res_nw_dj_aragpt2-base

This model is a fine-tuned version of aubmindlab/aragpt2-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0752
• Bleu: 0.0928
• Rouge1: 0.4343
• Rouge2: 0.2043
• Rougel: 0.4304

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: 8
• 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: 20.0

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Rouge1	Rouge2	Rougel
0.435	1.0	2679	0.0899	0.0225	0.2697	0.0681	0.2649
0.0908	2.0	5358	0.0822	0.0495	0.3438	0.1233	0.3394
0.0808	3.0	8037	0.0786	0.0670	0.3835	0.1582	0.3790
0.0738	4.0	10716	0.0765	0.0782	0.4066	0.1798	0.4025
0.0681	5.0	13395	0.0756	0.0880	0.4242	0.1964	0.4204
0.0632	6.0	16074	0.0752	0.0928	0.4343	0.2043	0.4304
0.059	7.0	18753	0.0755	0.0996	0.4439	0.2152	0.4401
0.0552	8.0	21432	0.0761	0.1015	0.4500	0.2217	0.4463
0.0517	9.0	24111	0.0766	0.1050	0.4527	0.2250	0.4489
0.0486	10.0	26790	0.0784	0.1093	0.4612	0.2338	0.4578
0.0458	11.0	29469	0.0798	0.1112	0.4634	0.2356	0.4600

Framework versions

• Transformers 4.45.0.dev0
• Pytorch 2.3.1+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1