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	Accuracy	Validation Loss
4.8202	0.2915	1000	0.2554	4.7525
4.3503	0.5830	2000	0.2982	4.2911
4.1524	0.8745	3000	0.3143	4.1055
4.0106	1.1659	4000	0.3238	4.0014
3.937	1.4574	5000	0.3306	3.9254
3.8757	1.7489	6000	0.3357	3.8667
3.7477	2.0402	7000	0.3399	3.8241
3.7649	2.3317	8000	0.3430	3.7917
3.746	2.6233	9000	0.3458	3.7633
3.7398	2.9148	10000	0.3479	3.7371
3.6517	3.2061	11000	0.3499	3.7230
3.6517	3.4976	12000	0.3520	3.7060
3.6558	3.7891	13000	0.3536	3.6876
3.5583	4.0805	14000	0.3549	3.6809
3.5754	4.3720	15000	0.3560	3.6693
3.5703	4.6635	16000	0.3572	3.6562
3.5782	4.9550	17000	0.3581	3.6433
3.502	5.2463	18000	0.3589	3.6448
3.5243	5.5378	19000	0.3599	3.6332
3.5438	5.8293	20000	0.3609	3.6211
3.4521	6.1207	21000	0.3614	3.6280
3.4677	6.4122	22000	0.3618	3.6181
3.5009	6.7037	23000	0.3625	3.6105
3.4934	6.9952	24000	0.3632	3.5996
3.4407	7.2866	25000	0.3633	3.6062
3.4512	7.5781	26000	0.3641	3.5965
3.4615	7.8696	27000	0.3646	3.5903
3.3907	8.1609	28000	0.3648	3.5984
3.4155	8.4524	29000	0.3652	3.5909
3.4241	8.7439	30000	0.3659	3.5819
3.3253	9.0353	31000	0.3661	3.5876
3.3841	9.3268	32000	0.3661	3.5874
3.39	9.6183	33000	0.3667	3.5803
3.4313	9.9098	34000	0.3675	3.5703
3.3379	10.2011	35000	0.3670	3.5815
3.3805	10.4927	36000	0.3675	3.5738
3.3939	10.7842	37000	0.3676	3.5656
3.2996	11.0755	38000	0.3680	3.5777
3.346	11.3670	39000	0.3680	3.5717
3.3718	11.6585	40000	0.3684	3.5659
3.377	11.9500	41000	0.3694	3.5564
3.3106	12.2414	42000	0.3686	3.5715
3.3468	12.5329	43000	0.3691	3.5657
3.3497	12.8244	44000	0.3696	3.5560
3.2763	13.1157	45000	0.3689	3.5708
3.3259	13.4072	46000	0.3694	3.5661
3.3433	13.6988	47000	0.3698	3.5552
3.349	13.9903	48000	0.3706	3.5488
3.2887	14.2816	49000	0.3696	3.5682
3.3125	14.5731	50000	0.3699	3.5570
3.3122	14.8646	51000	0.3706	3.5501
3.2578	15.1560	52000	0.3698	3.5673
3.2948	15.4475	53000	0.3703	3.5594
3.3098	15.7390	54000	0.3708	3.5498
3.208	16.0303	55000	0.3704	3.5624
3.2617	16.3218	56000	0.3704	3.5617
3.2922	16.6133	57000	0.3708	3.5562
3.2972	16.9049	58000	0.3713	3.5484
3.2232	17.1962	59000	0.3706	3.5616
3.2655	17.4877	60000	0.3708	3.5564
3.293	17.7792	61000	0.3717	3.5497
3.1985	18.0705	62000	0.3709	3.5640
3.2372	18.3621	63000	0.3712	3.5562
3.2506	18.6536	64000	0.3719	3.5508
3.2913	18.9451	65000	0.3720	3.5429
3.205	19.2364	66000	0.3716	3.5575
3.2488	19.5279	67000	0.3718	3.5558
3.2746	19.8194	68000	0.3722	3.5475
3.1991	20.1108	69000	0.3714	3.5593
3.2271	20.4023	70000	0.3718	3.5549
3.2509	20.6938	71000	0.3722	3.5489
3.2677	20.9853	72000	0.3728	3.5388
3.2061	21.2766	73000	0.3720	3.5555
3.2184	21.5682	74000	0.3724	3.5485
3.2365	21.8597	75000	0.3728	3.5423
3.1774	22.1510	76000	0.3717	3.5594
3.2043	22.4425	77000	0.3724	3.5513
3.2224	22.7340	78000	0.3727	3.5486
3.1395	23.0254	79000	0.3724	3.5592
3.1915	23.3169	80000	0.3722	3.5563
3.1811	23.6084	81000	3.5653	0.3719
3.2088	23.8999	82000	3.5498	0.3726
3.1596	24.1915	83000	3.5613	0.3720
3.195	24.4830	84000	3.5549	0.3725
3.2232	24.7745	85000	3.5471	0.3729
3.1365	25.0659	86000	3.5616	0.3721
3.1695	25.3574	87000	3.5583	0.3725
3.1984	25.6489	88000	3.5512	0.3731
3.2088	25.9404	89000	3.5413	0.3736
3.1446	26.2318	90000	3.5628	0.3725
3.1799	26.5233	91000	3.5528	0.3730
3.1989	26.8148	92000	3.5470	0.3734
3.1112	27.1061	93000	3.5628	0.3730
3.1655	27.3976	94000	3.5582	0.3729
3.1834	27.6891	95000	3.5454	0.3736
3.1862	27.9806	96000	3.5432	0.3739
3.1438	28.2720	97000	3.5582	0.3730
3.1684	28.5635	98000	3.5519	0.3735
3.1862	28.8550	99000	3.5458	0.3739
3.1138	29.1463	100000	3.5621	0.3731

Framework versions

Transformers 4.55.2
Pytorch 2.8.0+cu128
Datasets 4.0.0
Tokenizers 0.21.4

craa/exceptions_exp2_swap_0.7_last_to_hit_5039

exceptions_exp2_swap_0.7_last_to_hit_5039

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

craa/
exceptions_exp2_swap_0.7_last_to_hit_5039