exceptions_exp2_swap_last_to_drop_3591

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

• Loss: 3.5637
• Accuracy: 0.3686

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: 3591
• 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
4.8397	0.2915	1000	4.7697	0.2530
4.3447	0.5830	2000	4.2928	0.2975
4.1526	0.8744	3000	4.1036	0.3143
4.0167	1.1659	4000	3.9991	0.3242
3.9225	1.4573	5000	3.9228	0.3306
3.8868	1.7488	6000	3.8631	0.3354
3.7436	2.0402	7000	3.8193	0.3406
3.7519	2.3317	8000	3.7919	0.3427
3.7466	2.6232	9000	3.7611	0.3459
3.7297	2.9147	10000	3.7348	0.3482
3.6392	3.2061	11000	3.7210	0.3504
3.6527	3.4976	12000	3.7038	0.3520
3.655	3.7890	13000	3.6840	0.3538
3.559	4.0804	14000	3.6798	0.3550
3.5764	4.3719	15000	3.6675	0.3562
3.5791	4.6634	16000	3.6537	0.3573
3.5802	4.9549	17000	3.6406	0.3582
3.5149	5.2463	18000	3.6425	0.3589
3.5211	5.5378	19000	3.6322	0.3599
3.5469	5.8293	20000	3.6208	0.3607
3.4508	6.1207	21000	3.6234	0.3614
3.4785	6.4121	22000	3.6181	0.3615
3.4973	6.7036	23000	3.6064	0.3627
3.4985	6.9951	24000	3.5986	0.3635
3.4349	7.2865	25000	3.6063	0.3633
3.4554	7.5780	26000	3.5974	0.3640
3.4507	7.8695	27000	3.5870	0.3648
3.3821	8.1609	28000	3.5973	0.3646
3.4311	8.4524	29000	3.5911	0.3652
3.4313	8.7438	30000	3.5809	0.3659
3.3386	9.0353	31000	3.5859	0.3661
3.3808	9.3267	32000	3.5843	0.3662
3.4057	9.6182	33000	3.5756	0.3668
3.4165	9.9097	34000	3.5674	0.3673
3.3531	10.2011	35000	3.5814	0.3670
3.3798	10.4926	36000	3.5721	0.3673
3.4046	10.7841	37000	3.5651	0.3680
3.2851	11.0755	38000	3.5771	0.3676
3.3606	11.3670	39000	3.5732	0.3679
3.375	11.6584	40000	3.5637	0.3686
3.3827	11.9499	41000	3.5549	0.3688
3.3096	12.2413	42000	3.5724	0.3683
3.3418	12.5328	43000	3.5648	0.3691
3.3459	12.8243	44000	3.5559	0.3693
3.2627	13.1157	45000	3.5676	0.3693
3.3173	13.4072	46000	3.5611	0.3696
3.3405	13.6987	47000	3.5541	0.3699
3.3502	13.9901	48000	3.5477	0.3702
3.2876	14.2816	49000	3.5621	0.3701
3.3196	14.5730	50000	3.5537	0.3704
3.3248	14.8645	51000	3.5464	0.3706
3.2495	15.1559	52000	3.5632	0.3700
3.2921	15.4474	53000	3.5566	0.3706
3.3052	15.7389	54000	3.5463	0.3708
3.2087	16.0303	55000	3.5620	0.3703
3.275	16.3218	56000	3.5602	0.3708
3.2882	16.6133	57000	3.5513	0.3708
3.3085	16.9047	58000	3.5420	0.3717
3.2232	17.1962	59000	3.5616	0.3710
3.2553	17.4876	60000	3.5491	0.3717
3.2772	17.7791	61000	3.5456	0.3715
3.2025	18.0705	62000	3.5570	0.3712
3.2415	18.3620	63000	3.5561	0.3713
3.2605	18.6535	64000	3.5488	0.3720
3.2855	18.9450	65000	3.5376	0.3726
3.2222	19.2364	66000	3.5544	0.3716
3.2498	19.5279	67000	3.5489	0.3718
3.2732	19.8193	68000	3.5396	0.3725
3.189	20.1108	69000	3.5582	0.3717
3.2291	20.4022	70000	3.5530	0.3722
3.2326	20.6937	71000	3.5448	0.3726
3.2718	20.9852	72000	3.5384	0.3726
3.2028	21.2766	73000	3.5557	0.3721
3.2232	21.5681	74000	3.5476	0.3725
3.2406	21.8596	75000	3.5391	0.3730
3.1767	22.1510	76000	3.5607	0.3722
3.2126	22.4425	77000	3.5493	0.3725
3.2327	22.7339	78000	3.5400	0.3731
3.1332	23.0254	79000	3.5529	0.3728
3.1902	23.3168	80000	3.5519	0.3724
3.2092	23.6083	81000	3.5466	0.3732
3.2151	23.8998	82000	3.5412	0.3735
3.1647	24.1912	83000	3.5589	0.3728
3.2056	24.4827	84000	3.5529	0.3729
3.208	24.7742	85000	3.5441	0.3732

Framework versions

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