LongT5-Large-NSPCC

This model is a fine-tuned version of google/long-t5-tglobal-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.5481
• Rouge1: 0.4597
• Rouge2: 0.1665
• Rougel: 0.2562
• Rougelsum: 0.2557
• Gen Len: 250.6383

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: 0.0003
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.03
• num_epochs: 6

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
6.0521	1.0	188	2.8154	0.2268	0.0411	0.1627	0.1626	145.7447
2.5796	2.0	377	1.9961	0.3798	0.1115	0.2103	0.2101	220.234
2.0398	3.0	566	1.7703	0.4208	0.1319	0.2255	0.2258	299.6915
1.7329	4.0	755	1.5996	0.4427	0.1488	0.2423	0.2424	255.2553
1.5609	5.0	943	1.5510	0.4688	0.1726	0.2578	0.2576	289.2979
1.4733	5.98	1128	1.5481	0.4597	0.1665	0.2562	0.2557	250.6383

Framework versions

• Transformers 4.39.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2