Word-selector

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

• Loss: 2.0659
• Rouge1: 0.3077
• Rouge2: 0.0523
• Rougel: 0.2373
• Rougelsum: 0.2375
• Gen Len: 50.2144

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: 0.0002
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.7905	1.0	800	2.2300	0.1838	0.0234	0.1522	0.1525	58.5994
2.3488	2.0	1600	2.1395	0.2251	0.0339	0.1816	0.182	61.4869
2.2336	3.0	2400	2.1009	0.2553	0.0406	0.2052	0.2053	56.9838
2.1005	4.0	3200	2.0684	0.2777	0.0452	0.2161	0.2163	54.5738
2.0007	5.0	4000	2.0559	0.2907	0.0463	0.2247	0.2248	54.0806
1.9248	6.0	4800	2.0623	0.2981	0.0475	0.2306	0.2308	50.3856
1.8686	7.0	5600	2.0513	0.3013	0.0508	0.2326	0.2327	53.965
1.831	8.0	6400	2.0531	0.3083	0.0517	0.2355	0.2357	51.4162
1.7752	9.0	7200	2.0623	0.3022	0.0514	0.2326	0.2329	51.6419
1.7515	10.0	8000	2.0659	0.3077	0.0523	0.2373	0.2375	50.2144

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.1+cu121
• Datasets 3.0.1
• Tokenizers 0.15.1