AraT5v2-base-1024-finetuned-ar-wikilingua

This model is a fine-tuned version of UBC-NLP/AraT5v2-base-1024 on the wiki_lingua dataset. It achieves the following results on the evaluation set:

• Loss: 4.1591
• Rouge-1: 26.54
• Rouge-2: 10.4
• Rouge-l: 23.72
• Gen Len: 18.19
• Bertscore: 72.52

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: 5e-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
• lr_scheduler_warmup_steps: 250
• num_epochs: 8
• label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge-1	Rouge-2	Rouge-l	Gen Len	Bertscore
5.2884	1.0	4998	4.4307	23.0	8.16	20.56	17.66	70.77
4.6798	2.0	9996	4.2972	24.48	8.95	21.86	17.57	71.56
4.4355	3.0	14994	4.2313	24.85	9.17	22.23	17.68	71.7
4.2772	4.0	19992	4.1972	25.41	9.5	22.65	17.63	72.08
4.1551	5.0	24990	4.1724	25.43	9.44	22.58	17.68	72.08
4.0604	6.0	29988	4.1626	25.44	9.56	22.67	17.52	72.19
3.989	7.0	34986	4.1616	25.71	9.68	22.91	17.71	72.29
3.9467	8.0	39984	4.1591	25.81	9.81	23.03	17.67	72.33

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.13.3