exceptions_exp2_swap_0.3_resemble_to_drop_1032

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

• Loss: 3.5659
• Accuracy: 0.3685

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: 1032
• 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	Validation Loss	Accuracy
4.8372	0.2915	1000	4.7643	0.2534
4.3477	0.5830	2000	4.2867	0.2983
4.1598	0.8745	3000	4.1033	0.3144
3.9885	1.1659	4000	4.0003	0.3236
3.9421	1.4574	5000	3.9232	0.3303
3.898	1.7488	6000	3.8636	0.3360
3.7581	2.0402	7000	3.8225	0.3404
3.7538	2.3317	8000	3.7917	0.3433
3.755	2.6232	9000	3.7653	0.3457
3.733	2.9147	10000	3.7390	0.3486
3.6438	3.2061	11000	3.7249	0.3503
3.6564	3.4976	12000	3.7043	0.3523
3.6559	3.7891	13000	3.6871	0.3537
3.5495	4.0805	14000	3.6836	0.3545
3.5735	4.3719	15000	3.6702	0.3559
3.5772	4.6634	16000	3.6576	0.3569
3.5833	4.9549	17000	3.6453	0.3583
3.5291	5.2463	18000	3.6441	0.3590
3.5344	5.5378	19000	3.6324	0.3599
3.5421	5.8293	20000	3.6228	0.3607
3.4496	6.1207	21000	3.6267	0.3612
3.4801	6.4122	22000	3.6220	0.3617
3.4908	6.7037	23000	3.6100	0.3623
3.4982	6.9952	24000	3.6022	0.3633
3.4425	7.2865	25000	3.6104	0.3629
3.4742	7.5780	26000	3.6011	0.3638
3.4658	7.8695	27000	3.5919	0.3645
3.3838	8.1609	28000	3.6008	0.3643
3.4179	8.4524	29000	3.5926	0.3650
3.4366	8.7439	30000	3.5834	0.3655
3.3393	9.0353	31000	3.5909	0.3655
3.3892	9.3268	32000	3.5902	0.3659
3.4042	9.6183	33000	3.5814	0.3663
3.4278	9.9098	34000	3.5714	0.3671
3.3403	10.2011	35000	3.5843	0.3665
3.3798	10.4926	36000	3.5770	0.3673
3.4003	10.7841	37000	3.5692	0.3677
3.2965	11.0755	38000	3.5790	0.3678
3.3595	11.3670	39000	3.5782	0.3677
3.3747	11.6585	40000	3.5659	0.3685
3.3759	11.9500	41000	3.5604	0.3686
3.3052	12.2414	42000	3.5742	0.3683
3.3344	12.5329	43000	3.5673	0.3690
3.3684	12.8243	44000	3.5582	0.3695
3.2613	13.1157	45000	3.5724	0.3691
3.3154	13.4072	46000	3.5687	0.3690
3.3432	13.6987	47000	3.5585	0.3697
3.3539	13.9902	48000	3.5509	0.3701
3.2991	14.2816	49000	3.5669	0.3695
3.3089	14.5731	50000	3.5585	0.3700
3.3313	14.8646	51000	3.5511	0.3703
3.2787	15.1559	52000	3.5670	0.3695
3.2881	15.4474	53000	3.5626	0.3701
3.3128	15.7389	54000	3.5519	0.3708
3.2137	16.0303	55000	3.5637	0.3705
3.2641	16.3218	56000	3.5591	0.3706
3.2928	16.6133	57000	3.5520	0.3710
3.3101	16.9048	58000	3.5469	0.3717
3.2236	17.1962	59000	3.5636	0.3708
3.2727	17.4877	60000	3.5546	0.3712
3.2801	17.7792	61000	3.5492	0.3715
3.2099	18.0705	62000	3.5611	0.3710
3.2471	18.3620	63000	3.5570	0.3712
3.2681	18.6535	64000	3.5521	0.3716
3.2895	18.9450	65000	3.5407	0.3722
3.2241	19.2364	66000	3.5609	0.3715
3.2501	19.5279	67000	3.5536	0.3719
3.2767	19.8194	68000	3.5454	0.3722
3.2055	20.1108	69000	3.5618	0.3715
3.227	20.4023	70000	3.5563	0.3719
3.2437	20.6938	71000	3.5463	0.3723
3.2578	20.9853	72000	3.5410	0.3726
3.1969	21.2766	73000	3.5576	0.3719
3.2406	21.5681	74000	3.5547	0.3718
3.244	21.8596	75000	3.5418	0.3730
3.1779	22.1510	76000	3.5583	0.3723
3.2128	22.4425	77000	3.5538	0.3724
3.2358	22.7340	78000	3.5461	0.3725
3.1455	23.0254	79000	3.5582	0.3725
3.1967	23.3169	80000	3.5586	0.3720
3.2203	23.6083	81000	3.5515	0.3727
3.2325	23.8998	82000	3.5448	0.3732
3.1687	24.1912	83000	3.5581	0.3723
3.1897	24.4827	84000	3.5538	0.3730
3.2194	24.7742	85000	3.5467	0.3733

Framework versions

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