randomization_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: 1.2766
• Bleu: 0.0001
• Wer: 0.9507
• Rougel: 0.1324
• Gen Len: 18.9988

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
1.6112	0.16	1000	1.2949	0.0001	0.9506	0.1323	18.9988
1.5684	0.32	2000	1.2655	0.0001	0.9503	0.1328	18.9988
1.5378	0.48	3000	1.2461	0.0001	0.95	0.1334	18.9985
1.5183	0.64	4000	1.2246	0.0001	0.95	0.1334	18.9985
1.5065	0.8	5000	1.2108	0.0001	0.9499	0.1336	18.9985
1.4787	0.96	6000	1.1975	0.0001	0.9498	0.1338	18.9986
1.5081	1.12	7000	1.2173	0.0001	0.9498	0.134	18.9986
1.6302	1.28	8000	1.2801	0.0001	0.9507	0.1325	18.9988
1.6337	1.44	9000	1.2766	0.0001	0.9507	0.1324	18.9988
1.6333	1.6	10000	1.2766	0.0001	0.9507	0.1324	18.9988
1.6246	1.76	11000	1.2766	0.0001	0.9507	0.1324	18.9988
1.6219	1.92	12000	1.2766	0.0001	0.9507	0.1324	18.9988

Framework versions

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