exceptions_exp2_swap_0.7_last_to_push_1032

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

• Loss: 3.5644
• Accuracy: 0.3686

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
4.8199	0.2915	1000	0.2557	4.7489
4.3394	0.5830	2000	0.2985	4.2872
4.152	0.8745	3000	0.3142	4.1027
3.998	1.1659	4000	0.3234	3.9996
3.9535	1.4574	5000	0.3306	3.9238
3.8885	1.7489	6000	0.3357	3.8655
3.7583	2.0402	7000	0.3400	3.8238
3.7615	2.3317	8000	0.3430	3.7939
3.7454	2.6233	9000	0.3456	3.7623
3.7305	2.9148	10000	0.3484	3.7364
3.6492	3.2061	11000	0.3502	3.7243
3.65	3.4976	12000	0.3517	3.7054
3.6525	3.7891	13000	0.3534	3.6861
3.5452	4.0805	14000	0.3549	3.6810
3.5745	4.3720	15000	0.3562	3.6666
3.5906	4.6635	16000	0.3571	3.6557
3.5889	4.9550	17000	0.3584	3.6405
3.5194	5.2463	18000	0.3590	3.6444
3.5288	5.5378	19000	0.3600	3.6323
3.5341	5.8293	20000	0.3609	3.6224
3.4369	6.1207	21000	0.3612	3.6253
3.4869	6.4122	22000	0.3621	3.6204
3.4957	6.7037	23000	0.3627	3.6091
3.5051	6.9952	24000	0.3635	3.6000
3.4353	7.2866	25000	0.3633	3.6091
3.4666	7.5781	26000	0.3642	3.5999
3.4623	7.8696	27000	0.3650	3.5864
3.3899	8.1609	28000	0.3647	3.5995
3.4279	8.4524	29000	0.3652	3.5901
3.434	8.7439	30000	0.3659	3.5819
3.3392	9.0353	31000	0.3660	3.5891
3.3891	9.3268	32000	0.3659	3.5865
3.3997	9.6183	33000	0.3668	3.5792
3.4287	9.9098	34000	0.3676	3.5705
3.3463	10.2011	35000	0.3673	3.5800
3.3681	10.4927	36000	0.3674	3.5784
3.4045	10.7842	37000	0.3680	3.5678
3.2969	11.0755	38000	0.3678	3.5783
3.3605	11.3670	39000	0.3681	3.5736
3.3652	11.6585	40000	0.3686	3.5644
3.3896	11.9500	41000	0.3690	3.5596
3.3144	12.2414	42000	0.3682	3.5734
3.3353	12.5329	43000	0.3688	3.5665
3.3535	12.8244	44000	0.3695	3.5587
3.272	13.1157	45000	0.3693	3.5701
3.3143	13.4072	46000	0.3690	3.5667
3.3453	13.6988	47000	0.3699	3.5564
3.355	13.9903	48000	0.3706	3.5493
3.2878	14.2816	49000	0.3700	3.5643
3.3134	14.5731	50000	0.3702	3.5556
3.3318	14.8646	51000	0.3706	3.5504
3.2652	15.1560	52000	0.3700	3.5669
3.2886	15.4475	53000	0.3706	3.5605
3.3238	15.7390	54000	0.3711	3.5493
3.2186	16.0303	55000	0.3705	3.5647
3.2666	16.3218	56000	0.3708	3.5616
3.2879	16.6133	57000	0.3713	3.5543
3.3007	16.9049	58000	0.3717	3.5442
3.234	17.1962	59000	0.3706	3.5637
3.2701	17.4877	60000	0.3714	3.5541
3.2802	17.7792	61000	0.3720	3.5504
3.1942	18.0705	62000	0.3707	3.5671
3.2393	18.3621	63000	0.3711	3.5591
3.2598	18.6536	64000	0.3718	3.5515
3.271	18.9451	65000	0.3721	3.5439
3.2108	19.2364	66000	0.3715	3.5616
3.256	19.5279	67000	0.3719	3.5521
3.2692	19.8194	68000	0.3723	3.5431
3.1916	20.1108	69000	0.3715	3.5643
3.2157	20.4023	70000	0.3720	3.5574
3.2554	20.6938	71000	0.3724	3.5460
3.2615	20.9853	72000	0.3726	3.5405
3.2016	21.2766	73000	0.3719	3.5635
3.2166	21.5682	74000	0.3722	3.5502
3.243	21.8597	75000	0.3728	3.5441
3.1716	22.1510	76000	0.3721	3.5593
3.1998	22.4425	77000	0.3724	3.5544
3.2394	22.7340	78000	0.3729	3.5470
3.1309	23.0254	79000	0.3722	3.5588
3.1858	23.3169	80000	0.3722	3.5583
3.1992	23.6084	81000	3.5650	0.3722
3.2197	23.8999	82000	3.5497	0.3726
3.1587	24.1915	83000	3.5615	0.3723
3.1985	24.4830	84000	3.5550	0.3726
3.2164	24.7745	85000	3.5455	0.3733
3.1432	25.0659	86000	3.5623	0.3723
3.1808	25.3574	87000	3.5566	0.3728
3.2007	25.6489	88000	3.5481	0.3731
3.205	25.9404	89000	3.5415	0.3737
3.1654	26.2318	90000	3.5589	0.3727
3.1871	26.5233	91000	3.5475	0.3732
3.1968	26.8148	92000	3.5477	0.3735
3.1321	27.1061	93000	3.5612	0.3728
3.1537	27.3976	94000	3.5578	0.3731
3.1954	27.6891	95000	3.5478	0.3735
3.2019	27.9806	96000	3.5409	0.3739
3.1278	28.2720	97000	3.5578	0.3732
3.1623	28.5635	98000	3.5488	0.3737
3.1833	28.8550	99000	3.5416	0.3744
3.1149	29.1463	100000	3.5644	0.3731

Framework versions

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