mt5-large-commas-8b-20e-500k_unfiltered

This model is a fine-tuned version of google/mt5-large on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0219

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: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
0.0128	1.0	56250	0.0087
0.0097	2.0	112500	0.0081
0.007	3.0	168750	0.0085
0.0068	4.0	225000	0.0084
0.006	5.0	281250	0.0086
0.0046	6.0	337500	0.0096
0.0038	7.0	393750	0.0105
0.0038	8.0	450000	0.0108
0.0029	9.0	506250	0.0122
0.0026	10.0	562500	0.0131
0.0022	11.0	618750	0.0138
0.0015	12.0	675000	0.0151
0.0017	13.0	731250	0.0159
0.0012	14.0	787500	0.0175
0.001	15.0	843750	0.0179
0.0008	16.0	900000	0.0185
0.0006	17.0	956250	0.0210
0.0006	18.0	1012500	0.0214
0.0006	19.0	1068750	0.0216
0.0005	20.0	1125000	0.0219

Framework versions

• Transformers 4.48.0
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0