exceptions_exp2_swap_last_to_push_5039

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

• Loss: 3.5613
• Accuracy: 0.3689

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: 5039
• 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
4.827	0.2915	1000	4.7561	0.2547
4.3473	0.5830	2000	4.2888	0.2977
4.1507	0.8744	3000	4.1051	0.3144
4.0024	1.1659	4000	3.9950	0.3243
3.9542	1.4573	5000	3.9192	0.3311
3.8828	1.7488	6000	3.8607	0.3362
3.7609	2.0402	7000	3.8193	0.3405
3.758	2.3317	8000	3.7892	0.3433
3.7488	2.6232	9000	3.7577	0.3461
3.7284	2.9147	10000	3.7323	0.3486
3.6425	3.2061	11000	3.7196	0.3505
3.6524	3.4976	12000	3.7040	0.3524
3.6604	3.7890	13000	3.6831	0.3539
3.5523	4.0804	14000	3.6766	0.3556
3.565	4.3719	15000	3.6654	0.3563
3.5805	4.6634	16000	3.6526	0.3574
3.5797	4.9549	17000	3.6396	0.3584
3.5128	5.2463	18000	3.6400	0.3596
3.5289	5.5378	19000	3.6306	0.3601
3.5303	5.8293	20000	3.6181	0.3611
3.4462	6.1207	21000	3.6234	0.3613
3.473	6.4121	22000	3.6152	0.3624
3.4919	6.7036	23000	3.6045	0.3629
3.5037	6.9951	24000	3.5964	0.3636
3.4324	7.2865	25000	3.6029	0.3637
3.4514	7.5780	26000	3.5965	0.3646
3.4773	7.8695	27000	3.5874	0.3650
3.3931	8.1609	28000	3.5939	0.3650
3.4055	8.4524	29000	3.5862	0.3656
3.432	8.7438	30000	3.5800	0.3663
3.3373	9.0353	31000	3.5845	0.3660
3.382	9.3267	32000	3.5838	0.3663
3.4042	9.6182	33000	3.5730	0.3671
3.4188	9.9097	34000	3.5657	0.3676
3.3418	10.2011	35000	3.5763	0.3673
3.3722	10.4926	36000	3.5718	0.3673
3.3859	10.7841	37000	3.5641	0.3684
3.2956	11.0755	38000	3.5723	0.3682
3.3474	11.3670	39000	3.5705	0.3683
3.3665	11.6584	40000	3.5613	0.3689
3.3872	11.9499	41000	3.5527	0.3697
3.3103	12.2413	42000	3.5694	0.3690
3.3378	12.5328	43000	3.5595	0.3695
3.3613	12.8243	44000	3.5549	0.3698
3.2731	13.1157	45000	3.5644	0.3697
3.3164	13.4072	46000	3.5597	0.3699
3.3302	13.6987	47000	3.5525	0.3702
3.3409	13.9901	48000	3.5481	0.3706
3.2917	14.2816	49000	3.5600	0.3703
3.3057	14.5730	50000	3.5509	0.3708
3.3215	14.8645	51000	3.5454	0.3712
3.2559	15.1559	52000	3.5625	0.3704
3.2948	15.4474	53000	3.5552	0.3707
3.3043	15.7389	54000	3.5497	0.3715
3.2092	16.0303	55000	3.5574	0.3709
3.2667	16.3218	56000	3.5580	0.3711
3.2924	16.6133	57000	3.5493	0.3714
3.2971	16.9047	58000	3.5428	0.3718
3.2266	17.1962	59000	3.5594	0.3712
3.2635	17.4876	60000	3.5502	0.3716
3.2861	17.7791	61000	3.5422	0.3720
3.2064	18.0705	62000	3.5580	0.3714
3.2506	18.3620	63000	3.5526	0.3719
3.2589	18.6535	64000	3.5471	0.3721
3.284	18.9450	65000	3.5395	0.3725
3.2215	19.2364	66000	3.5595	0.3716
3.2425	19.5279	67000	3.5483	0.3722
3.2593	19.8193	68000	3.5429	0.3729
3.2034	20.1108	69000	3.5518	0.3721
3.2154	20.4022	70000	3.5472	0.3723
3.2438	20.6937	71000	3.5431	0.3727
3.2696	20.9852	72000	3.5344	0.3731
3.2131	21.2766	73000	3.5550	0.3723
3.2219	21.5681	74000	3.5457	0.3729
3.2434	21.8596	75000	3.5405	0.3731
3.1707	22.1510	76000	3.5526	0.3727
3.1933	22.4425	77000	3.5484	0.3728
3.2346	22.7339	78000	3.5404	0.3732
3.1206	23.0254	79000	3.5509	0.3729
3.1772	23.3168	80000	3.5538	0.3726
3.207	23.6083	81000	3.5405	0.3735
3.216	23.8998	82000	3.5379	0.3740
3.1572	24.1912	83000	3.5531	0.3726
3.1916	24.4827	84000	3.5453	0.3733
3.2191	24.7742	85000	3.5401	0.3737
3.1353	25.0656	86000	3.5544	0.3729
3.1683	25.3571	87000	3.5516	0.3733
3.2059	25.6485	88000	3.5426	0.3737
3.2038	25.9400	89000	3.5346	0.3742
3.1591	26.2314	90000	3.5534	0.3730
3.1674	26.5229	91000	3.5457	0.3736
3.2005	26.8144	92000	3.5387	0.3741

Framework versions

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