bart-cnn-pubmed-arxiv-pubmed-v3-e60

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

• Loss: 1.0969
• Rouge1: 60.5054
• Rouge2: 49.8345
• Rougel: 52.7857
• Rougelsum: 59.5625
• Gen Len: 142.0

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: 2e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 60
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
1.2541	1.0	795	0.9312	52.6474	33.219	35.3153	50.2117	142.0
0.7026	2.0	1590	0.8076	53.2385	34.2933	36.3889	50.8338	141.9815
0.5259	3.0	2385	0.7832	53.3407	33.8438	37.2622	50.8487	142.0
0.347	4.0	3180	0.7632	53.24	34.233	36.954	50.3872	142.0
0.2657	5.0	3975	0.7389	54.7175	36.9459	39.4874	52.5236	142.0
0.1799	6.0	4770	0.8152	53.7057	36.4558	39.2037	51.4357	141.8889
0.1425	7.0	5565	0.7632	56.4087	40.1423	44.2536	54.2827	141.8519
0.1161	8.0	6360	0.7787	57.048	41.4384	44.5318	55.001	141.9259
0.0936	9.0	7155	0.8074	55.9781	39.7293	42.2029	53.7465	142.0
0.0863	10.0	7950	0.8527	55.8303	40.3243	43.6105	53.7656	142.0
0.0669	11.0	8745	0.8699	57.0888	42.5994	45.3813	55.4823	142.0
0.0546	12.0	9540	0.8474	55.9644	41.4168	44.3511	54.27	141.8704
0.0473	13.0	10335	0.8369	56.3014	41.3835	44.6644	54.6368	142.0
0.0459	14.0	11130	0.8922	56.9204	42.6545	45.4635	55.3169	142.0
0.0379	15.0	11925	0.9166	57.783	44.3517	48.0052	55.9449	142.0
0.0333	16.0	12720	0.9346	57.7209	44.1832	47.634	56.0137	142.0
0.0304	17.0	13515	0.9046	57.2015	42.7752	46.4241	55.7707	142.0
0.0272	18.0	14310	0.9191	56.0557	41.6832	44.44	54.3098	142.0
0.0242	19.0	15105	0.9431	56.8941	42.662	46.147	55.1771	142.0
0.0208	20.0	15900	0.9127	58.5386	45.2057	48.5554	57.1466	142.0
0.02	21.0	16695	0.9537	57.8511	44.5897	47.8505	56.5768	142.0
0.018	22.0	17490	0.9576	57.5774	44.4534	47.6493	55.9042	142.0
0.0151	23.0	18285	1.0039	57.7678	43.6504	47.3487	55.9951	141.5926
0.0164	24.0	19080	0.9815	57.2684	44.4105	47.8775	55.9622	142.0
0.0131	25.0	19875	0.9932	58.0703	44.5521	47.9763	56.4451	142.0
0.0127	26.0	20670	0.9851	56.9139	43.707	46.8548	55.7885	142.0
0.0113	27.0	21465	0.9894	59.2224	46.5814	49.2356	58.0085	142.0
0.0107	28.0	22260	0.9845	58.6542	46.4524	49.3959	57.4585	142.0
0.0098	29.0	23055	1.0165	57.8297	44.7935	47.7898	56.5338	142.0
0.0093	30.0	23850	0.9844	58.6572	47.6771	50.309	57.4929	142.0
0.0094	31.0	24645	1.0083	57.9771	46.1191	49.7179	56.8376	142.0
0.0077	32.0	25440	0.9739	58.4251	46.2082	49.1364	57.1372	141.463
0.007	33.0	26235	1.0364	58.4724	46.2787	49.7396	57.203	142.0
0.0062	34.0	27030	1.0401	59.9105	48.5584	51.232	58.7889	142.0
0.007	35.0	27825	1.0477	58.3057	46.0506	49.7662	57.1383	142.0
0.0064	36.0	28620	1.0328	58.301	45.3733	48.1001	56.909	142.0
0.0049	37.0	29415	1.0488	58.8353	45.8655	48.7498	57.3955	142.0
0.0037	38.0	30210	1.0196	59.245	47.4285	50.9562	58.1597	142.0
0.0049	39.0	31005	1.0270	59.4799	48.1755	51.5027	58.3599	142.0
0.004	40.0	31800	1.0517	58.8698	46.8679	50.4378	57.7936	142.0
0.0034	41.0	32595	1.0787	59.2729	47.718	50.9233	57.9377	141.8148
0.0031	42.0	33390	1.0685	60.1618	48.1466	51.3451	58.978	142.0
0.0028	43.0	34185	1.0770	60.4238	50.1106	53.211	59.3799	142.0
0.0031	44.0	34980	1.0786	59.1729	47.6285	51.3243	58.0335	142.0
0.0024	45.0	35775	1.0829	59.4366	48.3836	51.7183	58.4366	142.0
0.0021	46.0	36570	1.0791	59.1313	47.6137	51.3465	58.048	142.0
0.002	47.0	37365	1.0630	58.8133	46.795	50.2249	57.496	141.9444
0.0016	48.0	38160	1.0800	58.7699	47.6953	50.1339	57.4936	142.0
0.0018	49.0	38955	1.0563	58.1134	46.3537	49.7251	56.7849	142.0
0.0013	50.0	39750	1.0819	59.3582	47.9255	51.1782	58.2925	142.0
0.0013	51.0	40545	1.0762	59.0797	48.0875	50.8556	57.9182	142.0
0.0013	52.0	41340	1.0906	60.0376	48.9763	51.9324	58.8537	142.0
0.0008	53.0	42135	1.1106	59.3213	48.7152	51.4854	58.2943	142.0
0.0009	54.0	42930	1.0845	59.8334	48.702	51.3005	58.921	142.0
0.0008	55.0	43725	1.1035	60.1754	48.9721	51.4863	59.0829	142.0
0.0008	56.0	44520	1.0872	59.8122	48.6515	51.8589	58.8101	142.0
0.0011	57.0	45315	1.0872	59.5352	48.1967	51.1626	58.3402	142.0
0.0005	58.0	46110	1.0937	59.4125	48.1826	51.5944	58.4618	142.0
0.0008	59.0	46905	1.0936	60.0138	49.1796	52.3896	59.0976	142.0
0.0005	60.0	47700	1.0969	60.5054	49.8345	52.7857	59.5625	142.0

Framework versions

• Transformers 4.19.2
• Pytorch 1.11.0+cu113
• Datasets 2.2.2
• Tokenizers 0.12.1