exceptions_exp2_swap_0.3_resemble_to_carry_3591

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

• Loss: 3.5627
• 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: 3591
• 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
4.8611	0.2915	1000	4.7759	0.2515
4.3524	0.5830	2000	4.2887	0.2982
4.1388	0.8745	3000	4.0997	0.3149
3.9954	1.1659	4000	3.9942	0.3244
3.9381	1.4574	5000	3.9175	0.3313
3.8796	1.7488	6000	3.8619	0.3364
3.74	2.0402	7000	3.8180	0.3409
3.7535	2.3317	8000	3.7894	0.3437
3.7476	2.6232	9000	3.7562	0.3464
3.7247	2.9147	10000	3.7313	0.3487
3.6212	3.2061	11000	3.7204	0.3508
3.6546	3.4976	12000	3.7027	0.3522
3.6366	3.7891	13000	3.6812	0.3541
3.539	4.0805	14000	3.6769	0.3553
3.5748	4.3719	15000	3.6634	0.3563
3.5716	4.6634	16000	3.6501	0.3576
3.5806	4.9549	17000	3.6361	0.3589
3.5099	5.2463	18000	3.6429	0.3595
3.535	5.5378	19000	3.6311	0.3601
3.512	5.8293	20000	3.6188	0.3613
3.4448	6.1207	21000	3.6224	0.3615
3.475	6.4122	22000	3.6144	0.3623
3.5005	6.7037	23000	3.6068	0.3628
3.5028	6.9952	24000	3.5979	0.3639
3.443	7.2865	25000	3.6031	0.3638
3.453	7.5780	26000	3.5940	0.3647
3.4619	7.8695	27000	3.5868	0.3650
3.4049	8.1609	28000	3.5938	0.3652
3.4264	8.4524	29000	3.5892	0.3652
3.4498	8.7439	30000	3.5790	0.3660
3.3344	9.0353	31000	3.5842	0.3663
3.3909	9.3268	32000	3.5841	0.3664
3.4037	9.6183	33000	3.5758	0.3669
3.4071	9.9098	34000	3.5660	0.3677
3.3266	10.2011	35000	3.5796	0.3672
3.3795	10.4926	36000	3.5754	0.3676
3.3768	10.7841	37000	3.5650	0.3682
3.3094	11.0755	38000	3.5742	0.3679
3.3421	11.3670	39000	3.5725	0.3681
3.3665	11.6585	40000	3.5627	0.3686
3.379	11.9500	41000	3.5567	0.3692
3.2988	12.2414	42000	3.5693	0.3690
3.346	12.5329	43000	3.5595	0.3694
3.3564	12.8243	44000	3.5514	0.3701
3.2784	13.1157	45000	3.5658	0.3694
3.3288	13.4072	46000	3.5639	0.3696
3.3299	13.6987	47000	3.5541	0.3700
3.3462	13.9902	48000	3.5466	0.3708
3.2894	14.2816	49000	3.5607	0.3700
3.3081	14.5731	50000	3.5544	0.3706
3.327	14.8646	51000	3.5506	0.3709
3.2482	15.1559	52000	3.5612	0.3704
3.2802	15.4474	53000	3.5576	0.3704
3.3049	15.7389	54000	3.5509	0.3709
3.2122	16.0303	55000	3.5596	0.3706
3.2658	16.3218	56000	3.5581	0.3710
3.2865	16.6133	57000	3.5490	0.3713
3.2986	16.9048	58000	3.5423	0.3717
3.2354	17.1962	59000	3.5566	0.3712
3.2642	17.4877	60000	3.5545	0.3714
3.2777	17.7792	61000	3.5483	0.3716
3.2038	18.0705	62000	3.5558	0.3715
3.2268	18.3620	63000	3.5541	0.3715
3.2634	18.6535	64000	3.5468	0.3721
3.2768	18.9450	65000	3.5386	0.3724
3.2221	19.2364	66000	3.5558	0.3717
3.243	19.5279	67000	3.5519	0.3722
3.2603	19.8194	68000	3.5451	0.3723
3.1875	20.1108	69000	3.5612	0.3717
3.2295	20.4023	70000	3.5522	0.3721
3.25	20.6938	71000	3.5442	0.3725
3.2594	20.9853	72000	3.5358	0.3734
3.2081	21.2766	73000	3.5549	0.3719
3.2195	21.5681	74000	3.5473	0.3726
3.2397	21.8596	75000	3.5381	0.3732
3.1756	22.1510	76000	3.5561	0.3721
3.2106	22.4425	77000	3.5478	0.3727
3.2176	22.7340	78000	3.5441	0.3728
3.1459	23.0254	79000	3.5531	0.3725
3.1883	23.3169	80000	3.5532	0.3726
3.2169	23.6083	81000	3.5488	0.3728
3.2302	23.8998	82000	3.5391	0.3733
3.1652	24.1912	83000	3.5565	0.3726
3.1966	24.4827	84000	3.5498	0.3731
3.2087	24.7742	85000	3.5395	0.3737
3.1184	25.0656	86000	3.5565	0.3727
3.1774	25.3571	87000	3.5512	0.3732
3.1971	25.6486	88000	3.5430	0.3735
3.2131	25.9401	89000	3.5372	0.3739
3.15	26.2314	90000	3.5535	0.3731
3.1785	26.5229	91000	3.5473	0.3733
3.183	26.8144	92000	3.5399	0.3740

Framework versions

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