bart-base-qds1

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

• Loss: 0.2346
• Rouge1: 57.4149
• Rouge2: 41.3439
• Rougel: 52.8013
• Rougelsum: 55.0522
• Gen Len: 18.0

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 12

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
0.249	1.0	1162	0.1781	59.6678	45.1815	55.6055	57.63	18.0
0.1626	2.0	2325	0.1744	59.0831	44.253	54.8112	57.0726	18.0
0.131	3.0	3487	0.1800	58.9286	43.5413	54.5349	56.7525	18.0
0.1071	4.0	4650	0.1855	59.7421	45.2636	55.6941	57.7737	18.0
0.0884	5.0	5812	0.1950	59.0777	43.5656	54.7076	56.861	18.0
0.0732	6.0	6975	0.2066	58.0715	42.4195	53.6034	55.7227	18.0
0.0611	7.0	8137	0.2109	58.2439	42.2049	53.6602	55.6642	18.0
0.0516	8.0	9300	0.2182	58.3266	42.403	53.8468	55.8914	18.0
0.0439	9.0	10462	0.2236	57.9343	41.6112	53.1966	55.2496	18.0
0.0382	10.0	11625	0.2272	57.5539	41.1889	52.9571	55.2649	18.0
0.0337	11.0	12787	0.2320	57.8289	41.8217	53.1707	55.3936	18.0
0.0304	11.99	13944	0.2346	57.4149	41.3439	52.8013	55.0522	18.0

Framework versions

• Transformers 4.36.1
• Pytorch 2.1.2
• Datasets 2.19.2
• Tokenizers 0.15.2