	11.076	3.0967	10.6128	10.6396	0.1301	0.3902	12.522
0.0727	2.0	4274	0.0707	11.288	3.2828	10.9125	10.9225	0.152	0.4878	12.161
0.0696	3.0	6411	0.0699	10.7512	3.3182	10.406	10.4123	0.2626	0.4878	12.1122
0.0719	4.0	8548	0.0696	10.803	3.2133	10.4337	10.4223	0.2626	0.4878	11.9698
0.07	5.0	10685	0.0695	10.7124	3.2292	10.375	10.3824	0.2626	0.4878	11.9668

Framework versions

• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1