rut5-base-absum-finetuned-summ

This model is a fine-tuned version of cointegrated/rut5-base-absum on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6039
• Rouge1: 97.0122
• Rouge2: 94.5148
• Rougel: 97.0189
• Rougelsum: 96.9668

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: 5.6e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
No log	1.0	15	0.9642	88.9988	70.6048	88.8684	88.9803
No log	2.0	30	0.7765	94.6938	86.9198	94.7219	94.6778
No log	3.0	45	0.6995	96.1002	90.9986	96.058	96.058
No log	4.0	60	0.6596	96.0421	92.2644	96.067	96.0107
No log	5.0	75	0.6294	96.5868	93.2489	96.5836	96.5625
No log	6.0	90	0.6172	96.4605	92.827	96.4538	96.4071
No log	7.0	105	0.6091	97.0122	94.5148	97.0189	96.9668
1.0079	8.0	120	0.6039	97.0122	94.5148	97.0189	96.9668

Framework versions

• Transformers 4.53.0
• Pytorch 2.2.1+cu118
• Datasets 4.0.0
• Tokenizers 0.21.4