exceptions_exp2_swap_0.7_last_to_push_3591

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

• Loss: 3.5795
• Accuracy: 0.3663

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
4.8245	0.2915	1000	4.7479	0.2552
4.3478	0.5830	2000	4.2903	0.2982
4.1604	0.8745	3000	4.1026	0.3142
3.9932	1.1659	4000	3.9907	0.3247
3.94	1.4574	5000	3.9161	0.3315
3.8846	1.7489	6000	3.8589	0.3363
3.7481	2.0402	7000	3.8169	0.3411
3.7434	2.3317	8000	3.7852	0.3440
3.7405	2.6233	9000	3.7572	0.3464
3.7195	2.9148	10000	3.7314	0.3489
3.6446	3.2061	11000	3.7183	0.3507
3.6418	3.4976	12000	3.6975	0.3526
3.638	3.7891	13000	3.6815	0.3546
3.5452	4.0805	14000	3.6747	0.3553
3.5639	4.3720	15000	3.6651	0.3565
3.5757	4.6635	16000	3.6487	0.3579
3.5794	4.9550	17000	3.6345	0.3592
3.4961	5.2463	18000	3.6378	0.3595
3.5291	5.5378	19000	3.6263	0.3604
3.5477	5.8293	20000	3.6187	0.3614
3.4426	6.1207	21000	3.6190	0.3619
3.4796	6.4122	22000	3.6126	0.3624
3.5024	6.7037	23000	3.6031	0.3634
3.496	6.9952	24000	3.5931	0.3640
3.4205	7.2866	25000	3.6021	0.3639
3.4508	7.5781	26000	3.5913	0.3647
3.453	7.8696	27000	3.5832	0.3655
3.3847	8.1609	28000	3.5938	0.3651
3.4229	8.4524	29000	3.5880	0.3657
3.4257	8.7439	30000	3.5795	0.3663
3.3431	9.0353	31000	3.5823	0.3665
3.3647	9.3268	32000	3.5805	0.3668
3.4024	9.6183	33000	3.5725	0.3671
3.4142	9.9098	34000	3.5659	0.3679
3.3529	10.2011	35000	3.5776	0.3674
3.3466	10.4927	36000	3.5697	0.3679
3.3845	10.7842	37000	3.5609	0.3684
3.2849	11.0755	38000	3.5759	0.3683
3.3503	11.3670	39000	3.5705	0.3686
3.3581	11.6585	40000	3.5622	0.3689
3.3793	11.9500	41000	3.5508	0.3697
3.3093	12.2414	42000	3.5660	0.3692
3.3391	12.5329	43000	3.5611	0.3695
3.3551	12.8244	44000	3.5527	0.3699
3.2685	13.1157	45000	3.5655	0.3696
3.2986	13.4072	46000	3.5590	0.3699
3.3224	13.6988	47000	3.5538	0.3703
3.35	13.9903	48000	3.5467	0.3709
3.2732	14.2816	49000	3.5643	0.3701
3.3097	14.5731	50000	3.5526	0.3708
3.322	14.8646	51000	3.5463	0.3712
3.2602	15.1560	52000	3.5608	0.3706
3.2819	15.4475	53000	3.5526	0.3710
3.3111	15.7390	54000	3.5436	0.3715
3.2119	16.0303	55000	3.5584	0.3710
3.2562	16.3218	56000	3.5544	0.3711
3.2765	16.6133	57000	3.5489	0.3718
3.2937	16.9049	58000	3.5420	0.3721
3.2387	17.1962	59000	3.5550	0.3716
3.2742	17.4877	60000	3.5513	0.3715
3.2854	17.7792	61000	3.5436	0.3723
3.2008	18.0705	62000	3.5575	0.3716
3.2348	18.3621	63000	3.5541	0.3717
3.2599	18.6536	64000	3.5466	0.3721
3.2804	18.9451	65000	3.5381	0.3726
3.2044	19.2364	66000	3.5542	0.3718
3.2455	19.5279	67000	3.5479	0.3724
3.249	19.8194	68000	3.5413	0.3728
3.188	20.1108	69000	3.5549	0.3718
3.2149	20.4023	70000	3.5513	0.3721
3.2423	20.6938	71000	3.5432	0.3729
3.2562	20.9853	72000	3.5370	0.3734
3.1981	21.2766	73000	3.5497	0.3724
3.224	21.5682	74000	3.5434	0.3732
3.2341	21.8597	75000	3.5370	0.3732
3.1725	22.1510	76000	3.5544	0.3725
3.2035	22.4425	77000	3.5506	0.3727
3.2282	22.7340	78000	3.5404	0.3731
3.1312	23.0254	79000	3.5542	0.3730
3.195	23.3169	80000	3.5524	0.3731
3.213	23.6084	81000	3.5435	0.3735
3.2259	23.8999	82000	3.5350	0.3739
3.1554	24.1912	83000	3.5550	0.3727
3.1986	24.4827	84000	3.5487	0.3733
3.2046	24.7743	85000	3.5418	0.3737
3.1261	25.0656	86000	3.5551	0.3732
3.1594	25.3571	87000	3.5527	0.3731
3.1924	25.6486	88000	3.5469	0.3735
3.1973	25.9401	89000	3.5385	0.3740
3.1322	26.2315	90000	3.5544	0.3730
3.1663	26.5230	91000	3.5484	0.3738
3.1819	26.8145	92000	3.5396	0.3743
3.1263	27.1058	93000	3.5573	0.3734
3.1514	27.3973	94000	3.5530	0.3733
3.1695	27.6888	95000	3.5472	0.3739
3.1772	27.9804	96000	3.5405	0.3744
3.144	28.2717	97000	3.5572	0.3736
3.1562	28.5632	98000	3.5473	0.3740
3.1756	28.8547	99000	3.5411	0.3742
3.1147	29.1460	100000	3.5560	0.3737
3.1334	29.4376	101000	3.5500	0.3739
3.1493	29.7291	102000	3.5447	0.3745

Framework versions

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