summarizer_samsum_model

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

• Loss: 0.3992
• Rouge1: 0.4144
• Rouge2: 0.1805
• Rougel: 0.3419
• Rougelsum: 0.3418
• Gen Len: 16.6732

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: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
0.4595	1.0	737	0.4170	0.3923	0.163	0.3243	0.3242	16.1826
0.4474	2.0	1474	0.4113	0.3991	0.1685	0.3304	0.3303	16.5925
0.4416	3.0	2211	0.4092	0.4021	0.1722	0.3337	0.3339	16.6023
0.4388	4.0	2948	0.4048	0.4062	0.1737	0.3361	0.3361	16.5731
0.4331	5.0	3685	0.4030	0.4093	0.1758	0.3379	0.338	16.696
0.4243	6.0	4422	0.4010	0.4111	0.1778	0.3396	0.3396	16.5728
0.4234	7.0	5159	0.4000	0.4129	0.1789	0.3406	0.3405	16.7139
0.425	8.0	5896	0.3996	0.4125	0.1797	0.3407	0.3407	16.7089
0.4247	9.0	6633	0.3993	0.4147	0.181	0.3421	0.3422	16.6943
0.4176	10.0	7370	0.3992	0.4144	0.1805	0.3419	0.3418	16.6732

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2