res_nw_irq_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: 0.2078
• Bleu: 0.0845
• Rouge1: 0.4018
• Rouge2: 0.1711
• Rougel: 0.3978

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
0.9334	1.0	1057	0.2461	0.0032	0.1593	0.0185	0.1533
0.0868	2.0	2114	0.2332	0.0149	0.2455	0.0510	0.2394
0.0767	3.0	3171	0.2342	0.0252	0.2961	0.0782	0.2910
0.0696	4.0	4228	0.2278	0.0404	0.3300	0.1050	0.3252
0.0636	5.0	5285	0.2219	0.0517	0.3536	0.1215	0.3480
0.0587	6.0	6342	0.2237	0.0590	0.3654	0.1348	0.3611
0.0542	7.0	7399	0.2194	0.0667	0.3755	0.1440	0.3712
0.0502	8.0	8456	0.2080	0.0715	0.3802	0.1521	0.3761
0.0468	9.0	9513	0.2123	0.0770	0.3931	0.1616	0.3889
0.0438	10.0	10570	0.2112	0.0812	0.3921	0.1648	0.3884
0.0408	11.0	11627	0.2102	0.0816	0.3967	0.1653	0.3936
0.0384	12.0	12684	0.2078	0.0845	0.4018	0.1711	0.3978
0.0363	13.0	13741	0.2145	0.0870	0.4023	0.1720	0.3986
0.0343	14.0	14798	0.2165	0.0878	0.4063	0.1757	0.4023
0.0327	15.0	15855	0.2169	0.0920	0.4049	0.1792	0.4014
0.0313	16.0	16912	0.2175	0.0920	0.4078	0.1821	0.4048
0.0301	17.0	17969	0.2191	0.0944	0.4103	0.1839	0.4066

Framework versions

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