t5_ner_task

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

• Loss: 0.0064

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: 0.0001
• train_batch_size: 48
• eval_batch_size: 48
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 192
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.0326	1.0	280	0.0247
0.0209	2.0	560	0.0170
0.0166	3.0	840	0.0141
0.0138	4.0	1120	0.0121
0.0123	5.0	1400	0.0110
0.0114	6.0	1680	0.0100
0.0103	7.0	1960	0.0090
0.0091	8.0	2240	0.0085
0.0084	9.0	2520	0.0084
0.0086	10.0	2800	0.0078
0.0082	11.0	3080	0.0077
0.0075	12.0	3360	0.0079
0.007	13.0	3640	0.0077
0.0067	14.0	3920	0.0075
0.0061	15.0	4200	0.0072
0.0064	16.0	4480	0.0069
0.006	17.0	4760	0.0071
0.0059	18.0	5040	0.0070
0.0055	19.0	5320	0.0067
0.0053	20.0	5600	0.0069
0.0053	21.0	5880	0.0067
0.0051	22.0	6160	0.0067
0.0052	23.0	6440	0.0067
0.0049	24.0	6720	0.0065
0.0049	25.0	7000	0.0065
0.0045	26.0	7280	0.0066
0.0048	27.0	7560	0.0064
0.0043	28.0	7840	0.0064
0.0046	29.0	8120	0.0064
0.0042	30.0	8400	0.0064

Framework versions

• Transformers 4.57.1
• Pytorch 2.11.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2