exceptions_exp2_swap_0.3_last_to_carry_2128

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.5511
• Accuracy: 0.3740

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: 0.0006
• train_batch_size: 16
• eval_batch_size: 16
• seed: 2128
• gradient_accumulation_steps: 5
• total_train_batch_size: 80
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.98) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 50.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Accuracy	Validation Loss
4.8538	0.2915	1000	0.2511	4.7815
4.3553	0.5830	2000	0.2974	4.2970
4.1572	0.8745	3000	0.3145	4.1083
4.0092	1.1659	4000	0.3239	4.0004
3.9442	1.4574	5000	0.3307	3.9270
3.8877	1.7488	6000	0.3357	3.8676
3.7633	2.0402	7000	0.3400	3.8234
3.7513	2.3317	8000	0.3428	3.7955
3.7464	2.6232	9000	0.3457	3.7640
3.7303	2.9147	10000	0.3482	3.7385
3.6382	3.2061	11000	0.3501	3.7256
3.6536	3.4976	12000	0.3515	3.7074
3.6604	3.7891	13000	0.3535	3.6882
3.5521	4.0805	14000	0.3546	3.6830
3.5792	4.3719	15000	0.3556	3.6708
3.5804	4.6634	16000	0.3571	3.6566
3.5891	4.9549	17000	0.3582	3.6436
3.5154	5.2463	18000	0.3590	3.6486
3.5389	5.5378	19000	0.3594	3.6370
3.5427	5.8293	20000	0.3607	3.6244
3.4396	6.1207	21000	0.3608	3.6279
3.4863	6.4122	22000	0.3618	3.6184
3.5084	6.7037	23000	0.3623	3.6104
3.5037	6.9952	24000	0.3629	3.6005
3.4393	7.2865	25000	0.3631	3.6119
3.4519	7.5780	26000	0.3639	3.6029
3.4802	7.8695	27000	0.3647	3.5909
3.3922	8.1609	28000	0.3647	3.5990
3.4089	8.4524	29000	0.3651	3.5945
3.4461	8.7439	30000	0.3655	3.5861
3.3276	9.0353	31000	0.3657	3.5891
3.383	9.3268	32000	0.3659	3.5871
3.4183	9.6183	33000	0.3664	3.5780
3.4229	9.9098	34000	0.3673	3.5723
3.3509	10.2011	35000	0.3664	3.5884
3.3786	10.4926	36000	0.3673	3.5770
3.391	10.7841	37000	0.3678	3.5691
3.2991	11.0755	38000	0.3673	3.5813
3.3412	11.3670	39000	0.3677	3.5766
3.3699	11.6585	40000	0.3684	3.5663
3.3945	11.9500	41000	0.3689	3.5602
3.3053	12.2414	42000	0.3682	3.5775
3.3438	12.5329	43000	0.3689	3.5682
3.3573	12.8243	44000	0.3690	3.5603
3.2803	13.1157	45000	0.3689	3.5727
3.3108	13.4072	46000	0.3690	3.5683
3.3295	13.6987	47000	0.3697	3.5559
3.3545	13.9902	48000	0.3703	3.5509
3.2864	14.2816	49000	0.3694	3.5652
3.3191	14.5731	50000	0.3701	3.5597
3.3265	14.8646	51000	0.3702	3.5527
3.2442	15.1559	52000	0.3701	3.5673
3.2987	15.4474	53000	0.3700	3.5612
3.3131	15.7389	54000	0.3708	3.5536
3.208	16.0303	55000	0.3706	3.5630
3.2701	16.3218	56000	0.3702	3.5624
3.2961	16.6133	57000	0.3711	3.5522
3.3121	16.9048	58000	0.3715	3.5455
3.2413	17.1962	59000	0.3708	3.5617
3.2822	17.4877	60000	0.3711	3.5555
3.2867	17.7792	61000	0.3718	3.5461
3.208	18.0705	62000	0.3712	3.5618
3.2382	18.3620	63000	0.3711	3.5593
3.271	18.6535	64000	0.3715	3.5497
3.2787	18.9450	65000	0.3719	3.5430
3.2242	19.2364	66000	0.3714	3.5594
3.2414	19.5279	67000	0.3716	3.5537
3.263	19.8194	68000	0.3722	3.5447
3.1872	20.1108	69000	0.3715	3.5611
3.2395	20.4023	70000	0.3718	3.5585
3.2498	20.6938	71000	0.3725	3.5478
3.2642	20.9853	72000	0.3729	3.5396
3.1952	21.2766	73000	0.3717	3.5607
3.2423	21.5681	74000	0.3725	3.5494
3.2351	21.8596	75000	0.3727	3.5432
3.1864	22.1510	76000	0.3721	3.5615
3.2091	22.4425	77000	0.3724	3.5523
3.2382	22.7340	78000	0.3728	3.5462
3.1495	23.0254	79000	0.3723	3.5588
3.1918	23.3169	80000	0.3722	3.5558
3.1792	23.6083	81000	3.5608	0.3722
3.2238	23.8998	82000	3.5523	0.3728
3.1709	24.1915	83000	3.5627	0.3723
3.2006	24.4830	84000	3.5556	0.3725
3.2189	24.7745	85000	3.5463	0.3729
3.1311	25.0659	86000	3.5606	0.3725
3.1849	25.3574	87000	3.5562	0.3729
3.1981	25.6489	88000	3.5496	0.3733
3.2249	25.9404	89000	3.5394	0.3739
3.1593	26.2317	90000	3.5611	0.3727
3.1831	26.5232	91000	3.5512	0.3731
3.2065	26.8147	92000	3.5430	0.3737
3.1175	27.1061	93000	3.5603	0.3733
3.1568	27.3976	94000	3.5579	0.3729
3.1802	27.6891	95000	3.5471	0.3735
3.188	27.9806	96000	3.5395	0.3740
3.1366	28.2720	97000	3.5598	0.3729
3.1722	28.5635	98000	3.5544	0.3733
3.189	28.8550	99000	3.5446	0.3740
3.114	29.1463	100000	3.5620	0.3734
3.1485	29.4378	101000	3.5540	0.3736
3.1665	29.7293	102000	3.5469	0.3740
3.0917	30.0207	103000	3.5574	0.3735
3.129	30.3122	104000	3.5562	0.3735
3.1507	30.6037	105000	3.5523	0.3737
3.1766	30.8952	106000	3.5448	0.3742
3.0988	31.1866	107000	3.5617	0.3732
3.1365	31.4781	108000	3.5554	0.3737
3.1547	31.7695	109000	3.5511	0.3740

Framework versions

• Transformers 4.55.2
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.21.4