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
4.8273	0.2915	1000	4.7552	0.2545
4.3431	0.5830	2000	4.2942	0.2975
4.1538	0.8744	3000	4.1027	0.3144
4.0175	1.1659	4000	3.9972	0.3241
3.9211	1.4573	5000	3.9206	0.3308
3.8863	1.7488	6000	3.8629	0.3353
3.7427	2.0402	7000	3.8191	0.3405
3.7511	2.3317	8000	3.7889	0.3429
3.7434	2.6232	9000	3.7594	0.3462
3.7283	2.9147	10000	3.7311	0.3486
3.637	3.2061	11000	3.7184	0.3505
3.6497	3.4976	12000	3.7015	0.3523
3.6512	3.7890	13000	3.6807	0.3540
3.5542	4.0804	14000	3.6765	0.3553
3.5731	4.3719	15000	3.6645	0.3565
3.576	4.6634	16000	3.6507	0.3579
3.5775	4.9549	17000	3.6382	0.3587
3.5114	5.2463	18000	3.6387	0.3592
3.5171	5.5378	19000	3.6289	0.3601
3.5438	5.8293	20000	3.6187	0.3609
3.4466	6.1207	21000	3.6228	0.3616
3.4752	6.4121	22000	3.6154	0.3620
3.4926	6.7036	23000	3.6044	0.3631
3.4953	6.9951	24000	3.5933	0.3639
3.4306	7.2865	25000	3.6031	0.3634
3.4505	7.5780	26000	3.5947	0.3643
3.4467	7.8695	27000	3.5850	0.3653
3.3792	8.1609	28000	3.5925	0.3650
3.4271	8.4524	29000	3.5885	0.3654
3.4277	8.7438	30000	3.5767	0.3663
3.3345	9.0353	31000	3.5829	0.3663
3.3772	9.3267	32000	3.5818	0.3666
3.4007	9.6182	33000	3.5740	0.3674
3.4129	9.9097	34000	3.5649	0.3679
3.3488	10.2011	35000	3.5771	0.3675
3.3752	10.4926	36000	3.5713	0.3679
3.4017	10.7841	37000	3.5627	0.3685
3.2817	11.0755	38000	3.5726	0.3682
3.3563	11.3670	39000	3.5677	0.3686
3.3707	11.6584	40000	3.5597	0.3690
3.3783	11.9499	41000	3.5514	0.3694
3.3043	12.2413	42000	3.5687	0.3688
3.3381	12.5328	43000	3.5612	0.3694
3.3436	12.8243	44000	3.5523	0.3700
3.2598	13.1157	45000	3.5681	0.3694
3.3126	13.4072	46000	3.5578	0.3698
3.3377	13.6987	47000	3.5518	0.3704
3.3456	13.9901	48000	3.5461	0.3706
3.2841	14.2816	49000	3.5617	0.3703
3.316	14.5730	50000	3.5507	0.3706
3.321	14.8645	51000	3.5442	0.3710
3.2468	15.1559	52000	3.5632	0.3702
3.2883	15.4474	53000	3.5542	0.3707
3.3014	15.7389	54000	3.5445	0.3712
3.2051	16.0303	55000	3.5582	0.3708
3.2714	16.3218	56000	3.5576	0.3710
3.2841	16.6133	57000	3.5493	0.3711
3.3063	16.9047	58000	3.5383	0.3719
3.2194	17.1962	59000	3.5591	0.3711
3.2508	17.4876	60000	3.5488	0.3718
3.2756	17.7791	61000	3.5422	0.3719
3.2002	18.0705	62000	3.5548	0.3715
3.238	18.3620	63000	3.5562	0.3714
3.2573	18.6535	64000	3.5457	0.3723
3.2814	18.9450	65000	3.5372	0.3728
3.2188	19.2364	66000	3.5508	0.3720
3.2464	19.5279	67000	3.5474	0.3722
3.2708	19.8193	68000	3.5366	0.3727
3.1871	20.1108	69000	3.5545	0.3720
3.2262	20.4022	70000	3.5519	0.3725
3.2309	20.6937	71000	3.5418	0.3727
3.2683	20.9852	72000	3.5362	0.3731
3.2	21.2766	73000	3.5545	0.3723
3.2195	21.5681	74000	3.5422	0.3731
3.2369	21.8596	75000	3.5387	0.3732
3.174	22.1510	76000	3.5572	0.3725
3.21	22.4425	77000	3.5461	0.3728
3.2286	22.7339	78000	3.5404	0.3733
3.1308	23.0254	79000	3.5521	0.3730
3.1869	23.3168	80000	3.5488	0.3730
3.2054	23.6083	81000	3.5411	0.3736
3.2129	23.8998	82000	3.5362	0.3740
3.1607	24.1912	83000	3.5607	0.3727
3.2039	24.4827	84000	3.5480	0.3732
3.2035	24.7742	85000	3.5417	0.3736
3.1279	25.0656	86000	3.5528	0.3731
3.1733	25.3571	87000	3.5499	0.3735
3.1876	25.6485	88000	3.5400	0.3740
3.21	25.9400	89000	3.5373	0.3740
3.1492	26.2314	90000	3.5567	0.3732
3.1763	26.5229	91000	3.5474	0.3736
3.2001	26.8144	92000	3.5379	0.3742
3.1198	27.1058	93000	3.5545	0.3735
3.1512	27.3973	94000	3.5536	0.3735
3.1749	27.6888	95000	3.5416	0.3740
3.1977	27.9802	96000	3.5375	0.3743
3.1372	28.2717	97000	3.5511	0.3737
3.1611	28.5631	98000	3.5464	0.3740
3.1694	28.8546	99000	3.5391	0.3744
3.1053	29.1460	100000	3.5543	0.3736
3.1358	29.4375	101000	3.5499	0.3738
3.1706	29.7290	102000	3.5407	0.3742

Framework versions

Transformers 4.55.2
Pytorch 2.8.0+cu128
Datasets 4.0.0
Tokenizers 0.21.4

craa/exceptions_exp2_swap_last_to_hit_3591

exceptions_exp2_swap_last_to_hit_3591

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

craa/
exceptions_exp2_swap_last_to_hit_3591