exceptions_exp2_swap_0.3_last_to_drop_1032

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

• Loss: 3.5508
• Accuracy: 0.3741

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: 1032
• 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
4.8235	0.2915	1000	0.2544	4.7551
4.3553	0.5830	2000	0.2984	4.2897
4.1484	0.8745	3000	0.3139	4.1042
4.0041	1.1659	4000	0.3239	3.9982
3.9468	1.4574	5000	0.3306	3.9197
3.8875	1.7488	6000	0.3360	3.8670
3.7501	2.0402	7000	0.3403	3.8202
3.7497	2.3317	8000	0.3434	3.7905
3.7424	2.6232	9000	0.3458	3.7606
3.729	2.9147	10000	0.3485	3.7343
3.6452	3.2061	11000	0.3502	3.7250
3.6466	3.4976	12000	0.3519	3.7026
3.6575	3.7891	13000	0.3535	3.6850
3.5567	4.0805	14000	0.3548	3.6800
3.5675	4.3719	15000	0.3561	3.6665
3.5869	4.6634	16000	0.3568	3.6547
3.5789	4.9549	17000	0.3582	3.6421
3.5123	5.2463	18000	0.3591	3.6440
3.512	5.5378	19000	0.3600	3.6330
3.5442	5.8293	20000	0.3609	3.6218
3.4481	6.1207	21000	0.3610	3.6227
3.4852	6.4122	22000	0.3617	3.6196
3.4867	6.7037	23000	0.3629	3.6085
3.5079	6.9952	24000	0.3632	3.6005
3.4287	7.2865	25000	0.3628	3.6080
3.4631	7.5780	26000	0.3641	3.5972
3.468	7.8695	27000	0.3646	3.5895
3.3731	8.1609	28000	0.3646	3.6000
3.4144	8.4524	29000	0.3654	3.5911
3.4449	8.7439	30000	0.3654	3.5835
3.3304	9.0353	31000	0.3658	3.5865
3.3887	9.3268	32000	0.3660	3.5869
3.3909	9.6183	33000	0.3667	3.5774
3.4112	9.9098	34000	0.3672	3.5700
3.3481	10.2011	35000	0.3669	3.5828
3.3647	10.4926	36000	0.3677	3.5744
3.3856	10.7841	37000	0.3680	3.5664
3.3035	11.0755	38000	0.3679	3.5784
3.345	11.3670	39000	0.3680	3.5718
3.3675	11.6585	40000	0.3688	3.5642
3.3699	11.9500	41000	0.3690	3.5599
3.3244	12.2414	42000	0.3686	3.5710
3.3428	12.5329	43000	0.3691	3.5675
3.3582	12.8243	44000	0.3697	3.5583
3.281	13.1157	45000	0.3687	3.5718
3.3096	13.4072	46000	0.3696	3.5636
3.3312	13.6987	47000	0.3698	3.5557
3.3427	13.9902	48000	0.3706	3.5487
3.2845	14.2816	49000	0.3697	3.5665
3.3226	14.5731	50000	0.3702	3.5575
3.3355	14.8646	51000	0.3706	3.5465
3.258	15.1559	52000	0.3701	3.5647
3.2921	15.4474	53000	0.3703	3.5605
3.3033	15.7389	54000	0.3708	3.5486
3.2211	16.0303	55000	0.3706	3.5614
3.2648	16.3218	56000	0.3708	3.5589
3.2845	16.6133	57000	0.3710	3.5573
3.3076	16.9048	58000	0.3716	3.5420
3.2284	17.1962	59000	0.3708	3.5640
3.2602	17.4877	60000	0.3710	3.5527
3.2906	17.7792	61000	0.3720	3.5479
3.2018	18.0705	62000	0.3713	3.5612
3.2406	18.3620	63000	0.3714	3.5581
3.2617	18.6535	64000	0.3720	3.5489
3.2833	18.9450	65000	0.3723	3.5409
3.2291	19.2364	66000	0.3710	3.5608
3.2556	19.5279	67000	0.3718	3.5511
3.2603	19.8194	68000	0.3723	3.5435
3.1786	20.1108	69000	0.3714	3.5614
3.2258	20.4023	70000	0.3718	3.5557
3.2437	20.6938	71000	0.3724	3.5454
3.2571	20.9853	72000	0.3726	3.5407
3.2064	21.2766	73000	0.3720	3.5581
3.2243	21.5681	74000	0.3722	3.5497
3.247	21.8596	75000	0.3732	3.5432
3.1678	22.1510	76000	0.3722	3.5580
3.2098	22.4425	77000	0.3727	3.5515
3.2263	22.7340	78000	0.3727	3.5462
3.137	23.0254	79000	0.3725	3.5595
3.1979	23.3169	80000	0.3724	3.5566
3.1956	23.6083	81000	3.5573	0.3724
3.2112	23.8998	82000	3.5530	0.3726
3.1659	24.1915	83000	3.5616	0.3724
3.1969	24.4830	84000	3.5556	0.3726
3.2202	24.7745	85000	3.5448	0.3730
3.1451	25.0659	86000	3.5617	0.3724
3.1759	25.3574	87000	3.5573	0.3726
3.1946	25.6489	88000	3.5505	0.3730
3.2097	25.9404	89000	3.5389	0.3739
3.1585	26.2317	90000	3.5570	0.3728
3.1781	26.5232	91000	3.5499	0.3732
3.2103	26.8147	92000	3.5439	0.3736
3.121	27.1061	93000	3.5603	0.3727
3.1601	27.3976	94000	3.5555	0.3731
3.1809	27.6891	95000	3.5464	0.3735
3.1921	27.9806	96000	3.5419	0.3740
3.1305	28.2720	97000	3.5572	0.3730
3.1726	28.5635	98000	3.5512	0.3734
3.1832	28.8550	99000	3.5419	0.3743
3.1176	29.1463	100000	3.5623	0.3733
3.1431	29.4378	101000	3.5572	0.3734
3.1734	29.7293	102000	3.5512	0.3736
3.0923	30.0207	103000	3.5627	0.3732
3.1347	30.3122	104000	3.5604	0.3733
3.1553	30.6037	105000	3.5488	0.3742
3.1699	30.8952	106000	3.5426	0.3745
3.1223	31.1866	107000	3.5599	0.3734
3.1372	31.4781	108000	3.5551	0.3739
3.153	31.7695	109000	3.5508	0.3741

Framework versions

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