t5-small-govReport-boardpapers-3072

This model is a fine-tuned version of RMWeerasinghe/t5-small-finetuned-govReport-3072 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.6701
• Rouge1: 0.0443
• Rouge2: 0.0194
• Rougel: 0.0382
• Rougelsum: 0.0443

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: 4e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
No log	0.67	1	3.9496	0.0584	0.0214	0.0482	0.0572
No log	2.0	3	3.9252	0.0562	0.0223	0.0463	0.0562
No log	2.67	4	3.9121	0.0597	0.0223	0.0485	0.0596
No log	4.0	6	3.8880	0.0597	0.0223	0.0485	0.0596
No log	4.67	7	3.8755	0.0597	0.0223	0.0485	0.0596
No log	6.0	9	3.8506	0.0597	0.0223	0.0485	0.0596
No log	6.67	10	3.8395	0.0553	0.0197	0.0441	0.0541
No log	8.0	12	3.8172	0.0582	0.0262	0.049	0.057
No log	8.67	13	3.8065	0.0582	0.0262	0.049	0.057
No log	10.0	15	3.7862	0.0582	0.0257	0.049	0.057
No log	10.67	16	3.7769	0.057	0.0262	0.049	0.0556
No log	12.0	18	3.7599	0.0577	0.0294	0.0495	0.0575
No log	12.67	19	3.7522	0.0487	0.0174	0.042	0.0474
4.3528	14.0	21	3.7378	0.048	0.0155	0.0406	0.0461
4.3528	14.67	22	3.7310	0.0536	0.0206	0.0421	0.0511
4.3528	16.0	24	3.7187	0.048	0.017	0.0394	0.0448
4.3528	16.67	25	3.7132	0.043	0.017	0.0374	0.041
4.3528	18.0	27	3.7031	0.043	0.017	0.0374	0.041
4.3528	18.67	28	3.6985	0.043	0.017	0.0374	0.041
4.3528	20.0	30	3.6905	0.043	0.017	0.0374	0.041
4.3528	20.67	31	3.6869	0.043	0.017	0.0374	0.041
4.3528	22.0	33	3.6807	0.0442	0.0194	0.0381	0.0423
4.3528	22.67	34	3.6781	0.0442	0.0194	0.0381	0.0423
4.3528	24.0	36	3.6740	0.0442	0.0194	0.0381	0.0423
4.3528	24.67	37	3.6725	0.0442	0.0194	0.0381	0.0423
4.3528	26.0	39	3.6705	0.0443	0.0194	0.0382	0.0443
4.0602	26.67	40	3.6701	0.0443	0.0194	0.0382	0.0443

Framework versions

• Transformers 4.37.0
• Pytorch 2.1.2
• Datasets 2.17.0
• Tokenizers 0.15.1