EUCO-agenda-classification

This model is a fine-tuned version of microsoft/mdeberta-v3-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.4480
• Accuracy: 0.7172
• F1 Macro: 0.5761
• Accuracy Balanced: 0.5714
• F1 Micro: 0.7172
• Precision Macro: 0.6144
• Recall Macro: 0.5714
• Precision Micro: 0.7172
• Recall Micro: 0.7172

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: 40
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.06
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	Accuracy Balanced	F1 Micro	Precision Macro	Recall Macro	Precision Micro	Recall Micro
No log	1.0	110	1.7007	0.3978	0.1235	0.1680	0.3978	0.0982	0.1680	0.3978	0.3978
No log	2.0	220	1.3681	0.5399	0.3020	0.3937	0.5399	0.3679	0.3937	0.5399	0.5399
No log	3.0	330	1.3401	0.6455	0.3707	0.3848	0.6455	0.4403	0.3848	0.6455	0.6455
No log	4.0	440	1.7684	0.6522	0.4350	0.4284	0.6522	0.4721	0.4284	0.6522	0.6522
2.9324	5.0	550	2.0030	0.6671	0.4491	0.4520	0.6671	0.4921	0.4520	0.6671	0.6671
2.9324	6.0	660	2.4103	0.6793	0.4226	0.4109	0.6793	0.4720	0.4109	0.6793	0.6793
2.9324	7.0	770	2.4339	0.6644	0.4307	0.4408	0.6644	0.4362	0.4408	0.6644	0.6644
2.9324	8.0	880	2.3396	0.6834	0.4262	0.4331	0.6834	0.4277	0.4331	0.6834	0.6834
2.9324	9.0	990	2.5157	0.6725	0.4312	0.4292	0.6725	0.4600	0.4292	0.6725	0.6725
0.0468	10.0	1100	2.5344	0.6793	0.4361	0.4297	0.6793	0.4731	0.4297	0.6793	0.6793
0.0468	11.0	1210	2.5214	0.6874	0.4903	0.4660	0.6874	0.5972	0.4660	0.6874	0.6874
0.0468	12.0	1320	2.6415	0.6698	0.4990	0.4670	0.6698	0.5933	0.4670	0.6698	0.6698
0.0468	13.0	1430	2.5394	0.6942	0.5624	0.5533	0.6942	0.6070	0.5533	0.6942	0.6942
0.0127	14.0	1540	2.5218	0.6969	0.5681	0.5641	0.6969	0.6075	0.5641	0.6969	0.6969
0.0127	15.0	1650	2.5401	0.6928	0.5612	0.5523	0.6928	0.6052	0.5523	0.6928	0.6928
0.0127	16.0	1760	2.5292	0.6942	0.5677	0.5639	0.6942	0.6079	0.5639	0.6942	0.6942
0.0127	17.0	1870	2.4163	0.7131	0.5759	0.5731	0.7131	0.6127	0.5731	0.7131	0.7131
0.0127	18.0	1980	2.4645	0.7158	0.5731	0.5652	0.7158	0.6174	0.5652	0.7158	0.7158
0.003	19.0	2090	2.4534	0.7199	0.5776	0.5718	0.7199	0.6170	0.5718	0.7199	0.7199
0.003	19.8238	2180	2.4480	0.7172	0.5761	0.5714	0.7172	0.6144	0.5714	0.7172	0.7172

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu121
• Datasets 2.14.7
• Tokenizers 0.21.0