exceptions_exp2_swap_0.3_resemble_to_carry_1032

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

• Loss: 3.5679
• Accuracy: 0.3680

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	Validation Loss	Accuracy
4.8421	0.2915	1000	4.7622	0.2537
4.351	0.5830	2000	4.2958	0.2972
4.1647	0.8745	3000	4.1061	0.3140
3.9922	1.1659	4000	4.0050	0.3228
3.9458	1.4574	5000	3.9261	0.3301
3.9024	1.7488	6000	3.8691	0.3356
3.7625	2.0402	7000	3.8261	0.3399
3.76	2.3317	8000	3.7980	0.3425
3.7595	2.6232	9000	3.7659	0.3454
3.7362	2.9147	10000	3.7422	0.3482
3.6473	3.2061	11000	3.7267	0.3500
3.6615	3.4976	12000	3.7076	0.3516
3.6616	3.7891	13000	3.6918	0.3532
3.5536	4.0805	14000	3.6852	0.3543
3.5772	4.3719	15000	3.6746	0.3552
3.5823	4.6634	16000	3.6603	0.3568
3.5882	4.9549	17000	3.6488	0.3579
3.5332	5.2463	18000	3.6485	0.3584
3.5385	5.5378	19000	3.6355	0.3595
3.5465	5.8293	20000	3.6257	0.3604
3.4535	6.1207	21000	3.6306	0.3609
3.4857	6.4122	22000	3.6228	0.3613
3.4948	6.7037	23000	3.6119	0.3619
3.5026	6.9952	24000	3.6045	0.3630
3.4461	7.2865	25000	3.6139	0.3626
3.4777	7.5780	26000	3.6036	0.3635
3.47	7.8695	27000	3.5940	0.3641
3.3878	8.1609	28000	3.6021	0.3640
3.4226	8.4524	29000	3.5953	0.3648
3.4401	8.7439	30000	3.5858	0.3651
3.3421	9.0353	31000	3.5925	0.3654
3.3924	9.3268	32000	3.5909	0.3657
3.407	9.6183	33000	3.5823	0.3660
3.4309	9.9098	34000	3.5736	0.3667
3.3439	10.2011	35000	3.5858	0.3663
3.384	10.4926	36000	3.5802	0.3668
3.4041	10.7841	37000	3.5720	0.3673
3.3012	11.0755	38000	3.5814	0.3674
3.3628	11.3670	39000	3.5814	0.3671
3.3785	11.6585	40000	3.5679	0.3680
3.3792	11.9500	41000	3.5610	0.3683
3.3091	12.2414	42000	3.5781	0.3678
3.3372	12.5329	43000	3.5688	0.3685
3.3697	12.8243	44000	3.5607	0.3692
3.2652	13.1157	45000	3.5765	0.3685
3.3184	13.4072	46000	3.5694	0.3687
3.3471	13.6987	47000	3.5620	0.3693
3.3563	13.9902	48000	3.5529	0.3697
3.3027	14.2816	49000	3.5678	0.3692
3.3122	14.5731	50000	3.5612	0.3695
3.3345	14.8646	51000	3.5553	0.3699
3.2818	15.1559	52000	3.5676	0.3694
3.291	15.4474	53000	3.5633	0.3699
3.3165	15.7389	54000	3.5556	0.3704
3.2157	16.0303	55000	3.5641	0.3703
3.2672	16.3218	56000	3.5646	0.3700
3.295	16.6133	57000	3.5535	0.3704
3.313	16.9048	58000	3.5489	0.3711
3.2262	17.1962	59000	3.5674	0.3703
3.2764	17.4877	60000	3.5588	0.3706
3.2842	17.7792	61000	3.5504	0.3715
3.2146	18.0705	62000	3.5637	0.3708
3.2502	18.3620	63000	3.5584	0.3708
3.2715	18.6535	64000	3.5525	0.3713
3.2926	18.9450	65000	3.5440	0.3716
3.2279	19.2364	66000	3.5632	0.3711
3.2537	19.5279	67000	3.5535	0.3717
3.2813	19.8194	68000	3.5452	0.3717
3.2086	20.1108	69000	3.5654	0.3710
3.2297	20.4023	70000	3.5586	0.3716
3.2479	20.6938	71000	3.5473	0.3719
3.2615	20.9853	72000	3.5412	0.3722
3.2009	21.2766	73000	3.5620	0.3714
3.2455	21.5681	74000	3.5530	0.3717
3.2481	21.8596	75000	3.5443	0.3725
3.1822	22.1510	76000	3.5620	0.3716
3.2181	22.4425	77000	3.5539	0.3721
3.2393	22.7340	78000	3.5481	0.3722
3.1498	23.0254	79000	3.5591	0.3718
3.1997	23.3169	80000	3.5581	0.3718
3.2237	23.6083	81000	3.5510	0.3725
3.2372	23.8998	82000	3.5435	0.3730
3.1724	24.1912	83000	3.5590	0.3719
3.1946	24.4827	84000	3.5525	0.3727
3.2219	24.7742	85000	3.5479	0.3729
3.1396	25.0656	86000	3.5610	0.3724
3.1926	25.3571	87000	3.5571	0.3722
3.2133	25.6486	88000	3.5530	0.3728
3.2397	25.9401	89000	3.5443	0.3731
3.1539	26.2314	90000	3.5618	0.3725
3.1987	26.5229	91000	3.5529	0.3729
3.1964	26.8144	92000	3.5463	0.3734

Framework versions

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