bart-large-cnn-prompt_generation

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

• Loss: 2.6454
• Rouge1: 40.6908
• Rouge2: 16.1706
• Rougel: 25.6927
• Rougelsum: 25.6588
• Gen Len: 77.2

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: 3e-07
• 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: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	15	3.6562	25.0903	5.3158	16.4265	16.3853	67.42
No log	2.0	30	3.5539	24.9011	4.9854	16.5812	16.5697	65.28
No log	3.0	45	3.3930	24.9983	5.2373	17.0342	16.993	65.8
No log	4.0	60	3.2928	24.8418	4.7159	16.929	16.907	66.0
No log	5.0	75	3.1723	26.012	5.5696	17.4002	17.4621	66.84
No log	6.0	90	3.0813	26.9443	5.8262	17.8297	17.8673	67.52
No log	7.0	105	3.0169	27.7155	6.4297	18.4479	18.4913	66.78
No log	8.0	120	2.9700	27.2858	6.5437	18.5185	18.4731	67.78
No log	9.0	135	2.9340	28.0747	7.3049	18.7045	18.718	67.34
No log	10.0	150	2.9044	28.4417	7.34	18.7805	18.8377	66.44
No log	11.0	165	2.8795	28.8704	7.4119	18.7748	18.849	67.02
No log	12.0	180	2.8558	28.5338	7.1929	18.7993	18.859	67.02
No log	13.0	195	2.8351	30.3984	8.3546	19.8864	19.918	68.18
No log	14.0	210	2.8170	30.934	8.8637	20.6051	20.6574	67.74
No log	15.0	225	2.8016	33.611	10.3334	22.0692	22.11	67.94
No log	16.0	240	2.7867	34.4518	11.2186	22.5517	22.5979	67.36
No log	17.0	255	2.7737	33.8745	10.9904	22.0985	22.1333	68.98
No log	18.0	270	2.7617	35.1795	11.6458	22.3628	22.3954	68.1
No log	19.0	285	2.7502	35.3137	11.7688	22.7397	22.7986	67.24
No log	20.0	300	2.7402	35.8673	12.3602	23.4671	23.481	67.32
No log	21.0	315	2.7312	37.2112	13.6711	24.0348	24.0426	68.58
No log	22.0	330	2.7228	37.521	14.1801	24.1826	24.2038	68.46
No log	23.0	345	2.7148	37.4877	13.7803	24.2369	24.189	70.18
No log	24.0	360	2.7074	38.2158	14.3195	24.4562	24.4262	69.56
No log	25.0	375	2.7012	38.0379	14.2362	24.5273	24.4723	70.7
No log	26.0	390	2.6955	37.4245	13.8152	24.4203	24.4188	69.52
No log	27.0	405	2.6905	37.4296	13.4741	24.569	24.5797	70.7
No log	28.0	420	2.6859	38.7617	14.3506	25.0565	25.0256	71.56
No log	29.0	435	2.6815	39.3441	15.2271	25.4611	25.4251	73.38
No log	30.0	450	2.6774	38.6753	14.4202	24.7802	24.8057	72.94
No log	31.0	465	2.6732	39.7278	15.0554	25.4741	25.4578	74.02
No log	32.0	480	2.6697	39.9498	15.0412	25.4949	25.5039	74.2
No log	33.0	495	2.6668	40.0256	15.1986	25.4401	25.436	75.14
2.6871	34.0	510	2.6638	39.8616	15.249	25.4639	25.4979	75.54
2.6871	35.0	525	2.6613	39.9678	15.1735	25.7189	25.719	75.8
2.6871	36.0	540	2.6593	40.3261	15.4175	25.6158	25.6426	75.0
2.6871	37.0	555	2.6572	40.6307	15.3666	25.6045	25.6245	76.06
2.6871	38.0	570	2.6551	41.2257	15.55	26.0762	26.0547	75.74
2.6871	39.0	585	2.6535	41.2164	15.981	26.068	26.0566	76.16
2.6871	40.0	600	2.6520	41.3161	15.9648	26.0276	26.0199	76.14
2.6871	41.0	615	2.6508	41.1103	15.7775	25.2761	25.237	77.28
2.6871	42.0	630	2.6496	41.4765	16.2494	26.021	26.0026	76.68
2.6871	43.0	645	2.6488	41.725	16.3547	26.1039	26.067	75.88
2.6871	44.0	660	2.6478	41.3649	16.3576	26.0133	25.9943	76.08
2.6871	45.0	675	2.6472	41.1901	16.4955	26.0594	26.0468	76.34
2.6871	46.0	690	2.6466	41.0942	16.2436	25.8578	25.853	75.92
2.6871	47.0	705	2.6461	40.6232	16.1631	25.6709	25.6473	76.46
2.6871	48.0	720	2.6458	41.1453	16.3914	25.946	25.9199	76.1
2.6871	49.0	735	2.6455	41.0364	16.3432	25.8202	25.7964	76.18
2.6871	50.0	750	2.6454	40.6908	16.1706	25.6927	25.6588	77.2

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0