bart-large-finetuned-bart

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

• Loss: 0.8644
• Rouge1: 80.883
• Rouge2: 72.0268
• Rougel: 77.0146
• Rougelsum: 77.3408
• Gen Len: 19.2969

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	12	0.8158	70.5188	55.0313	61.668	62.0697	18.8281
No log	2.0	24	0.6610	78.9718	67.9238	71.2508	71.5416	19.2188
No log	3.0	36	0.6257	79.2603	69.4298	73.031	73.4136	19.1875
No log	4.0	48	0.6354	79.0425	69.7772	73.6893	73.8356	19.3125
No log	5.0	60	0.6126	79.9003	69.6148	73.3622	73.7154	19.2656
No log	6.0	72	0.6537	80.0209	69.9125	73.6876	73.9664	19.1875
No log	7.0	84	0.7168	80.7559	71.8251	75.5574	75.6627	19.2031
No log	8.0	96	0.6980	80.9116	72.2951	75.9015	76.2205	19.2656
No log	9.0	108	0.7569	80.1034	70.8352	75.1723	75.4102	19.25
No log	10.0	120	0.7523	80.0436	71.0855	75.8337	76.1002	19.3281
No log	11.0	132	0.7742	80.5982	71.713	75.9081	76.1564	19.4375
No log	12.0	144	0.7570	79.4243	70.5775	75.3521	75.9334	19.4844
No log	13.0	156	0.8225	80.2529	72.6598	76.6931	76.8326	19.7344
No log	14.0	168	0.8696	79.821	71.2369	75.836	76.0734	19.4688
No log	15.0	180	0.8820	80.9234	72.5022	76.7848	77.0368	19.375
No log	16.0	192	0.8400	80.5926	72.0275	76.8664	77.0773	19.2969
No log	17.0	204	0.8461	80.6004	72.2808	76.8693	77.0517	19.375
No log	18.0	216	0.8577	81.0069	73.1954	77.467	77.7083	19.3906
No log	19.0	228	0.8695	81.0333	72.5324	77.4248	77.5971	19.2969
No log	20.0	240	0.8644	80.883	72.0268	77.0146	77.3408	19.2969

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.0+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3