bart-cnn-pubmed-arxiv-pubmed-v3-e64

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: 1.0630
• Rouge1: 58.7
• Rouge2: 47.8042
• Rougel: 50.6967
• Rougelsum: 57.5543
• 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: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 64
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	398	0.9499	53.8396	34.0954	35.6734	51.3453	142.0
1.1219	2.0	796	0.8223	53.0414	33.3193	35.7448	50.1675	142.0
0.6681	3.0	1194	0.7689	53.6684	35.3651	37.7087	51.1441	142.0
0.4393	4.0	1592	0.7694	53.9066	35.3925	38.8917	51.6172	142.0
0.4393	5.0	1990	0.7597	54.0746	36.1026	39.1318	51.9272	142.0
0.2947	6.0	2388	0.8284	53.1168	34.7428	38.0573	50.9563	142.0
0.2016	7.0	2786	0.7951	55.7222	39.0458	42.5265	53.5359	142.0
0.1422	8.0	3184	0.7793	56.2376	40.3348	43.435	54.3228	142.0
0.1096	9.0	3582	0.8260	55.0372	39.0552	42.5403	53.0694	142.0
0.1096	10.0	3980	0.8397	53.849	37.519	40.674	52.1357	141.7037
0.0881	11.0	4378	0.8504	56.4835	41.0484	44.9407	54.3557	142.0
0.0693	12.0	4776	0.8285	55.7705	39.8585	43.722	53.7607	142.0
0.0572	13.0	5174	0.8327	57.932	43.5378	46.8233	55.8739	142.0
0.0461	14.0	5572	0.8720	57.6733	42.9742	45.8698	56.018	142.0
0.0461	15.0	5970	0.8723	57.6072	42.6946	45.2551	55.8486	142.0
0.0416	16.0	6368	0.8764	57.1973	43.1931	46.4492	55.3842	142.0
0.0343	17.0	6766	0.8638	57.4474	43.3544	46.3026	55.7863	142.0
0.03	18.0	7164	0.9234	57.9166	43.8551	46.6473	56.3895	142.0
0.0252	19.0	7562	0.9393	58.2908	45.2321	47.1398	56.6618	142.0
0.0252	20.0	7960	0.8966	59.2798	46.381	49.3514	57.6061	142.0
0.024	21.0	8358	0.9056	57.8409	44.2048	47.3329	56.2568	142.0
0.0195	22.0	8756	0.9424	57.551	44.6847	47.2771	56.2391	142.0
0.0182	23.0	9154	0.9361	59.1078	46.4704	49.4178	57.6796	142.0
0.0169	24.0	9552	0.9456	56.7966	43.3135	46.4208	55.4646	142.0
0.0169	25.0	9950	0.9867	59.5561	47.4638	50.0725	58.2388	141.8519
0.0147	26.0	10348	0.9727	58.2574	44.9904	47.2701	56.4274	142.0
0.0125	27.0	10746	0.9589	58.6792	45.8465	48.0781	57.0755	142.0
0.0117	28.0	11144	0.9635	59.1118	46.6614	50.0552	57.6153	142.0
0.0103	29.0	11542	0.9623	58.2517	45.6401	48.5888	56.7733	142.0
0.0103	30.0	11940	0.9752	59.0707	47.203	49.7992	57.6216	142.0
0.0096	31.0	12338	0.9610	57.6781	44.0504	47.6718	56.1201	142.0
0.0089	32.0	12736	0.9705	58.5592	45.7397	48.681	57.0302	142.0
0.008	33.0	13134	0.9989	58.1997	45.6345	48.2551	56.8571	141.7778
0.0075	34.0	13532	0.9880	57.9632	44.7845	47.8763	56.3979	142.0
0.0075	35.0	13930	1.0041	58.1316	46.2737	49.5986	56.8263	142.0
0.0061	36.0	14328	0.9923	58.4686	46.1735	49.1299	57.0331	142.0
0.0066	37.0	14726	1.0157	58.4277	45.6559	49.1739	56.8198	141.6481
0.0052	38.0	15124	1.0220	58.5166	46.3883	50.0964	57.0104	142.0
0.0049	39.0	15522	0.9949	59.3697	47.0609	50.2733	58.1388	142.0
0.0049	40.0	15920	1.0368	59.9537	48.4059	51.8185	58.8002	142.0
0.0039	41.0	16318	1.0228	58.2093	46.4807	49.54	56.9994	142.0
0.0041	42.0	16716	1.0218	57.6376	45.4951	49.003	56.4606	142.0
0.0035	43.0	17114	1.0381	57.2845	43.9593	46.779	55.6106	142.0
0.0059	44.0	17512	1.0316	58.5506	46.2111	49.4844	56.9506	142.0
0.0059	45.0	17910	1.0388	58.8383	47.6053	50.6187	57.7125	142.0
0.0028	46.0	18308	1.0068	59.3198	47.6888	50.2478	58.0	142.0
0.0028	47.0	18706	1.0446	58.8938	46.7524	49.5642	57.3659	142.0
0.0022	48.0	19104	1.0347	59.8253	48.3871	51.3949	58.5652	142.0
0.0024	49.0	19502	1.0294	60.655	50.2339	53.1662	59.3333	142.0
0.0024	50.0	19900	1.0225	58.5131	47.3009	50.1642	57.2287	142.0
0.0022	51.0	20298	1.0320	59.6101	47.4104	50.5291	58.075	142.0
0.0018	52.0	20696	1.0507	58.7957	46.8893	50.2996	57.3662	142.0
0.0015	53.0	21094	1.0599	58.9064	47.9433	51.3082	57.6871	142.0
0.0015	54.0	21492	1.0636	59.6607	48.5737	51.2361	58.333	142.0
0.0013	55.0	21890	1.0452	58.7026	46.5286	49.9672	57.2521	142.0
0.0012	56.0	22288	1.0418	58.9452	47.7209	50.657	57.7103	142.0
0.0011	57.0	22686	1.0578	58.485	46.0691	49.811	57.2591	142.0
0.0009	58.0	23084	1.0561	59.2268	48.1987	50.1948	57.8871	142.0
0.0009	59.0	23482	1.0548	59.6307	48.1778	50.9934	58.2098	142.0
0.0009	60.0	23880	1.0498	59.5054	48.8866	51.5977	58.1868	142.0
0.0008	61.0	24278	1.0583	60.0232	49.2518	52.2297	58.6774	142.0
0.0007	62.0	24676	1.0659	59.1755	48.4144	51.5157	58.0416	142.0
0.0007	63.0	25074	1.0622	59.1023	47.74	50.5188	57.9707	142.0
0.0007	64.0	25472	1.0630	58.7	47.8042	50.6967	57.5543	142.0

Framework versions

• Transformers 4.18.0
• Pytorch 1.11.0+cu113
• Datasets 2.1.0
• Tokenizers 0.12.1