asril-pegasus-xlsum-skripsi

This model is a fine-tuned version of google/pegasus-xsum on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.6919

Model description

this model is spesifically for indonsesian abstractive news article summarization wich has been fine tuning in more than 48k dataset this model fine-tuned using pegasus model

Intended uses & limitations

More information needed

Training and evaluation data

xlsum/indonesian

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 1
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
4.5256	0.1046	1000	3.4857
3.699	0.2092	2000	3.1625
3.4046	0.3138	3000	2.9968
3.2456	0.4184	4000	2.8834
3.126	0.5230	5000	2.8127
3.055	0.6275	6000	2.7644
3.005	0.7321	7000	2.7281
2.9597	0.8367	8000	2.7060
2.9627	0.9413	9000	2.6919

Framework versions

• Transformers 4.40.0
• Pytorch 2.1.0+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1