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
4.8627	0.2915	1000	4.7844	0.2501
4.3539	0.5830	2000	4.2921	0.2980
4.1413	0.8745	3000	4.0998	0.3147
3.9978	1.1659	4000	3.9957	0.3238
3.9403	1.4574	5000	3.9182	0.3310
3.8819	1.7488	6000	3.8626	0.3362
3.7431	2.0402	7000	3.8177	0.3404
3.7531	2.3317	8000	3.7886	0.3437
3.7484	2.6232	9000	3.7565	0.3464
3.7252	2.9147	10000	3.7308	0.3487
3.6226	3.2061	11000	3.7193	0.3506
3.6547	3.4976	12000	3.7019	0.3523
3.637	3.7891	13000	3.6811	0.3542
3.5408	4.0805	14000	3.6759	0.3556
3.5757	4.3719	15000	3.6660	0.3564
3.5739	4.6634	16000	3.6493	0.3576
3.5802	4.9549	17000	3.6362	0.3588
3.5099	5.2463	18000	3.6397	0.3594
3.5369	5.5378	19000	3.6304	0.3601
3.5126	5.8293	20000	3.6168	0.3615
3.4452	6.1207	21000	3.6223	0.3615
3.4752	6.4122	22000	3.6119	0.3623
3.501	6.7037	23000	3.6024	0.3630
3.5034	6.9952	24000	3.5956	0.3639
3.4441	7.2865	25000	3.6024	0.3641
3.4538	7.5780	26000	3.5936	0.3644
3.4636	7.8695	27000	3.5863	0.3651
3.405	8.1609	28000	3.5926	0.3653
3.4264	8.4524	29000	3.5872	0.3654
3.4503	8.7439	30000	3.5775	0.3662
3.3354	9.0353	31000	3.5844	0.3660
3.3905	9.3268	32000	3.5852	0.3663
3.4035	9.6183	33000	3.5740	0.3671
3.4073	9.9098	34000	3.5659	0.3679
3.3267	10.2011	35000	3.5812	0.3673
3.3804	10.4926	36000	3.5730	0.3680
3.3781	10.7841	37000	3.5659	0.3682
3.3095	11.0755	38000	3.5744	0.3681
3.3431	11.3670	39000	3.5715	0.3683
3.3663	11.6585	40000	3.5625	0.3686
3.3794	11.9500	41000	3.5569	0.3694
3.3	12.2414	42000	3.5677	0.3689
3.3442	12.5329	43000	3.5605	0.3693
3.3568	12.8243	44000	3.5496	0.3703
3.2791	13.1157	45000	3.5656	0.3694
3.328	13.4072	46000	3.5615	0.3698
3.3301	13.6987	47000	3.5516	0.3703
3.3459	13.9902	48000	3.5439	0.3710
3.2893	14.2816	49000	3.5609	0.3701
3.3083	14.5731	50000	3.5533	0.3707
3.327	14.8646	51000	3.5483	0.3710
3.2497	15.1559	52000	3.5610	0.3705
3.2812	15.4474	53000	3.5555	0.3707
3.3065	15.7389	54000	3.5494	0.3711
3.2118	16.0303	55000	3.5592	0.3707
3.2646	16.3218	56000	3.5571	0.3711
3.2876	16.6133	57000	3.5480	0.3714
3.3007	16.9048	58000	3.5421	0.3717
3.2349	17.1962	59000	3.5551	0.3715
3.265	17.4877	60000	3.5531	0.3715
3.2788	17.7792	61000	3.5488	0.3715
3.2039	18.0705	62000	3.5559	0.3716
3.2284	18.3620	63000	3.5550	0.3717
3.2642	18.6535	64000	3.5453	0.3721
3.2778	18.9450	65000	3.5394	0.3723
3.2227	19.2364	66000	3.5561	0.3716
3.2441	19.5279	67000	3.5491	0.3721
3.2613	19.8194	68000	3.5433	0.3726
3.1872	20.1108	69000	3.5609	0.3718
3.231	20.4023	70000	3.5553	0.3721
3.2504	20.6938	71000	3.5449	0.3724
3.2609	20.9853	72000	3.5352	0.3732
3.2085	21.2766	73000	3.5548	0.3720
3.2198	21.5681	74000	3.5482	0.3726
3.2399	21.8596	75000	3.5386	0.3732
3.1757	22.1510	76000	3.5555	0.3720
3.2121	22.4425	77000	3.5502	0.3728
3.2182	22.7340	78000	3.5434	0.3731
3.1447	23.0254	79000	3.5546	0.3724
3.1889	23.3169	80000	3.5531	0.3727
3.2171	23.6083	81000	3.5461	0.3730
3.2313	23.8998	82000	3.5389	0.3735
3.1645	24.1912	83000	3.5569	0.3728
3.1987	24.4827	84000	3.5473	0.3732
3.209	24.7742	85000	3.5431	0.3733
3.1183	25.0656	86000	3.5545	0.3730
3.1783	25.3571	87000	3.5518	0.3733
3.1976	25.6486	88000	3.5480	0.3733
3.2137	25.9401	89000	3.5378	0.3740
3.1519	26.2314	90000	3.5551	0.3730
3.1801	26.5229	91000	3.5485	0.3735
3.1844	26.8144	92000	3.5414	0.3741

Framework versions

Transformers 4.55.2
Pytorch 2.8.0+cu128
Datasets 4.0.0
Tokenizers 0.21.4

craa/exceptions_exp2_swap_0.3_resemble_to_drop_3591

exceptions_exp2_swap_0.3_resemble_to_drop_3591

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

craa/
exceptions_exp2_swap_0.3_resemble_to_drop_3591