exceptions_exp2_swap_0.3_resemble_to_push_3591

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

• Loss: 3.5638
• Accuracy: 0.3684

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
4.8456	0.2915	1000	4.7632	0.2537
4.3515	0.5830	2000	4.2880	0.2985
4.1415	0.8745	3000	4.1011	0.3145
3.9964	1.1659	4000	3.9946	0.3243
3.9395	1.4574	5000	3.9203	0.3310
3.8823	1.7488	6000	3.8624	0.3360
3.7414	2.0402	7000	3.8180	0.3406
3.7534	2.3317	8000	3.7893	0.3436
3.7482	2.6232	9000	3.7577	0.3463
3.7257	2.9147	10000	3.7313	0.3488
3.623	3.2061	11000	3.7211	0.3506
3.6538	3.4976	12000	3.7016	0.3524
3.637	3.7891	13000	3.6805	0.3542
3.5399	4.0805	14000	3.6761	0.3555
3.5753	4.3719	15000	3.6643	0.3562
3.5721	4.6634	16000	3.6509	0.3577
3.5806	4.9549	17000	3.6367	0.3589
3.5109	5.2463	18000	3.6407	0.3594
3.5357	5.5378	19000	3.6300	0.3600
3.5129	5.8293	20000	3.6196	0.3611
3.4459	6.1207	21000	3.6218	0.3615
3.4753	6.4122	22000	3.6140	0.3623
3.5009	6.7037	23000	3.6081	0.3628
3.5039	6.9952	24000	3.5967	0.3640
3.4436	7.2865	25000	3.6063	0.3638
3.4535	7.5780	26000	3.5947	0.3644
3.4635	7.8695	27000	3.5890	0.3650
3.4052	8.1609	28000	3.5960	0.3652
3.4284	8.4524	29000	3.5885	0.3654
3.45	8.7439	30000	3.5783	0.3663
3.3362	9.0353	31000	3.5852	0.3662
3.3917	9.3268	32000	3.5855	0.3661
3.4043	9.6183	33000	3.5772	0.3669
3.4076	9.9098	34000	3.5673	0.3676
3.3278	10.2011	35000	3.5792	0.3671
3.3805	10.4926	36000	3.5740	0.3679
3.3777	10.7841	37000	3.5666	0.3681
3.3095	11.0755	38000	3.5736	0.3678
3.343	11.3670	39000	3.5730	0.3681
3.3675	11.6585	40000	3.5638	0.3684
3.3795	11.9500	41000	3.5558	0.3692
3.2994	12.2414	42000	3.5705	0.3688
3.3445	12.5329	43000	3.5615	0.3694
3.357	12.8243	44000	3.5531	0.3702
3.2793	13.1157	45000	3.5676	0.3693
3.327	13.4072	46000	3.5620	0.3698
3.3319	13.6987	47000	3.5525	0.3700
3.3461	13.9902	48000	3.5467	0.3707
3.2891	14.2816	49000	3.5616	0.3702
3.3087	14.5731	50000	3.5573	0.3702
3.3283	14.8646	51000	3.5486	0.3709
3.2493	15.1559	52000	3.5638	0.3704
3.2813	15.4474	53000	3.5574	0.3706
3.3062	15.7389	54000	3.5498	0.3708
3.2134	16.0303	55000	3.5585	0.3706
3.2661	16.3218	56000	3.5572	0.3710
3.2874	16.6133	57000	3.5512	0.3715
3.2994	16.9048	58000	3.5412	0.3718
3.2365	17.1962	59000	3.5584	0.3712
3.264	17.4877	60000	3.5513	0.3716
3.2788	17.7792	61000	3.5473	0.3718
3.2045	18.0705	62000	3.5568	0.3715
3.2283	18.3620	63000	3.5566	0.3712
3.2642	18.6535	64000	3.5466	0.3719
3.2771	18.9450	65000	3.5403	0.3724
3.2224	19.2364	66000	3.5570	0.3716
3.2433	19.5279	67000	3.5476	0.3720
3.2618	19.8194	68000	3.5449	0.3723
3.1872	20.1108	69000	3.5594	0.3719
3.2301	20.4023	70000	3.5535	0.3721
3.2507	20.6938	71000	3.5458	0.3725
3.2607	20.9853	72000	3.5355	0.3732
3.2097	21.2766	73000	3.5544	0.3721
3.2209	21.5681	74000	3.5488	0.3725
3.2395	21.8596	75000	3.5400	0.3731
3.1757	22.1510	76000	3.5577	0.3718
3.2105	22.4425	77000	3.5517	0.3729
3.2181	22.7340	78000	3.5448	0.3732
3.145	23.0254	79000	3.5561	0.3724
3.1886	23.3169	80000	3.5567	0.3725
3.2171	23.6083	81000	3.5480	0.3730
3.2298	23.8998	82000	3.5412	0.3735
3.1648	24.1912	83000	3.5565	0.3726
3.196	24.4827	84000	3.5495	0.3732
3.2086	24.7742	85000	3.5433	0.3735
3.1187	25.0656	86000	3.5575	0.3728
3.1785	25.3571	87000	3.5511	0.3732
3.198	25.6486	88000	3.5422	0.3735
3.2131	25.9401	89000	3.5396	0.3739
3.1507	26.2314	90000	3.5545	0.3731
3.1798	26.5229	91000	3.5481	0.3733
3.1843	26.8144	92000	3.5411	0.3740

Framework versions

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