T5_base_NASA_ADS_title_v5

This model is a fine-tuned version of t5-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.7910
• Rouge1: 0.4285
• Rouge2: 0.2147
• Rougel: 0.3712
• Rougelsum: 0.3711
• Gen Len: 16.3966

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.1209	1.0	2535	1.9129	0.4133	0.2034	0.3581	0.3581	16.0037
1.9984	2.0	5070	1.8568	0.4138	0.2051	0.3608	0.3608	15.9065
1.9004	3.0	7605	1.8307	0.4222	0.2109	0.3675	0.3674	15.9252
1.8491	4.0	10140	1.8110	0.4199	0.209	0.3638	0.3638	16.1813
1.7823	5.0	12675	1.7972	0.4231	0.2107	0.3665	0.3664	16.2793
1.7302	6.0	15210	1.7924	0.427	0.2145	0.3704	0.3705	16.3819
1.6996	7.0	17745	1.7918	0.4284	0.2157	0.3712	0.3713	16.0767
1.6705	8.0	20280	1.7851	0.4252	0.2134	0.3682	0.3682	16.3485
1.6338	9.0	22815	1.7853	0.4291	0.2146	0.3721	0.372	16.398
1.6193	10.0	25350	1.7850	0.4297	0.2158	0.3726	0.3725	16.3821
1.5808	11.0	27885	1.7837	0.4274	0.2144	0.3708	0.3706	16.303
1.574	12.0	30420	1.7895	0.428	0.2151	0.3708	0.3709	16.3568
1.5544	13.0	32955	1.7898	0.4282	0.2152	0.3711	0.3711	16.412
1.5377	14.0	35490	1.7902	0.4287	0.2147	0.3709	0.3709	16.3917
1.5304	15.0	38025	1.7910	0.4285	0.2147	0.3712	0.3711	16.3966

Framework versions

• Transformers 4.40.0
• Pytorch 2.1.2
• Datasets 2.1.0
• Tokenizers 0.19.1