10pos_model

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

• Loss: 0.3032
• Bleu: 0.0001
• Wer: 0.9323
• Rougel: 0.1628
• Gen Len: 18.9983

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: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Wer	Rougel	Gen Len
0.8158	0.16	1000	0.3327	0.0001	0.933	0.1616	18.999
0.3791	0.32	2000	0.3029	0.0001	0.9326	0.1623	18.9988
0.3536	0.48	3000	0.2901	0.0001	0.9325	0.1624	18.9988
0.3353	0.64	4000	0.2822	0.0001	0.9324	0.1626	18.9987
0.3333	0.8	5000	0.2774	0.0001	0.9323	0.1627	18.9987
0.3258	0.96	6000	0.2742	0.0001	0.9322	0.1627	18.9987
0.3267	1.12	7000	0.2811	0.0001	0.9323	0.1627	18.9984
0.3532	1.28	8000	0.2960	0.0001	0.9322	0.1629	18.9983
0.3859	1.44	9000	0.3083	0.0001	0.9323	0.1629	18.9983
0.3867	1.6	10000	0.3054	0.0001	0.9323	0.1628	18.9983
0.3807	1.76	11000	0.3038	0.0001	0.9323	0.1628	18.9983
0.3807	1.92	12000	0.3032	0.0001	0.9323	0.1628	18.9983

Framework versions

• Transformers 4.37.1
• Pytorch 2.3.0.dev20240122+cu121
• Datasets 2.16.1
• Tokenizers 0.15.1