KETI-AIR-Downstream/long-ke-t5-base-summarization_e10

summarization_all

This model is a fine-tuned version of KETI-AIR/long-ke-t5-base on the jsonl_dataset_sum.py dataset. It achieves the following results on the evaluation set:

• Loss: 1.1442
• Rouge1: 21.9857
• Rouge2: 10.2876
• Rougel: 21.4026
• Rougelsum: 21.4278
• Gen Len: 86.2560

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: 0.001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• total_train_batch_size: 8
• total_eval_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
1.2503	1.0	184670	1.2439	20.2525	9.1467	19.7454	19.771	87.1766
1.1629	2.0	369340	1.1773	21.0068	9.6691	20.4565	20.4888	89.6074
1.1087	3.0	554010	1.1431	21.0216	9.6545	20.489	20.5108	85.5895
1.056	4.0	738680	1.1247	21.6776	10.1424	21.09	21.1168	89.6576
1.0199	5.0	923350	1.1179	21.6563	10.0965	21.0814	21.1056	89.2454
0.9652	6.0	1108020	1.1122	21.6209	10.0725	21.0623	21.0864	86.7079
0.92	7.0	1292690	1.1136	21.9396	10.2734	21.3465	21.3745	86.5547
0.8804	8.0	1477360	1.1228	21.8457	10.1858	21.2552	21.278	87.6413
0.8447	9.0	1662030	1.1327	21.92	10.2635	21.3415	21.3633	86.4453
0.7678	10.0	1846700	1.1442	21.9857	10.2876	21.4026	21.4278	86.2560

Framework versions

• Transformers 4.25.1
• Pytorch 1.12.0
• Datasets 2.8.0
• Tokenizers 0.13.2