exceptions_exp2_swap_0.3_cost_to_hit_2128

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

• Loss: 3.5869
• Accuracy: 0.3658

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: 2128
• 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
4.8577	0.2915	1000	4.7865	0.2499
4.3688	0.5831	2000	4.3115	0.2963
4.1518	0.8746	3000	4.1113	0.3134
4.0023	1.1662	4000	4.0043	0.3232
3.9528	1.4577	5000	3.9268	0.3302
3.896	1.7493	6000	3.8698	0.3357
3.7661	2.0408	7000	3.8236	0.3401
3.763	2.3324	8000	3.7941	0.3429
3.7367	2.6239	9000	3.7634	0.3457
3.7362	2.9155	10000	3.7357	0.3481
3.6461	3.2070	11000	3.7256	0.3497
3.6601	3.4985	12000	3.7075	0.3514
3.6589	3.7901	13000	3.6882	0.3535
3.5515	4.0816	14000	3.6810	0.3549
3.5739	4.3732	15000	3.6697	0.3558
3.5846	4.6647	16000	3.6576	0.3569
3.5828	4.9563	17000	3.6435	0.3582
3.5042	5.2478	18000	3.6474	0.3589
3.5314	5.5394	19000	3.6348	0.3598
3.5384	5.8309	20000	3.6236	0.3607
3.4452	6.1224	21000	3.6262	0.3606
3.4801	6.4140	22000	3.6221	0.3615
3.504	6.7055	23000	3.6095	0.3624
3.5034	6.9971	24000	3.6003	0.3631
3.4405	7.2886	25000	3.6093	0.3632
3.469	7.5802	26000	3.5990	0.3641
3.466	7.8717	27000	3.5924	0.3648
3.3821	8.1633	28000	3.6006	0.3645
3.4233	8.4548	29000	3.5922	0.3654
3.4214	8.7464	30000	3.5869	0.3658
3.3404	9.0379	31000	3.5900	0.3658
3.3875	9.3294	32000	3.5873	0.3657
3.4062	9.6210	33000	3.5791	0.3666
3.4131	9.9125	34000	3.5721	0.3670
3.3526	10.2041	35000	3.5822	0.3669
3.3653	10.4956	36000	3.5747	0.3674
3.3866	10.7872	37000	3.5640	0.3681
3.311	11.0787	38000	3.5761	0.3676
3.3562	11.3703	39000	3.5761	0.3678
3.3653	11.6618	40000	3.5677	0.3683
3.3683	11.9534	41000	3.5553	0.3690
3.3254	12.2449	42000	3.5713	0.3687
3.3447	12.5364	43000	3.5648	0.3689
3.3614	12.8280	44000	3.5551	0.3692
3.2762	13.1195	45000	3.5709	0.3690
3.3092	13.4111	46000	3.5662	0.3691
3.3317	13.7026	47000	3.5579	0.3699
3.3512	13.9942	48000	3.5488	0.3702
3.3039	14.2857	49000	3.5654	0.3697
3.3128	14.5773	50000	3.5592	0.3700
3.3292	14.8688	51000	3.5495	0.3707
3.2652	15.1603	52000	3.5653	0.3701
3.2876	15.4519	53000	3.5595	0.3703
3.3246	15.7434	54000	3.5502	0.3710
3.217	16.0350	55000	3.5615	0.3705
3.2564	16.3265	56000	3.5590	0.3707
3.2843	16.6181	57000	3.5543	0.3709
3.2962	16.9096	58000	3.5463	0.3714
3.2252	17.2012	59000	3.5604	0.3707
3.2623	17.4927	60000	3.5543	0.3711
3.2851	17.7843	61000	3.5469	0.3719
3.2008	18.0758	62000	3.5635	0.3708
3.2415	18.3673	63000	3.5552	0.3715
3.2789	18.6589	64000	3.5461	0.3717
3.2861	18.9504	65000	3.5444	0.3720
3.2243	19.2420	66000	3.5592	0.3714
3.2582	19.5335	67000	3.5547	0.3718
3.2691	19.8251	68000	3.5432	0.3726
3.1809	20.1166	69000	3.5597	0.3716
3.2223	20.4082	70000	3.5548	0.3721
3.2442	20.6997	71000	3.5455	0.3723
3.2531	20.9913	72000	3.5378	0.3728
3.2093	21.2828	73000	3.5569	0.3720
3.2377	21.5743	74000	3.5463	0.3724
3.2261	21.8659	75000	3.5394	0.3727
3.1731	22.1574	76000	3.5594	0.3719
3.2154	22.4490	77000	3.5518	0.3724
3.2484	22.7405	78000	3.5413	0.3728
3.1377	23.0321	79000	3.5568	0.3723
3.1774	23.3236	80000	3.5564	0.3725
3.2135	23.6152	81000	3.5461	0.3728
3.2283	23.9067	82000	3.5397	0.3731
3.1617	24.1983	83000	3.5572	0.3725
3.1902	24.4898	84000	3.5499	0.3727
3.2133	24.7813	85000	3.5429	0.3735
3.138	25.0729	86000	3.5567	0.3730
3.1705	25.3644	87000	3.5544	0.3729
3.1969	25.6560	88000	3.5516	0.3731
3.1931	25.9475	89000	3.5386	0.3737
3.1508	26.2391	90000	3.5576	0.3729
3.1739	26.5306	91000	3.5491	0.3734
3.1847	26.8222	92000	3.5425	0.3738

Framework versions

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