sum_model_lr1e_3_20_30epoch

From epoch20 I use lr = 1e-4 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: 1.9732
• Rouge1: 0.2208
• Rouge2: 0.0917
• Rougel: 0.1817
• Rougelsum: 0.1817
• Gen Len: 18.986

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: 0.0001
• 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	1.8899	0.222	0.0926	0.1826	0.1826	18.982
1.3783	2.0	670	1.9074	0.2183	0.0902	0.1801	0.1801	18.984
1.3451	3.0	1005	1.9165	0.2214	0.0924	0.1821	0.182	18.9807
1.3451	4.0	1340	1.9231	0.2203	0.0923	0.1816	0.1816	18.982
1.3128	5.0	1675	1.9327	0.2208	0.0924	0.1815	0.1818	18.98
1.2892	6.0	2010	1.9408	0.2205	0.0922	0.1816	0.1816	18.982
1.2892	7.0	2345	1.9524	0.2201	0.0918	0.1813	0.1814	18.9787
1.2675	8.0	2680	1.9617	0.2205	0.0924	0.1816	0.1815	18.986
1.252	9.0	3015	1.9690	0.2204	0.092	0.1816	0.1816	18.984
1.252	10.0	3350	1.9732	0.2208	0.0917	0.1817	0.1817	18.986

Framework versions

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