flan-t5-large-extraction-all-dm_2000-ep10-nonstop

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

• Loss: 1.6131
• Hint Hit Num: 2.4754
• Hint Precision: 0.4686
• Num: 5.2796
• Gen Len: 18.5498

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: 12
• eval_batch_size: 96
• seed: 1799
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Hint Hit Num	Hint Precision	Num	Gen Len
2.2839	1.2	200	1.6280	2.639	0.4798	5.535	18.8242
2.0028	2.4	400	1.5794	2.6424	0.478	5.5434	18.8458
1.8677	3.59	600	1.5796	2.543	0.4677	5.444	18.7314
1.8106	4.79	800	1.5944	2.4596	0.4624	5.3132	18.62
1.7414	5.99	1000	1.5957	2.4998	0.4721	5.299	18.5504
1.6858	7.19	1200	1.6058	2.4596	0.4643	5.2926	18.5856
1.6744	8.38	1400	1.6113	2.4936	0.4668	5.333	18.5994
1.6563	9.58	1600	1.6130	2.4752	0.4685	5.2802	18.548

Framework versions

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