pegasus-cnn_dailymail-finetuned-scope-summarization

This model is a fine-tuned version of google/pegasus-cnn_dailymail on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1923
• Rouge1: 56.9116
• Rouge2: 45.4236
• Rougel: 49.8645
• Rougelsum: 49.71

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: 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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
0.5864	1.0	158	0.3056	41.1501	21.8098	34.2505	34.2057
0.3517	2.0	316	0.2787	45.9535	26.9084	37.7287	37.7213
0.2835	3.0	474	0.2655	49.4653	30.6584	40.5201	40.4494
0.2683	4.0	632	0.2528	50.2066	32.8862	40.4058	40.2244
0.2557	5.0	790	0.2469	50.3451	33.536	41.7433	41.6118
0.2493	6.0	948	0.2382	51.9053	36.1533	42.0343	41.8884
0.2406	7.0	1106	0.2330	53.2105	38.1	43.434	43.2194
0.235	8.0	1264	0.2267	51.9642	38.1903	44.4502	44.3851
0.2296	9.0	1422	0.2237	53.5609	38.9875	44.7145	44.6146
0.2246	10.0	1580	0.2195	54.6691	41.5464	45.7506	45.6856
0.221	11.0	1738	0.2141	54.4114	41.2748	45.9992	45.8182
0.2145	12.0	1896	0.2097	55.3852	42.9342	48.376	48.5267
0.2115	13.0	2054	0.2060	55.9251	43.4806	48.0303	47.9584
0.2081	14.0	2212	0.2017	55.8426	43.1239	47.8006	47.8356
0.2042	15.0	2370	0.1997	55.4631	42.78	47.307	47.3142
0.2031	16.0	2528	0.1970	57.0004	44.4252	49.6236	49.5213
0.1996	17.0	2686	0.1953	55.438	43.8797	48.536	48.3506
0.1991	18.0	2844	0.1939	56.1102	44.5176	48.5553	48.4163
0.1963	19.0	3002	0.1925	56.6366	45.3753	49.4421	49.3468
0.1955	20.0	3160	0.1923	56.9116	45.4236	49.8645	49.71

Framework versions

• Transformers 4.40.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1