exceptions_exp2_swap_0.7_cost_to_drop_2128

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

• Loss: 3.5663
• Accuracy: 0.3683

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: 2128
• 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
4.832	0.2917	1000	4.7527	0.2543
4.3421	0.5834	2000	4.2906	0.2985
4.1495	0.8750	3000	4.1056	0.3146
3.9908	1.1665	4000	3.9981	0.3238
3.9542	1.4582	5000	3.9231	0.3306
3.8839	1.7499	6000	3.8660	0.3357
3.7578	2.0414	7000	3.8234	0.3402
3.7666	2.3331	8000	3.7898	0.3433
3.7525	2.6248	9000	3.7616	0.3458
3.7258	2.9165	10000	3.7372	0.3483
3.6466	3.2080	11000	3.7261	0.3501
3.6628	3.4996	12000	3.7039	0.3521
3.6379	3.7913	13000	3.6881	0.3536
3.5461	4.0828	14000	3.6816	0.3549
3.583	4.3745	15000	3.6699	0.3557
3.5925	4.6662	16000	3.6561	0.3569
3.5858	4.9579	17000	3.6425	0.3582
3.5167	5.2494	18000	3.6457	0.3588
3.5331	5.5411	19000	3.6345	0.3598
3.5265	5.8327	20000	3.6231	0.3605
3.4474	6.1243	21000	3.6285	0.3609
3.4822	6.4159	22000	3.6201	0.3614
3.5014	6.7076	23000	3.6109	0.3624
3.5041	6.9993	24000	3.5990	0.3632
3.4445	7.2908	25000	3.6086	0.3634
3.4609	7.5825	26000	3.6004	0.3639
3.4543	7.8742	27000	3.5917	0.3644
3.3872	8.1657	28000	3.5976	0.3646
3.4203	8.4574	29000	3.5922	0.3651
3.429	8.7490	30000	3.5787	0.3658
3.3283	9.0405	31000	3.5919	0.3655
3.396	9.3322	32000	3.5867	0.3661
3.4124	9.6239	33000	3.5808	0.3667
3.4131	9.9156	34000	3.5718	0.3670
3.358	10.2071	35000	3.5841	0.3668
3.3717	10.4988	36000	3.5771	0.3671
3.3815	10.7905	37000	3.5663	0.3678
3.303	11.0820	38000	3.5782	0.3675
3.356	11.3736	39000	3.5755	0.3675
3.3704	11.6653	40000	3.5663	0.3683
3.3764	11.9570	41000	3.5599	0.3689
3.3064	12.2485	42000	3.5762	0.3682
3.3481	12.5402	43000	3.5674	0.3686
3.3554	12.8319	44000	3.5573	0.3693
3.2778	13.1234	45000	3.5729	0.3690
3.3194	13.4151	46000	3.5645	0.3693
3.3408	13.7067	47000	3.5554	0.3699
3.3448	13.9984	48000	3.5526	0.3703
3.2808	14.2899	49000	3.5677	0.3694
3.3114	14.5816	50000	3.5577	0.3699
3.3368	14.8733	51000	3.5524	0.3704
3.2588	15.1648	52000	3.5674	0.3698
3.2919	15.4565	53000	3.5593	0.3702
3.313	15.7482	54000	3.5527	0.3707
3.216	16.0397	55000	3.5626	0.3705
3.2571	16.3313	56000	3.5595	0.3706
3.2919	16.6230	57000	3.5552	0.3711
3.299	16.9147	58000	3.5429	0.3714
3.2411	17.2062	59000	3.5603	0.3706
3.2522	17.4979	60000	3.5590	0.3709
3.3022	17.7896	61000	3.5454	0.3718
3.2061	18.0811	62000	3.5610	0.3709
3.2462	18.3728	63000	3.5575	0.3711
3.2517	18.6644	64000	3.5495	0.3718
3.2749	18.9561	65000	3.5402	0.3721
3.2242	19.2476	66000	3.5596	0.3712
3.2399	19.5393	67000	3.5523	0.3718
3.2707	19.8310	68000	3.5454	0.3725
3.1815	20.1225	69000	3.5601	0.3715
3.2264	20.4142	70000	3.5547	0.3718
3.2509	20.7059	71000	3.5468	0.3723
3.2589	20.9975	72000	3.5424	0.3725
3.2001	21.2891	73000	3.5601	0.3718
3.2212	21.5807	74000	3.5515	0.3722
3.2545	21.8724	75000	3.5412	0.3725
3.171	22.1639	76000	3.5623	0.3718
3.2161	22.4556	77000	3.5560	0.3722
3.2172	22.7473	78000	3.5494	0.3725
3.1352	23.0388	79000	3.5588	0.3724
3.1928	23.3305	80000	3.5580	0.3721
3.2016	23.6222	81000	3.5519	0.3725
3.228	23.9138	82000	3.5407	0.3732
3.158	24.2053	83000	3.5605	0.3724
3.1928	24.4970	84000	3.5524	0.3727
3.2238	24.7887	85000	3.5443	0.3733

Framework versions

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