flan-t5-xl-summarization-epoch20

This model is a fine-tuned version of google/flan-t5-xl on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5008
• Rouge1: 48.3084
• Rouge2: 27.2658
• Rougel: 37.9769
• Rougelsum: 41.5848
• Gen Len: 52.1176

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• 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	Gen Len
No log	1.0	40	0.9558	33.7961	16.1287	27.1659	28.0049	27.5294
No log	2.0	80	0.7329	41.1727	26.7202	35.6927	37.9856	62.6471
No log	3.0	120	0.5996	39.7001	21.6984	29.3765	34.738	82.0588
No log	4.0	160	0.5612	41.4021	23.9875	32.7841	36.4756	67.5294
No log	5.0	200	0.5494	42.9379	24.0227	33.2609	37.7189	67.5882
No log	6.0	240	0.5344	44.3145	24.7379	34.5022	38.7382	58.1176
No log	7.0	280	0.5264	48.3821	28.1406	36.8146	40.9602	54.1765
No log	8.0	320	0.5193	48.5669	28.7554	37.2762	41.4076	55.8235
No log	9.0	360	0.5129	48.4222	25.9534	35.4387	40.3668	57.7059
No log	10.0	400	0.5109	48.1639	27.399	37.7239	40.9771	51.0588
No log	11.0	440	0.5093	50.4094	29.8618	39.3303	42.7215	53.0
No log	12.0	480	0.5060	50.3864	27.8568	37.5365	42.3323	53.3529
0.8091	13.0	520	0.5073	48.0328	26.5537	36.7542	41.2961	55.1765
0.8091	14.0	560	0.5049	47.2298	26.6774	36.8165	40.5404	52.5294
0.8091	15.0	600	0.5008	48.3084	27.2658	37.9769	41.5848	52.1176
0.8091	16.0	640	0.5017	47.6969	27.0742	37.3415	41.0155	54.9412
0.8091	17.0	680	0.5022	48.3553	27.5197	38.2598	41.5044	54.0588
0.8091	18.0	720	0.5018	48.474	27.5343	37.7907	41.5528	56.0
0.8091	19.0	760	0.5010	48.474	27.5343	37.7907	41.5528	56.0
0.8091	20.0	800	0.5009	48.474	27.5343	37.7907	41.5528	56.0

Framework versions

• PEFT 0.12.0
• Transformers 4.44.2
• Pytorch 2.2.1+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1