bart-large-cnn-finetuned-roundup-16

This model is a fine-tuned version of facebook/bart-large-cnn on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.8957
• Rouge1: 49.4097
• Rouge2: 29.3516
• Rougel: 31.527
• Rougelsum: 46.4241
• Gen Len: 141.9

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: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 16
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	132	1.3170	48.412	29.2017	31.6679	45.494	141.85
No log	2.0	264	1.2292	49.0133	29.6645	30.7612	46.1673	142.0
No log	3.0	396	1.2670	49.183	29.4104	31.573	46.7082	142.0
0.9596	4.0	528	1.3059	47.3854	26.6865	28.4666	44.4934	141.8
0.9596	5.0	660	1.3288	48.1189	26.9242	31.2938	45.3462	142.0
0.9596	6.0	792	1.4084	47.5713	26.7488	29.2959	45.1764	141.3
0.9596	7.0	924	1.5043	46.5407	26.0995	29.9007	43.9335	142.0
0.3369	8.0	1056	1.5115	49.6891	29.0514	32.33	46.9357	142.0
0.3369	9.0	1188	1.6131	47.5773	27.6348	30.5294	45.1151	142.0
0.3369	10.0	1320	1.6837	46.5699	26.3805	29.8581	43.5252	142.0
0.3369	11.0	1452	1.7874	47.1383	26.535	30.1724	44.2508	142.0
0.148	12.0	1584	1.7776	49.8061	30.1994	33.2405	47.6102	142.0
0.148	13.0	1716	1.8144	48.4451	28.2949	30.9026	45.6614	142.0
0.148	14.0	1848	1.8646	50.1964	30.4426	32.8156	47.4134	142.0
0.148	15.0	1980	1.8829	48.8129	29.2358	32.3247	46.2233	142.0
0.0726	16.0	2112	1.8957	49.4097	29.3516	31.527	46.4241	141.9

Framework versions

• Transformers 4.18.0
• Pytorch 1.11.0+cu113
• Datasets 2.1.0
• Tokenizers 0.12.1