exceptions_exp2_swap_0.3_last_to_drop_5039

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

• Loss: 3.5637
• Accuracy: 0.3729

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	Accuracy	Validation Loss
4.844	0.2915	1000	0.2538	4.7601
4.3495	0.5830	2000	0.2986	4.2927
4.1551	0.8745	3000	0.3143	4.1093
3.9998	1.1659	4000	0.3239	3.9958
3.9319	1.4574	5000	0.3311	3.9190
3.8792	1.7488	6000	0.3364	3.8598
3.7431	2.0402	7000	0.3403	3.8187
3.7605	2.3317	8000	0.3436	3.7870
3.7396	2.6232	9000	0.3460	3.7570
3.7255	2.9147	10000	0.3487	3.7307
3.6305	3.2061	11000	0.3508	3.7204
3.6487	3.4976	12000	0.3524	3.7005
3.645	3.7891	13000	0.3538	3.6833
3.5469	4.0805	14000	0.3551	3.6754
3.5691	4.3719	15000	0.3563	3.6666
3.5799	4.6634	16000	0.3577	3.6527
3.58	4.9549	17000	0.3586	3.6404
3.5093	5.2463	18000	0.3593	3.6407
3.5156	5.5378	19000	0.3600	3.6298
3.5457	5.8293	20000	0.3609	3.6194
3.4538	6.1207	21000	0.3616	3.6228
3.4903	6.4122	22000	0.3622	3.6140
3.4974	6.7037	23000	0.3629	3.6066
3.4914	6.9952	24000	0.3639	3.5955
3.4382	7.2865	25000	0.3636	3.6061
3.4515	7.5780	26000	0.3643	3.5944
3.4607	7.8695	27000	0.3652	3.5857
3.3944	8.1609	28000	0.3649	3.5955
3.4335	8.4524	29000	0.3651	3.5917
3.4359	8.7439	30000	0.3661	3.5826
3.3406	9.0353	31000	0.3661	3.5865
3.3901	9.3268	32000	0.3662	3.5857
3.4008	9.6183	33000	0.3670	3.5784
3.4189	9.9098	34000	0.3673	3.5705
3.3465	10.2011	35000	0.3670	3.5792
3.3774	10.4926	36000	0.3675	3.5732
3.3979	10.7841	37000	0.3682	3.5661
3.3066	11.0755	38000	0.3681	3.5756
3.3449	11.3670	39000	0.3682	3.5711
3.3609	11.6585	40000	0.3688	3.5632
3.3863	11.9500	41000	0.3692	3.5563
3.3216	12.2414	42000	0.3689	3.5702
3.3446	12.5329	43000	0.3691	3.5624
3.3568	12.8243	44000	0.3697	3.5551
3.2827	13.1157	45000	0.3693	3.5678
3.3219	13.4072	46000	0.3695	3.5620
3.3323	13.6987	47000	0.3700	3.5527
3.3479	13.9902	48000	0.3704	3.5476
3.2759	14.2816	49000	0.3696	3.5653
3.3082	14.5731	50000	0.3701	3.5561
3.3429	14.8646	51000	0.3708	3.5464
3.2589	15.1559	52000	0.3701	3.5644
3.2771	15.4474	53000	0.3705	3.5592
3.3108	15.7389	54000	0.3711	3.5496
3.2005	16.0303	55000	0.3706	3.5599
3.2652	16.3218	56000	0.3707	3.5581
3.2774	16.6133	57000	0.3711	3.5528
3.3011	16.9048	58000	0.3718	3.5431
3.2297	17.1962	59000	0.3708	3.5613
3.2647	17.4877	60000	0.3714	3.5512
3.2735	17.7792	61000	0.3723	3.5429
3.1887	18.0705	62000	0.3712	3.5588
3.237	18.3620	63000	0.3712	3.5561
3.2584	18.6535	64000	0.3719	3.5469
3.2791	18.9450	65000	0.3726	3.5381
3.2059	19.2364	66000	0.3719	3.5552
3.2428	19.5279	67000	0.3719	3.5494
3.2512	19.8194	68000	0.3726	3.5434
3.1956	20.1108	69000	0.3719	3.5579
3.2246	20.4023	70000	0.3717	3.5530
3.2391	20.6938	71000	0.3722	3.5478
3.2655	20.9853	72000	0.3729	3.5408
3.217	21.2766	73000	0.3719	3.5548
3.2157	21.5681	74000	0.3724	3.5482
3.2408	21.8596	75000	0.3730	3.5432
3.1699	22.1510	76000	0.3723	3.5580
3.2108	22.4425	77000	0.3727	3.5529
3.2314	22.7340	78000	0.3730	3.5459
3.1423	23.0254	79000	0.3724	3.5562
3.1849	23.3169	80000	0.3725	3.5558
3.1659	23.6083	81000	3.5575	0.3726
3.1987	23.8998	82000	3.5504	0.3731
3.1785	24.1915	83000	3.5625	0.3724
3.1864	24.4830	84000	3.5524	0.3727
3.211	24.7745	85000	3.5439	0.3735
3.1256	25.0659	86000	3.5594	0.3727
3.1657	25.3574	87000	3.5546	0.3730
3.2027	25.6489	88000	3.5469	0.3737
3.2195	25.9404	89000	3.5392	0.3738
3.1591	26.2317	90000	3.5588	0.3730
3.1871	26.5232	91000	3.5502	0.3733
3.1886	26.8147	92000	3.5429	0.3737
3.1272	27.1061	93000	3.5568	0.3730
3.1554	27.3976	94000	3.5534	0.3731
3.1832	27.6891	95000	3.5500	0.3735
3.1958	27.9806	96000	3.5399	0.3741
3.1378	28.2720	97000	3.5549	0.3732
3.154	28.5635	98000	3.5492	0.3738
3.1631	28.8550	99000	3.5431	0.3740
3.1171	29.1463	100000	3.5637	0.3729

Framework versions

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