exceptions_exp2_swap_0.7_last_to_drop_40817

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

• Loss: 3.5645
• 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: 40817
• 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
4.8345	0.2915	1000	0.2522	4.7701
4.3512	0.5830	2000	0.2990	4.2900
4.1533	0.8745	3000	0.3146	4.1023
4.0173	1.1659	4000	0.3243	3.9956
3.9437	1.4574	5000	0.3312	3.9228
3.8689	1.7489	6000	0.3361	3.8647
3.7489	2.0402	7000	0.3405	3.8199
3.7471	2.3317	8000	0.3434	3.7898
3.7405	2.6233	9000	0.3459	3.7604
3.735	2.9148	10000	0.3488	3.7319
3.6385	3.2061	11000	0.3505	3.7230
3.642	3.4976	12000	0.3522	3.7026
3.6426	3.7891	13000	0.3540	3.6841
3.5402	4.0805	14000	0.3551	3.6770
3.5763	4.3720	15000	0.3563	3.6650
3.5938	4.6635	16000	0.3575	3.6533
3.5962	4.9550	17000	0.3590	3.6389
3.504	5.2463	18000	0.3592	3.6414
3.5211	5.5378	19000	0.3601	3.6316
3.5471	5.8293	20000	0.3610	3.6205
3.4456	6.1207	21000	0.3618	3.6224
3.4751	6.4122	22000	0.3620	3.6152
3.4816	6.7037	23000	0.3629	3.6071
3.4942	6.9952	24000	0.3637	3.5960
3.4289	7.2866	25000	0.3637	3.6038
3.4503	7.5781	26000	0.3644	3.5971
3.4669	7.8696	27000	0.3653	3.5867
3.3922	8.1609	28000	0.3649	3.5937
3.4176	8.4524	29000	0.3655	3.5879
3.4341	8.7439	30000	0.3661	3.5808
3.3273	9.0353	31000	0.3662	3.5866
3.3879	9.3268	32000	0.3663	3.5840
3.4128	9.6183	33000	0.3668	3.5762
3.4164	9.9098	34000	0.3676	3.5688
3.3424	10.2011	35000	0.3673	3.5790
3.3844	10.4927	36000	0.3673	3.5735
3.3913	10.7842	37000	0.3683	3.5659
3.2964	11.0755	38000	0.3680	3.5758
3.3432	11.3670	39000	0.3681	3.5749
3.3693	11.6585	40000	0.3686	3.5645
3.3762	11.9500	41000	0.3694	3.5550
3.3334	12.2414	42000	0.3684	3.5708
3.3411	12.5329	43000	0.3693	3.5627
3.3449	12.8244	44000	0.3697	3.5568
3.2707	13.1157	45000	0.3692	3.5677
3.3034	13.4072	46000	0.3694	3.5622
3.3426	13.6988	47000	0.3702	3.5562
3.3491	13.9903	48000	0.3703	3.5490
3.281	14.2816	49000	0.3700	3.5613
3.3029	14.5731	50000	0.3705	3.5537
3.3212	14.8646	51000	0.3707	3.5477
3.2535	15.1560	52000	0.3702	3.5648
3.2847	15.4475	53000	0.3707	3.5570
3.3045	15.7390	54000	0.3711	3.5483
3.2177	16.0303	55000	0.3706	3.5641
3.2647	16.3218	56000	0.3708	3.5583
3.2833	16.6133	57000	0.3712	3.5505
3.3074	16.9049	58000	0.3718	3.5425
3.2219	17.1962	59000	0.3709	3.5612
3.2688	17.4877	60000	0.3715	3.5518
3.3008	17.7792	61000	0.3720	3.5442
3.1956	18.0705	62000	0.3715	3.5577
3.2442	18.3621	63000	0.3712	3.5578
3.2687	18.6536	64000	0.3717	3.5480
3.2761	18.9451	65000	0.3723	3.5432
3.2104	19.2364	66000	0.3715	3.5592
3.2417	19.5279	67000	0.3720	3.5504
3.2588	19.8194	68000	0.3723	3.5423
3.1794	20.1108	69000	0.3720	3.5576
3.2109	20.4023	70000	0.3720	3.5561
3.2441	20.6938	71000	0.3724	3.5467
3.2459	20.9853	72000	0.3731	3.5375
3.2019	21.2766	73000	0.3722	3.5560
3.2187	21.5682	74000	0.3728	3.5467
3.2485	21.8597	75000	0.3730	3.5402
3.176	22.1510	76000	0.3720	3.5589
3.2122	22.4425	77000	0.3724	3.5536
3.2284	22.7340	78000	0.3730	3.5429
3.1336	23.0254	79000	0.3726	3.5554
3.197	23.3169	80000	0.3723	3.5569
3.1915	23.6084	81000	3.5587	0.3724
3.2062	23.8999	82000	3.5512	0.3728
3.173	24.1915	83000	3.5606	0.3725
3.1911	24.4830	84000	3.5545	0.3727
3.2223	24.7745	85000	3.5455	0.3733
3.1325	25.0659	86000	3.5570	0.3730
3.1702	25.3574	87000	3.5559	0.3728
3.1968	25.6489	88000	3.5477	0.3733
3.2104	25.9404	89000	3.5387	0.3740
3.1293	26.2318	90000	3.5597	0.3730
3.1762	26.5233	91000	3.5529	0.3733
3.1947	26.8148	92000	3.5433	0.3738
3.127	27.1061	93000	3.5590	0.3726
3.1511	27.3976	94000	3.5501	0.3735
3.1798	27.6891	95000	3.5434	0.3737
3.1795	27.9806	96000	3.5402	0.3742
3.1285	28.2720	97000	3.5596	0.3730
3.1533	28.5635	98000	3.5491	0.3736
3.1743	28.8550	99000	3.5459	0.3737
3.1123	29.1463	100000	3.5596	0.3732

Framework versions

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