t5-small_multinews_model

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

• Loss: 2.6269
• Rouge Rouge1: 0.1471
• Rouge Rouge2: 0.0483
• Rouge Rougel: 0.1131
• Rouge Rougelsum: 0.1131
• Bleu Bleu: 0.0003
• Bleu Precisions: [0.5848502090652357, 0.18492208339182928, 0.08486295668446923, 0.04842115016777968]
• Bleu Brevity Penalty: 0.0022
• Bleu Length Ratio: 0.1408
• Bleu Translation Length: 191567
• Bleu Reference Length: 1360656

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge Rouge1	Rouge Rouge2	Rouge Rougel	Rouge Rougelsum	Bleu Bleu	Bleu Precisions	Bleu Brevity Penalty	Bleu Length Ratio	Bleu Translation Length	Bleu Reference Length
2.9189	1.0	7870	2.6869	0.1448	0.0474	0.1117	0.1117	0.0003	[0.5827522821123012, 0.1820493433028088, 0.08242051182628926, 0.04574874477953644]	0.0023	0.1411	192037	1360656
2.8435	2.0	15740	2.6535	0.1460	0.0474	0.1122	0.1122	0.0003	[0.5809636959568958, 0.18126278620071182, 0.08254004826406995, 0.04636911719064694]	0.0023	0.1410	191907	1360656
2.7922	3.0	23610	2.6389	0.1461	0.0477	0.1124	0.1124	0.0003	[0.581669805398619, 0.18257649098318213, 0.08343485040444401, 0.0471782007379682]	0.0022	0.1405	191160	1360656
2.814	4.0	31480	2.6280	0.1468	0.0478	0.1129	0.1129	0.0003	[0.5844809737428239, 0.18360803285143726, 0.08381524001996615, 0.04753093788548009]	0.0022	0.1406	191262	1360656
2.7869	5.0	39350	2.6269	0.1471	0.0483	0.1131	0.1131	0.0003	[0.5848502090652357, 0.18492208339182928, 0.08486295668446923, 0.04842115016777968]	0.0022	0.1408	191567	1360656

Framework versions

• Transformers 4.32.1
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3