DanSumT5-largeV_84227V_29665

This model is a fine-tuned version of emilstabil/DanSumT5-largeV_84227 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.4261
• Rouge1: 32.3558
• Rouge2: 8.5749
• Rougel: 19.1338
• Rougelsum: 29.7211
• Gen Len: 122.67

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-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 4
• 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	Gen Len
No log	1.0	200	2.3019	31.5752	8.1099	18.3519	29.1739	124.3
No log	2.0	400	2.3070	31.9164	8.7613	19.0514	29.6226	124.78
1.7988	3.0	600	2.3043	32.1339	8.5015	18.9179	29.2816	124.63
1.7988	4.0	800	2.3136	31.6595	8.1675	18.7604	29.2322	124.76
1.6472	5.0	1000	2.3283	32.2728	8.3023	18.9738	29.8167	124.88
1.6472	6.0	1200	2.3351	32.0678	8.5944	18.7781	29.595	123.61
1.6472	7.0	1400	2.3524	32.3119	8.2027	19.0197	29.8383	123.66
1.5261	8.0	1600	2.3595	32.3179	8.2699	18.869	29.7615	124.09
1.5261	9.0	1800	2.3669	31.6759	8.2793	18.7806	29.3134	123.97
1.4329	10.0	2000	2.3767	32.3248	8.6537	19.0477	29.9719	124.87
1.4329	11.0	2200	2.3960	31.6656	8.2021	18.932	29.5141	125.2
1.4329	12.0	2400	2.4107	31.8985	8.0749	18.8192	29.4881	124.02
1.3681	13.0	2600	2.4135	31.5421	8.233	18.9548	29.2292	125.34
1.3681	14.0	2800	2.4136	32.1472	8.0262	18.8115	29.7394	123.28
1.3366	15.0	3000	2.4125	31.8903	8.4487	18.9192	29.3627	124.31
1.3366	16.0	3200	2.4199	31.9141	8.1944	18.8633	29.6124	123.63
1.3366	17.0	3400	2.4205	32.0654	8.333	18.6209	29.5897	124.23
1.3135	18.0	3600	2.4248	32.1559	8.4127	18.9745	29.8147	122.88
1.3135	19.0	3800	2.4244	32.5409	8.7406	19.0785	29.9493	123.22
1.3006	20.0	4000	2.4261	32.3558	8.5749	19.1338	29.7211	122.67

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.0
• Datasets 2.12.0
• Tokenizers 0.13.3