flan-t5-base-extraction-cnndm_8000-all-hint_precision-ep50

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

• Loss: 1.7529
• Hint Hit Num: 2.3661
• Hint Precision: 0.432
• Num: 5.4575
• Gen Len: 18.9999

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: 32
• eval_batch_size: 100
• seed: 1799
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Hint Hit Num	Hint Precision	Num	Gen Len
2.3309	0.8	200	1.8686	2.2144	0.415	5.3428	18.9925
2.0891	1.6	400	1.8016	2.2762	0.4225	5.3918	18.999
2.021	2.4	600	1.7832	2.2796	0.4222	5.3946	18.9994
1.9813	3.2	800	1.7706	2.319	0.4271	5.4208	18.9998
1.9465	4.0	1000	1.7521	2.327	0.4267	5.4423	18.9999
1.9139	4.8	1200	1.7491	2.3122	0.4255	5.412	18.9993
1.8886	5.6	1400	1.7506	2.3072	0.426	5.3995	18.9998
1.8753	6.4	1600	1.7529	2.3661	0.432	5.4575	18.9999
1.8432	7.2	1800	1.7428	2.3228	0.4249	5.4502	18.9998
1.8294	8.0	2000	1.7465	2.3463	0.4293	5.4482	18.9996
1.8072	8.8	2200	1.7403	2.3396	0.4271	5.4617	18.9993
1.7819	9.6	2400	1.7429	2.3223	0.4244	5.4575	18.9996
1.7732	10.4	2600	1.7446	2.3258	0.4254	5.4462	18.999

Framework versions

• Transformers 4.18.0
• Pytorch 1.10.0+cu111
• Datasets 2.5.1
• Tokenizers 0.12.1