exceptions_exp2_swap_0.7_last_to_hit_40817

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

• Loss: 3.5634
• Accuracy: 0.3687

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
4.82	0.2915	1000	0.2546	4.7491
4.3452	0.5830	2000	0.2995	4.2827
4.1511	0.8745	3000	0.3149	4.1005
4.0159	1.1659	4000	0.3246	3.9934
3.9427	1.4574	5000	0.3316	3.9212
3.8685	1.7489	6000	0.3363	3.8624
3.7477	2.0402	7000	0.3407	3.8195
3.7461	2.3317	8000	0.3436	3.7886
3.7386	2.6233	9000	0.3461	3.7604
3.7339	2.9148	10000	0.3488	3.7331
3.6372	3.2061	11000	0.3506	3.7208
3.6421	3.4976	12000	0.3522	3.7021
3.6399	3.7891	13000	0.3540	3.6850
3.5387	4.0805	14000	0.3549	3.6801
3.5763	4.3720	15000	0.3564	3.6654
3.5914	4.6635	16000	0.3574	3.6534
3.5959	4.9550	17000	0.3590	3.6384
3.5027	5.2463	18000	0.3594	3.6431
3.5203	5.5378	19000	0.3602	3.6316
3.5458	5.8293	20000	0.3610	3.6181
3.4446	6.1207	21000	0.3617	3.6232
3.4736	6.4122	22000	0.3623	3.6165
3.4812	6.7037	23000	0.3628	3.6074
3.493	6.9952	24000	0.3639	3.5957
3.4286	7.2866	25000	0.3640	3.6034
3.4475	7.5781	26000	0.3644	3.5973
3.4659	7.8696	27000	0.3653	3.5870
3.3908	8.1609	28000	0.3649	3.5949
3.4168	8.4524	29000	0.3653	3.5919
3.4338	8.7439	30000	0.3663	3.5821
3.3249	9.0353	31000	0.3661	3.5867
3.3866	9.3268	32000	0.3665	3.5841
3.4114	9.6183	33000	0.3671	3.5765
3.4155	9.9098	34000	0.3676	3.5680
3.3406	10.2011	35000	0.3673	3.5785
3.3818	10.4927	36000	0.3679	3.5723
3.3918	10.7842	37000	0.3686	3.5650
3.2962	11.0755	38000	0.3680	3.5764
3.3402	11.3670	39000	0.3684	3.5720
3.3688	11.6585	40000	0.3687	3.5634
3.3751	11.9500	41000	0.3695	3.5535
3.3316	12.2414	42000	0.3684	3.5711
3.3401	12.5329	43000	0.3694	3.5637
3.3441	12.8244	44000	0.3698	3.5547
3.2688	13.1157	45000	0.3695	3.5672
3.3015	13.4072	46000	0.3699	3.5612
3.3412	13.6988	47000	0.3704	3.5539
3.348	13.9903	48000	0.3707	3.5471
3.2793	14.2816	49000	0.3699	3.5617
3.302	14.5731	50000	0.3707	3.5510
3.3196	14.8646	51000	0.3709	3.5470
3.2517	15.1560	52000	0.3703	3.5652
3.282	15.4475	53000	0.3709	3.5575
3.3025	15.7390	54000	0.3712	3.5482
3.2162	16.0303	55000	0.3710	3.5597
3.2625	16.3218	56000	0.3709	3.5593
3.2817	16.6133	57000	0.3715	3.5498
3.3056	16.9049	58000	0.3723	3.5401
3.2195	17.1962	59000	0.3711	3.5591
3.267	17.4877	60000	0.3718	3.5491
3.2985	17.7792	61000	0.3721	3.5426
3.1927	18.0705	62000	0.3714	3.5552
3.2428	18.3621	63000	0.3715	3.5540
3.2681	18.6536	64000	0.3719	3.5466
3.2754	18.9451	65000	0.3727	3.5389
3.2074	19.2364	66000	0.3719	3.5573
3.2401	19.5279	67000	0.3721	3.5492
3.257	19.8194	68000	0.3725	3.5415
3.1764	20.1108	69000	0.3722	3.5544
3.2093	20.4023	70000	0.3724	3.5530
3.2427	20.6938	71000	0.3727	3.5470
3.2435	20.9853	72000	0.3735	3.5348
3.1998	21.2766	73000	0.3723	3.5532
3.2156	21.5682	74000	0.3728	3.5487
3.2468	21.8597	75000	0.3733	3.5367
3.174	22.1510	76000	0.3723	3.5572
3.2102	22.4425	77000	0.3726	3.5492
3.2254	22.7340	78000	0.3735	3.5386
3.1307	23.0254	79000	0.3730	3.5528
3.1952	23.3169	80000	0.3727	3.5524
3.1882	23.6084	81000	3.5551	0.3726
3.2042	23.8999	82000	3.5464	0.3731
3.1701	24.1915	83000	3.5567	0.3728
3.1877	24.4830	84000	3.5519	0.3730
3.2204	24.7745	85000	3.5413	0.3737
3.13	25.0659	86000	3.5559	0.3730
3.1678	25.3574	87000	3.5530	0.3731
3.1946	25.6489	88000	3.5448	0.3736
3.2083	25.9404	89000	3.5375	0.3741
3.1252	26.2318	90000	3.5572	0.3735
3.1743	26.5233	91000	3.5456	0.3737
3.1933	26.8148	92000	3.5410	0.3740
3.1236	27.1061	93000	3.5582	0.3728
3.1469	27.3976	94000	3.5510	0.3736
3.1768	27.6891	95000	3.5412	0.3739
3.1761	27.9806	96000	3.5355	0.3746
3.1261	28.2720	97000	3.5564	0.3734
3.1522	28.5635	98000	3.5489	0.3739
3.1705	28.8550	99000	3.5439	0.3743
3.1081	29.1463	100000	3.5558	0.3736

Framework versions

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