check-amount-deverbalizer-flan-t5-small

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

• Loss: 0.0685
• Parse Rate: 1.0
• Dollar Accuracy: 0.9298
• Cents Accuracy: 0.9991
• Digit Count Accuracy: 1.0
• Perfect Match: 0.9296

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: 128
• eval_batch_size: 128
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Parse Rate	Dollar Accuracy	Cents Accuracy	Digit Count Accuracy	Perfect Match
1.0914	0.2128	200	0.7008	0.9987	0.4598	0.8981	0.36	0.1972
0.6574	0.4255	400	0.3127	0.9999	0.6983	0.9889	0.8008	0.6573
0.4343	0.6383	600	0.1917	1.0	0.8006	0.9978	0.9763	0.7880
0.3524	0.8511	800	0.1484	1.0	0.8458	0.9979	0.9923	0.8407
0.2875	1.0638	1000	0.1249	1.0	0.8716	0.9982	0.9950	0.8682
0.2608	1.2766	1200	0.1089	1.0	0.8845	0.9985	0.9983	0.8827
0.2545	1.4894	1400	0.0974	1.0	0.8915	0.9981	0.9998	0.8909
0.2112	1.7021	1600	0.0895	1.0	0.8998	0.9988	1.0	0.8995
0.2074	1.9149	1800	0.0800	1.0	0.9128	0.9985	1.0	0.9124
0.1984	2.1277	2000	0.0766	1.0	0.9166	0.9987	1.0	0.9165
0.1837	2.3404	2200	0.0725	1.0	0.9246	0.9986	1.0	0.9245
0.1835	2.5532	2400	0.0705	1.0	0.9263	0.9991	1.0	0.9261
0.1743	2.7660	2600	0.0686	1.0	0.9313	0.9991	1.0	0.9311
0.1871	2.9787	2800	0.0685	1.0	0.9298	0.9991	1.0	0.9296

Framework versions

• Transformers 4.48.2
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0