flan-t5-small-samsum

This model is a fine-tuned version of google/flan-t5-small on the samsum dataset. It achieves the following results on the evaluation set:

• Loss: 1.6732
• Rouge1: 42.4655
• Rouge2: 18.4875
• Rougel: 35.2198
• Rougelsum: 38.6465
• Gen Len: 16.8486

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
1.8853	0.22	100	1.7046	42.3969	18.365	35.0091	38.6527	16.6703
1.8463	0.43	200	1.6954	42.5607	18.4425	35.1088	38.8749	17.3565
1.8549	0.65	300	1.6794	42.5148	18.4716	35.1769	38.7018	17.1123
1.8361	0.87	400	1.6775	42.3899	18.4343	35.134	38.5732	16.6215
1.8132	1.08	500	1.6732	42.4655	18.4875	35.2198	38.6465	16.8486
1.8073	1.3	600	1.6708	42.4741	18.3824	35.1819	38.6066	16.9475
1.7973	1.52	700	1.6686	42.8206	18.7011	35.3874	38.9173	16.7595
1.798	1.74	800	1.6666	42.7779	18.6627	35.323	38.9467	16.9389
1.79	1.95	900	1.6668	42.8071	18.7113	35.2872	38.8641	16.9426

Framework versions

• Transformers 4.36.0
• Pytorch 2.0.0
• Datasets 2.15.0
• Tokenizers 0.15.0