bart-large-cnn-finetuned-roundup-4-8

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: 0.7882
• Rouge1: 54.2292
• Rouge2: 37.3874
• Rougel: 40.3261
• Rougelsum: 52.2155
• Gen Len: 141.8889

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: 8
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	398	0.9442	53.1634	33.7744	35.5688	50.7523	142.0
1.1285	2.0	796	0.8305	54.0713	35.7079	37.5147	51.6285	142.0
0.6796	3.0	1194	0.7735	52.6656	34.0198	36.8075	50.1502	142.0
0.4572	4.0	1592	0.7759	53.6269	35.4308	38.3735	51.1369	141.7222
0.4572	5.0	1990	0.7527	54.4206	36.0907	38.0818	51.7885	142.0
0.3171	6.0	2388	0.7755	54.9642	38.0459	41.6383	52.8847	142.0
0.2269	7.0	2786	0.7801	54.1637	35.9853	39.5262	51.6562	142.0
0.1686	8.0	3184	0.7882	54.2292	37.3874	40.3261	52.2155	141.8889

Framework versions

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