CSE4022_NLP_EPJ_model

This model is a fine-tuned version of t5-small on the billsum dataset. It achieves the following results on the evaluation set:

• Loss: 2.1741
• Rouge1: 0.2276
• Rouge2: 0.1147
• Rougel: 0.1939
• Rougelsum: 0.1939
• Gen Len: 19.0

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	62	2.6085	0.1452	0.0463	0.1198	0.1193	19.0
No log	2.0	124	2.4702	0.1683	0.059	0.1373	0.1372	19.0
No log	3.0	186	2.3930	0.1982	0.0824	0.1636	0.1635	19.0
No log	4.0	248	2.3491	0.2221	0.1036	0.1856	0.1856	19.0
No log	5.0	310	2.3202	0.2281	0.1081	0.1903	0.1905	19.0
No log	6.0	372	2.2934	0.2255	0.1072	0.1891	0.1892	19.0
No log	7.0	434	2.2722	0.2272	0.111	0.1913	0.1913	19.0
No log	8.0	496	2.2516	0.2284	0.1119	0.1922	0.1924	19.0
2.598	9.0	558	2.2386	0.2301	0.1133	0.194	0.1942	19.0
2.598	10.0	620	2.2276	0.229	0.1134	0.1936	0.1938	19.0
2.598	11.0	682	2.2165	0.2278	0.1145	0.1939	0.194	19.0
2.598	12.0	744	2.2056	0.2281	0.1147	0.194	0.1942	19.0
2.598	13.0	806	2.1976	0.2268	0.1143	0.1931	0.1933	19.0
2.598	14.0	868	2.1923	0.2269	0.1142	0.1939	0.194	19.0
2.598	15.0	930	2.1873	0.2289	0.1154	0.1952	0.1951	19.0
2.598	16.0	992	2.1805	0.2287	0.1153	0.1953	0.1953	19.0
2.3476	17.0	1054	2.1790	0.2286	0.1153	0.1951	0.1951	19.0
2.3476	18.0	1116	2.1768	0.2277	0.1146	0.1942	0.1942	19.0
2.3476	19.0	1178	2.1746	0.2278	0.1149	0.1942	0.1942	19.0
2.3476	20.0	1240	2.1741	0.2276	0.1147	0.1939	0.1939	19.0

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0