bart-large-cnn-finetuned-roundup-32

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: 2.2324
• Rouge1: 46.462
• Rouge2: 25.9506
• Rougel: 29.4584
• Rougelsum: 44.1863
• 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: 32
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	132	1.3139	48.8247	29.2173	31.7628	45.8992	142.0
No log	2.0	264	1.2287	47.9398	29.4061	30.9133	44.9142	140.9
No log	3.0	396	1.2676	49.2743	30.4469	32.8893	46.6208	142.0
0.9578	4.0	528	1.3218	47.315	26.7303	30.5007	44.7654	142.0
0.9578	5.0	660	1.3173	47.1476	25.9408	29.4257	44.4956	142.0
0.9578	6.0	792	1.4283	47.5836	27.1572	29.8553	44.8858	142.0
0.9578	7.0	924	1.5005	46.6839	26.2214	30.1895	43.8753	140.75
0.3306	8.0	1056	1.5316	47.7611	27.1105	30.8142	44.7598	142.0
0.3306	9.0	1188	1.6295	48.4416	27.6912	30.3409	45.317	142.0
0.3306	10.0	1320	1.6564	46.5751	27.2306	29.7265	43.7327	142.0
0.3306	11.0	1452	1.7471	47.9684	27.5739	30.7018	44.6852	141.75
0.145	12.0	1584	1.7700	47.9274	28.5129	31.129	45.1009	142.0
0.145	13.0	1716	1.8391	49.8091	30.1597	33.6004	47.2007	141.95
0.145	14.0	1848	1.9212	45.2195	25.033	27.4181	42.6161	142.0
0.145	15.0	1980	1.9267	48.4959	28.1	31.2796	46.2758	142.0
0.0723	16.0	2112	1.9130	47.0765	27.4929	30.6862	44.1458	142.0
0.0723	17.0	2244	1.9514	48.5354	28.4909	31.8966	45.7116	142.0
0.0723	18.0	2376	2.0064	47.9339	28.6862	32.4472	45.3704	142.0
0.042	19.0	2508	2.0210	48.3169	28.1579	30.2681	45.3831	141.3
0.042	20.0	2640	2.0377	46.8156	26.0122	28.817	43.9383	142.0
0.042	21.0	2772	2.0587	46.3813	27.3555	29.875	43.6605	142.0
0.042	22.0	2904	2.0695	45.6728	26.0639	29.5653	42.3772	142.0
0.025	23.0	3036	2.1617	46.7283	26.2082	28.52	43.3304	142.0
0.025	24.0	3168	2.1375	48.1347	28.3444	31.7509	45.4907	142.0
0.025	25.0	3300	2.1911	47.3358	27.1479	29.4923	44.0087	142.0
0.025	26.0	3432	2.1806	47.2218	26.8421	30.03	44.2417	142.0
0.0153	27.0	3564	2.1890	46.3745	27.0095	29.7274	43.3372	142.0
0.0153	28.0	3696	2.2235	50.1274	30.8817	32.8766	46.7486	141.5
0.0153	29.0	3828	2.2236	50.1785	30.8079	32.8886	46.9888	142.0
0.0153	30.0	3960	2.2312	46.7468	26.4272	30.1175	43.9132	142.0
0.0096	31.0	4092	2.2287	47.558	26.3933	29.9122	44.5752	142.0
0.0096	32.0	4224	2.2324	46.462	25.9506	29.4584	44.1863	142.0

Framework versions

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