fine-tuned-bart-10-epochs-1500-input-512-output

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

• Loss: 0.8625
• Rouge1: 0.165
• Rouge2: 0.0393
• Rougel: 0.1287
• Rougelsum: 0.1279
• Gen Len: 33.11

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	151	2.4626	0.0074	0.0	0.0072	0.0071	4.31
No log	2.0	302	1.0378	0.0601	0.0109	0.0554	0.0557	7.64
No log	3.0	453	0.9493	0.0832	0.0204	0.0639	0.0638	21.16
3.1513	4.0	604	0.9068	0.1305	0.0335	0.1007	0.1004	29.73
3.1513	5.0	755	0.8867	0.1605	0.0399	0.1177	0.1176	61.33
3.1513	6.0	906	0.8760	0.1603	0.0306	0.1332	0.133	31.76
0.7578	7.0	1057	0.8677	0.1549	0.0431	0.1168	0.1168	34.18
0.7578	8.0	1208	0.8624	0.161	0.0403	0.1172	0.1168	37.92
0.7578	9.0	1359	0.8639	0.1542	0.0374	0.1219	0.1213	31.73
0.5902	10.0	1510	0.8625	0.165	0.0393	0.1287	0.1279	33.11

Framework versions

• Transformers 4.36.2
• Pytorch 1.12.1+cu113
• Datasets 2.16.1
• Tokenizers 0.15.0