flan-t5-large-da-multiwoz_1000

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: 0.3538
• Accuracy: 41.3747
• Num: 3689
• Gen Len: 15.5115

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: 8
• eval_batch_size: 24
• seed: 1799
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Num	Gen Len
1.3315	0.24	200	0.5697	25.9543	3689	14.556
0.6418	0.48	400	0.4645	30.0503	3689	14.9314
0.5433	0.72	600	0.4307	31.9506	3689	16.1515
0.4909	0.95	800	0.4177	34.7593	3689	15.418
0.4769	1.19	1000	0.3996	35.0943	3689	14.9607
0.4491	1.43	1200	0.3881	36.2741	3689	15.543
0.4531	1.67	1400	0.3820	35.7704	3689	14.1583
0.4322	1.91	1600	0.3726	37.4853	3689	15.961
0.4188	2.15	1800	0.3699	38.4117	3689	15.0773
0.4085	2.38	2000	0.3674	38.5353	3689	15.4012
0.4063	2.62	2200	0.3606	40.0046	3689	15.3546
0.3977	2.86	2400	0.3570	40.6543	3689	15.704
0.3992	3.1	2600	0.3549	40.4284	3689	15.7446
0.3828	3.34	2800	0.3538	41.3747	3689	15.5115
0.3792	3.58	3000	0.3539	39.8513	3689	14.7951
0.3914	3.81	3200	0.3498	41.0388	3689	15.4153
0.3707	4.05	3400	0.3498	40.9596	3689	16.3136

Framework versions

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