sum_model_lr1e_3_20_30epoch_use2e-5from20epoch

This model is a fine-tuned version of weny22/sum_model_lr1e_3_20epoch on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.9485
• Rouge1: 0.219
• Rouge2: 0.0909
• Rougel: 0.1806
• Rougelsum: 0.1803
• Gen Len: 18.9827

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: 64
• eval_batch_size: 64
• 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
No log	1.0	335	1.8977	0.2197	0.0921	0.1821	0.182	18.9807
1.3728	2.0	670	1.9029	0.2188	0.091	0.1806	0.1805	18.9793
1.3461	3.0	1005	1.9100	0.221	0.0926	0.1827	0.1825	18.9833
1.3461	4.0	1340	1.9138	0.2197	0.0917	0.1818	0.1816	18.982
1.3229	5.0	1675	1.9248	0.2192	0.0908	0.1807	0.1804	18.9813
1.3041	6.0	2010	1.9285	0.2191	0.0912	0.1807	0.1805	18.9793
1.3041	7.0	2345	1.9345	0.2192	0.0914	0.1811	0.1808	18.982
1.2879	8.0	2680	1.9411	0.2182	0.0907	0.1803	0.18	18.982
1.2748	9.0	3015	1.9458	0.2193	0.0913	0.1809	0.1805	18.982
1.2748	10.0	3350	1.9485	0.219	0.0909	0.1806	0.1803	18.9827

Framework versions

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