Tag_recommendaiton_model_kh_en

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

• Loss: 0.0422
• Tag Set Precision: 0.9495
• Tag Set Recall: 0.9431
• Tag Set F1: 0.9455
• Exact Set Match: 0.7633

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: 0.0003
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• 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
• lr_scheduler_warmup_steps: 0.06
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Exact Set Match	Validation Loss	Tag Set F1	Tag Set Precision	Tag Set Recall
2.9061	0.1058	2000	0.0523	0.4014	0.5352	0.5993	0.4971
0.7883	0.2115	4000	0.5647	0.1045	0.8685	0.9008	0.8528
0.3851	0.3173	6000	0.6913	0.0612	0.9136	0.9365	0.9065
0.2798	0.4230	8000	0.725	0.0520	0.9242	0.9508	0.9102
0.2432	0.5288	10000	0.7327	0.0506	0.9279	0.9401	0.9255
0.2287	0.6346	12000	0.7427	0.0488	0.9330	0.9483	0.9262
0.1885	0.7403	14000	0.7473	0.0474	0.9383	0.9502	0.9325
0.2049	0.8461	16000	0.755	0.0455	0.9405	0.9478	0.9377
0.1912	0.9518	18000	0.7493	0.0450	0.9380	0.9438	0.9374
0.1813	1.0576	20000	0.758	0.0442	0.9426	0.9467	0.9423
0.1743	1.1633	22000	0.765	0.0437	0.9446	0.9475	0.9450
0.1767	1.2691	24000	0.0429	0.9489	0.9423	0.9440	0.762
0.1716	1.3749	26000	0.0424	0.9519	0.9424	0.9459	0.7603
0.1700	1.4806	28000	0.0426	0.9462	0.9457	0.9447	0.7647
0.1678	1.5864	30000	0.0422	0.9495	0.9431	0.9455	0.7633

Framework versions

• Transformers 5.14.1
• Pytorch 2.10.0+cu128
• Datasets 5.0.1
• Tokenizers 0.22.2