exceptions_exp2_swap_take_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: 3.5549
• Accuracy: 0.3699

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
4.8575	0.2911	1000	0.2535	4.7630
4.3442	0.5822	2000	0.2993	4.2880
4.1481	0.8733	3000	0.3151	4.0998
4.0028	1.1642	4000	0.3253	3.9876
3.9368	1.4553	5000	0.3324	3.9141
3.8855	1.7464	6000	0.3376	3.8546
3.7554	2.0373	7000	0.3417	3.8121
3.7443	2.3284	8000	0.3446	3.7834
3.743	2.6195	9000	0.3474	3.7528
3.7265	2.9106	10000	0.3496	3.7286
3.636	3.2014	11000	0.3513	3.7155
3.6463	3.4925	12000	0.3534	3.6948
3.6462	3.7837	13000	0.3549	3.6790
3.5388	4.0745	14000	0.3563	3.6700
3.5512	4.3656	15000	0.3575	3.6593
3.5794	4.6567	16000	0.3587	3.6458
3.577	4.9478	17000	0.3600	3.6332
3.4963	5.2387	18000	0.3601	3.6360
3.5062	5.5298	19000	0.3615	3.6241
3.5301	5.8209	20000	0.3620	3.6132
3.4355	6.1118	21000	0.3626	3.6153
3.4721	6.4029	22000	0.3631	3.6084
3.4879	6.6940	23000	0.3640	3.5990
3.4872	6.9851	24000	0.3651	3.5880
3.4348	7.2760	25000	0.3653	3.5969
3.442	7.5671	26000	0.3656	3.5887
3.4488	7.8582	27000	0.3662	3.5801
3.3946	8.1490	28000	0.3659	3.5885
3.3966	8.4401	29000	0.3665	3.5812
3.4248	8.7313	30000	0.3672	3.5737
3.3156	9.0221	31000	0.3675	3.5799
3.3859	9.3132	32000	0.3676	3.5768
3.3917	9.6043	33000	0.3682	3.5672
3.4102	9.8954	34000	0.3689	3.5580
3.329	10.1863	35000	0.3685	3.5709
3.361	10.4774	36000	0.3686	3.5663
3.3766	10.7685	37000	0.3694	3.5590
3.2923	11.0594	38000	0.3693	3.5660
3.3265	11.3505	39000	0.3695	3.5644
3.3713	11.6416	40000	0.3699	3.5549
3.3622	11.9327	41000	0.3706	3.5480
3.3006	12.2236	42000	0.3702	3.5634
3.3349	12.5147	43000	0.3708	3.5522
3.3434	12.8058	44000	0.3710	3.5475
3.2814	13.0966	45000	0.3702	3.5631
3.295	13.3878	46000	0.3708	3.5528
3.3066	13.6789	47000	0.3713	3.5490
3.3447	13.9700	48000	0.3718	3.5389
3.2718	14.2608	49000	0.3709	3.5539
3.2971	14.5519	50000	0.3717	3.5504
3.3174	14.8430	51000	0.3721	3.5420
3.2361	15.1339	52000	0.3714	3.5559
3.2771	15.4250	53000	0.3715	3.5537
3.2981	15.7161	54000	0.3723	3.5413
3.25	16.0070	55000	0.3720	3.5527
3.2385	16.2981	56000	0.3720	3.5493
3.2749	16.5892	57000	0.3723	3.5437
3.2776	16.8803	58000	0.3731	3.5353
3.2252	17.1712	59000	0.3723	3.5516
3.2539	17.4623	60000	0.3727	3.5474
3.2669	17.7534	61000	0.3733	3.5358
3.1814	18.0442	62000	0.3726	3.5520
3.2274	18.3354	63000	0.3727	3.5478
3.2574	18.6265	64000	0.3729	3.5397
3.2732	18.9176	65000	0.3738	3.5327
3.1984	19.2084	66000	0.3729	3.5487
3.2259	19.4995	67000	0.3733	3.5450
3.2471	19.7906	68000	0.3737	3.5362
3.1672	20.0815	69000	0.3728	3.5517
3.217	20.3726	70000	0.3731	3.5466
3.2198	20.6637	71000	0.3738	3.5368
3.2521	20.9548	72000	0.3742	3.5312
3.1796	21.2457	73000	0.3733	3.5484
3.2055	21.5368	74000	0.3738	3.5416
3.2252	21.8279	75000	0.3743	3.5332
3.1683	22.1188	76000	0.3733	3.5511
3.1948	22.4099	77000	0.3735	3.5473
3.2098	22.7010	78000	0.3740	3.5374
3.2212	22.9921	79000	0.3743	3.5343
3.1864	23.2830	80000	0.3737	3.5496
3.1843	23.5741	81000	3.5557	0.3735
3.2067	23.8652	82000	3.5417	0.3742
3.1414	24.1563	83000	3.5560	0.3731
3.1811	24.4474	84000	3.5489	0.3740
3.2035	24.7385	85000	3.5399	0.3744
3.1086	25.0294	86000	3.5526	0.3738
3.169	25.3205	87000	3.5482	0.3742
3.1861	25.6116	88000	3.5393	0.3747
3.2032	25.9027	89000	3.5314	0.3749
3.1414	26.1936	90000	3.5537	0.3741
3.1603	26.4847	91000	3.5435	0.3744
3.1793	26.7758	92000	3.5379	0.3748
3.1234	27.0667	93000	3.5519	0.3744
3.1479	27.3578	94000	3.5480	0.3744
3.1803	27.6489	95000	3.5397	0.3750
3.1758	27.9400	96000	3.5335	0.3752
3.1265	28.2308	97000	3.5515	0.3742
3.1609	28.5219	98000	3.5414	0.3750
3.173	28.8131	99000	3.5380	0.3751
3.0872	29.1039	100000	3.5472	0.3750

Framework versions

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