ModernBERT-base_en-tr_jobs_sentence-transformer_isco08-clf_part-05

This model is a fine-tuned version of ai4jobs/ModernBERT-base_en-tr_jobs_sentence-transformer on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.3499
• F1 Weighted: 0.7700
• F1 Macro: 0.5328
• Accuracy: 0.7734

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Use adamw_bnb_8bit 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: 0.1
• num_epochs: 6

Training results

Training Loss	Epoch	Step	Validation Loss	F1 Weighted	F1 Macro	Accuracy
6.4761	0.1290	1000	1.4466	0.5991	0.2762	0.6474
2.5152	0.2580	2000	1.2032	0.6414	0.3803	0.669
2.2911	0.3870	3000	1.1419	0.6668	0.4173	0.689
2.2391	0.5160	4000	1.0739	0.6692	0.4298	0.6926
2.06	0.6450	5000	1.0828	0.6734	0.4386	0.695
1.931	0.7740	6000	0.9855	0.6906	0.4503	0.708
1.8233	0.9030	7000	0.9522	0.7108	0.4622	0.7248
1.6038	1.0320	8000	0.9133	0.7238	0.5015	0.7398
1.3756	1.1610	9000	0.9152	0.7165	0.4857	0.7316
1.3543	1.2900	10000	0.8855	0.7318	0.4949	0.7426
1.3413	1.4190	11000	0.8926	0.7333	0.4953	0.7434
1.2833	1.5480	12000	0.8498	0.7408	0.5274	0.7516
1.2525	1.6770	13000	0.8164	0.7480	0.5159	0.7594
1.207	1.8060	14000	0.8246	0.7483	0.5309	0.7542
1.16	1.9350	15000	0.8083	0.7557	0.5330	0.7636
0.9448	2.0640	16000	0.8621	0.7440	0.5263	0.7528
0.7289	2.1930	17000	0.8537	0.7561	0.5367	0.759
0.7117	2.3220	18000	0.8620	0.7553	0.5312	0.7634
0.7167	2.4510	19000	0.8982	0.7529	0.5286	0.7584
0.7129	2.5800	20000	0.8877	0.7555	0.5352	0.7634
0.6961	2.7090	21000	0.8876	0.7538	0.5426	0.7594
0.689	2.8380	22000	0.8947	0.7579	0.5431	0.7636
0.6488	2.9670	23000	0.8763	0.7614	0.5458	0.7694
0.4092	3.0960	24000	0.9449	0.7638	0.5466	0.7698
0.3237	3.2250	25000	0.9822	0.7623	0.5275	0.7686
0.3155	3.3540	26000	1.0121	0.7534	0.5248	0.759
0.3185	3.4830	27000	1.0267	0.7562	0.5342	0.7602
0.3125	3.6120	28000	1.0373	0.7575	0.5282	0.7634
0.3151	3.7410	29000	1.0390	0.7623	0.5309	0.766
0.2953	3.8700	30000	1.0490	0.7621	0.5453	0.7652
0.2851	3.9990	31000	1.0769	0.7587	0.5301	0.763
0.1055	4.1280	32000	1.1409	0.7670	0.5468	0.771
0.1024	4.2570	33000	1.1721	0.7619	0.5361	0.7668
0.1014	4.3860	34000	1.1987	0.7632	0.5392	0.7652
0.1079	4.5150	35000	1.1890	0.7707	0.5411	0.7726
0.1115	4.6440	36000	1.2438	0.7670	0.5311	0.7724
0.0988	4.7730	37000	1.2474	0.7668	0.5438	0.7702
0.1038	4.9020	38000	1.2408	0.7679	0.5401	0.7716
0.0751	5.0310	39000	1.2840	0.7643	0.5368	0.769
0.0261	5.1600	40000	1.3044	0.7658	0.5244	0.7704
0.0348	5.2890	41000	1.3229	0.7685	0.5309	0.7728
0.0351	5.4180	42000	1.3458	0.7695	0.5332	0.7726
0.0362	5.5470	43000	1.3494	0.7678	0.5301	0.7718
0.0353	5.6760	44000	1.3500	0.7682	0.5352	0.7714
0.0368	5.8050	45000	1.3499	0.7700	0.5328	0.7734

Framework versions

• Transformers 4.48.3
• Pytorch 2.6.0+cu124
• Datasets 3.3.0
• Tokenizers 0.21.0