mbart-large-50-clara-med

This model is a fine-tuned version of facebook/mbart-large-50 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.2121
• Rouge1: 49.1001
• Rouge2: 31.2516
• Rougel: 44.0446
• Rougelsum: 44.1075

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: 5.6e-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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
No log	1.0	190	1.8633	44.8593	28.0451	40.7724	40.8654
No log	2.0	380	1.6667	46.8654	29.5857	42.6056	42.7844
3.317	3.0	570	1.6847	48.1605	30.163	43.1965	43.3317
3.317	4.0	760	1.7845	48.7615	30.8887	43.6946	43.8016
0.7441	5.0	950	2.0090	48.4207	30.64	43.654	43.7979
0.7441	6.0	1140	2.2425	49.1967	31.2644	44.0566	44.2112
0.7441	7.0	1330	2.4520	47.0568	28.7501	41.8219	41.9605
0.2396	8.0	1520	2.5336	47.969	30.0618	42.9924	43.1481
0.2396	9.0	1710	2.6153	47.2037	28.9732	42.0939	42.2242
0.1112	10.0	1900	2.7299	48.3657	30.3342	43.2025	43.3223
0.1112	11.0	2090	2.7696	48.0929	30.0156	42.9385	43.026
0.1112	12.0	2280	2.8627	48.1979	30.2714	43.0959	43.2027
0.0938	13.0	2470	2.8788	47.7685	29.5733	42.7561	42.9112
0.0938	14.0	2660	2.9128	47.5374	29.8217	42.7097	42.7803
0.0394	15.0	2850	2.9470	48.6385	30.1425	43.3326	43.3963
0.0394	16.0	3040	3.0039	48.6657	30.6642	43.471	43.592
0.0394	17.0	3230	3.0380	48.2351	30.5653	43.257	43.3788
0.023	18.0	3420	3.0289	48.6593	30.6916	43.7861	43.9098
0.023	19.0	3610	3.0733	49.2114	31.2737	44.0852	44.1993
0.0122	20.0	3800	3.1089	48.5431	30.5305	43.4128	43.5288
0.0122	21.0	3990	3.0684	48.4197	30.4005	43.2305	43.3214
0.0122	22.0	4180	3.1252	48.6007	30.5594	43.4008	43.5336
0.0071	23.0	4370	3.1572	48.7297	30.7028	43.436	43.5106
0.0071	24.0	4560	3.1716	48.9335	30.9918	43.7764	43.8044
0.0041	25.0	4750	3.1687	48.8731	31.1055	43.8021	43.8987
0.0041	26.0	4940	3.1845	48.9432	31.0766	43.8628	43.9726
0.0041	27.0	5130	3.2133	49.2016	31.1265	44.052	44.1427
0.0025	28.0	5320	3.2146	49.1473	31.3109	44.0372	44.1189
0.0025	29.0	5510	3.2121	49.2815	31.4258	44.1661	44.2436
0.0019	30.0	5700	3.2121	49.1001	31.2516	44.0446	44.1075

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.0
• Datasets 2.8.0
• Tokenizers 0.12.1