emilstabil/DanSumT5-baseV_91333

DanSumT5-base-finetuned-test_6887

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

• Loss: 2.5277
• Rouge1: 31.3188
• Rouge2: 7.8236
• Rougel: 17.8296
• Rougelsum: 28.6162
• Gen Len: 127.0

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: 3
• eval_batch_size: 3
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 12
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	0.99	66	2.7401	29.9221	6.3861	16.5877	27.0611	125.09
No log	1.99	133	2.6815	30.6874	6.9413	16.9609	27.8851	125.98
No log	3.0	200	2.6440	31.2045	7.4012	17.7421	28.3497	126.63
No log	4.0	267	2.6199	31.3329	7.4574	17.8549	28.643	126.98
No log	4.99	333	2.5984	31.5184	7.7763	17.9153	29.0627	127.0
No log	5.99	400	2.5822	31.8839	7.9755	18.0572	29.2282	126.65
No log	7.0	467	2.5677	31.5939	7.9515	17.865	29.2019	126.45
2.8684	8.0	534	2.5587	31.4931	7.6042	17.6853	28.8366	126.79
2.8684	8.99	600	2.5496	31.105	7.6714	17.5128	28.5242	126.78
2.8684	9.99	667	2.5423	31.6087	8.0358	17.9956	28.9514	126.78
2.8684	11.0	734	2.5364	31.411	7.9534	17.895	28.7595	127.0
2.8684	12.0	801	2.5326	31.4648	7.9777	17.9589	28.8168	127.0
2.8684	12.99	867	2.5296	31.374	7.8341	17.8341	28.8146	127.0
2.8684	13.99	934	2.5278	31.2822	7.7789	17.7983	28.5903	127.0
2.8684	14.83	990	2.5277	31.3188	7.8236	17.8296	28.6162	127.0

Framework versions

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