HAR_model

This model is a fine-tuned version of t5-small on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.5524
• Rouge1: 0.3529
• Rouge2: 0.1071
• Rougel: 0.2263
• Rougelsum: 0.2263
• Gen Len: 86.0

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	1	2.9579	0.312	0.0738	0.2003	0.2003	75.0
No log	2.0	2	2.8855	0.312	0.0738	0.2003	0.2003	75.0
No log	3.0	3	2.8381	0.3376	0.0808	0.205	0.205	77.5
No log	4.0	4	2.7929	0.3383	0.0903	0.2018	0.2018	74.5
No log	5.0	5	2.7389	0.3383	0.0903	0.2018	0.2018	74.5
No log	6.0	6	2.6640	0.3383	0.0903	0.2018	0.2018	74.5
No log	7.0	7	2.6333	0.3422	0.0916	0.1961	0.1961	72.0
No log	8.0	8	2.6110	0.3383	0.0903	0.2018	0.2018	74.5
No log	9.0	9	2.5951	0.3529	0.1071	0.2263	0.2263	86.0
No log	10.0	10	2.5826	0.3529	0.1071	0.2263	0.2263	86.0
No log	11.0	11	2.5732	0.3529	0.1071	0.2263	0.2263	86.0
No log	12.0	12	2.5632	0.3529	0.1071	0.2263	0.2263	86.0
No log	13.0	13	2.5632	0.3529	0.1071	0.2263	0.2263	86.0
No log	14.0	14	2.5562	0.3529	0.1071	0.2263	0.2263	86.0
No log	15.0	15	2.5524	0.3529	0.1071	0.2263	0.2263	86.0

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.1