mt5-summarize-full

This model is a fine-tuned version of lunarlist/mt5-summarize on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.8640
• Rouge1: 0.3352
• Rouge2: 0.1212
• Rougel: 0.2748
• Rougelsum: 0.4747

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: 0.0005
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 90
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
4.0732	1.0667	100	3.1187	0.3331	0.1146	0.2648	0.5137
3.6546	2.1333	200	2.9872	0.3410	0.1256	0.2894	0.4943
3.3308	3.2	300	2.9373	0.3430	0.1278	0.2881	0.4743
3.276	4.2667	400	2.8782	0.3355	0.1163	0.2793	0.4801
3.1345	5.3333	500	2.9083	0.3354	0.1216	0.2835	0.4758
3.0736	6.4	600	2.8588	0.3531	0.1353	0.2900	0.4991
3.0168	7.4667	700	2.8592	0.3436	0.1229	0.2893	0.4863
2.969	8.5333	800	2.8739	0.3528	0.1297	0.2863	0.4968
2.9677	9.6	900	2.8640	0.3352	0.1212	0.2748	0.4747

Framework versions

• Transformers 4.42.3
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1