faqsum_fine-tune_vietai_vipubmedt5-base_4epochs_v1

This model is a fine-tuned version of htdung167/VietAI_ViPubmedT5_t5x_converted on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0506
• Rouge1: 0.725
• Rouge2: 0.5785
• Rougel: 0.684
• Rougelsum: 0.6849
• Gen Len: 16.2157

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: 0.0002
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
1.1558	0.4006	133	1.1219	0.7022	0.5538	0.6592	0.6598	28.6471
1.136	0.8012	266	1.0128	0.7085	0.5623	0.6703	0.6706	25.417
0.8667	1.2018	399	1.0023	0.724	0.5806	0.6855	0.6859	15.9434
0.7045	1.6024	532	0.9938	0.7213	0.5804	0.6837	0.6843	19.3982
0.8247	2.0030	665	0.9750	0.727	0.5841	0.6878	0.6877	17.0151
0.6886	2.4036	798	0.9992	0.7242	0.5792	0.6826	0.6829	16.7345
0.7363	2.8042	931	0.9992	0.7234	0.5756	0.6812	0.6815	17.7134
0.4797	3.2048	1064	1.0674	0.7295	0.585	0.6907	0.6912	15.6961
0.4926	3.6054	1197	1.0506	0.725	0.5785	0.684	0.6849	16.2157

Framework versions

• Transformers 4.51.3
• Pytorch 2.7.0+cu126
• Datasets 3.6.0
• Tokenizers 0.21.1