t5-text2sql_v1

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

• Loss: 0.0772
• Rouge2 Precision: 0.8835
• Rouge2 Recall: 0.39
• Rouge2 Fmeasure: 0.5088

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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge2 Precision	Rouge2 Recall	Rouge2 Fmeasure
No log	1.0	11	1.9420	0.0755	0.022	0.0323
No log	2.0	22	1.2731	0.0912	0.0263	0.039
No log	3.0	33	0.8717	0.0993	0.0284	0.0424
No log	4.0	44	0.5705	0.1014	0.032	0.0464
No log	5.0	55	0.3929	0.4151	0.1528	0.2149
No log	6.0	66	0.2911	0.7778	0.351	0.4594
No log	7.0	77	0.2290	0.781	0.3305	0.4395
No log	8.0	88	0.1995	0.7381	0.2992	0.4018
No log	9.0	99	0.1768	0.752	0.3147	0.4202
No log	10.0	110	0.1554	0.7242	0.3136	0.412
No log	11.0	121	0.1446	0.8128	0.3583	0.4694
No log	12.0	132	0.1337	0.8194	0.3653	0.478
No log	13.0	143	0.1264	0.8088	0.3564	0.4675
No log	14.0	154	0.1170	0.8036	0.3502	0.462
No log	15.0	165	0.1078	0.8851	0.3981	0.5188
No log	16.0	176	0.1046	0.8716	0.3864	0.5054
No log	17.0	187	0.1007	0.8753	0.3851	0.5042
No log	18.0	198	0.0951	0.8756	0.3941	0.5126
No log	19.0	209	0.0928	0.8414	0.3565	0.4708
No log	20.0	220	0.0894	0.854	0.3642	0.4808
No log	21.0	231	0.0863	0.8954	0.3954	0.5168
No log	22.0	242	0.0832	0.888	0.3931	0.5122
No log	23.0	253	0.0828	0.8835	0.39	0.5088
No log	24.0	264	0.0820	0.8835	0.39	0.5088
No log	25.0	275	0.0803	0.8835	0.39	0.5088
No log	26.0	286	0.0792	0.8835	0.39	0.5088
No log	27.0	297	0.0784	0.8761	0.3886	0.5066
No log	28.0	308	0.0775	0.8835	0.39	0.5088
No log	29.0	319	0.0772	0.8835	0.39	0.5088
No log	30.0	330	0.0772	0.8835	0.39	0.5088

Framework versions

• Transformers 4.24.0
• Pytorch 1.12.1+cu113
• Datasets 2.7.1
• Tokenizers 0.13.2