exceptions_exp2_swap_0.3_cost_to_push_5039

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

• Loss: 3.5839
• Accuracy: 0.3657

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: 5039
• 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
4.8347	0.2915	1000	4.7477	0.2551
4.3492	0.5831	2000	4.2936	0.2980
4.1447	0.8746	3000	4.1108	0.3142
4.0065	1.1662	4000	3.9987	0.3237
3.947	1.4577	5000	3.9255	0.3307
3.8832	1.7493	6000	3.8682	0.3357
3.762	2.0408	7000	3.8264	0.3399
3.7664	2.3324	8000	3.7946	0.3426
3.7444	2.6239	9000	3.7642	0.3459
3.7319	2.9155	10000	3.7383	0.3481
3.6478	3.2070	11000	3.7237	0.3501
3.6472	3.4985	12000	3.7052	0.3519
3.6504	3.7901	13000	3.6869	0.3535
3.5369	4.0816	14000	3.6815	0.3548
3.5946	4.3732	15000	3.6687	0.3559
3.5779	4.6647	16000	3.6556	0.3572
3.5883	4.9563	17000	3.6456	0.3580
3.5046	5.2478	18000	3.6475	0.3589
3.5406	5.5394	19000	3.6340	0.3596
3.5537	5.8309	20000	3.6241	0.3607
3.4495	6.1224	21000	3.6262	0.3615
3.4806	6.4140	22000	3.6182	0.3621
3.4962	6.7055	23000	3.6080	0.3628
3.5033	6.9971	24000	3.5971	0.3636
3.4345	7.2886	25000	3.6087	0.3631
3.4486	7.5802	26000	3.5977	0.3640
3.4794	7.8717	27000	3.5907	0.3648
3.3944	8.1633	28000	3.5990	0.3646
3.4335	8.4548	29000	3.5924	0.3654
3.4418	8.7464	30000	3.5839	0.3657
3.3402	9.0379	31000	3.5898	0.3657
3.3896	9.3294	32000	3.5870	0.3660
3.4004	9.6210	33000	3.5777	0.3669
3.4144	9.9125	34000	3.5710	0.3674
3.355	10.2041	35000	3.5801	0.3669
3.38	10.4956	36000	3.5774	0.3674
3.3902	10.7872	37000	3.5692	0.3679
3.3013	11.0787	38000	3.5764	0.3676
3.3484	11.3703	39000	3.5735	0.3680
3.3765	11.6618	40000	3.5695	0.3681
3.3952	11.9534	41000	3.5555	0.3694
3.3156	12.2449	42000	3.5733	0.3684
3.3349	12.5364	43000	3.5639	0.3691
3.3619	12.8280	44000	3.5589	0.3694
3.2689	13.1195	45000	3.5710	0.3691
3.3035	13.4111	46000	3.5665	0.3693
3.3422	13.7026	47000	3.5574	0.3697
3.3608	13.9942	48000	3.5482	0.3705
3.2932	14.2857	49000	3.5640	0.3698
3.3201	14.5773	50000	3.5547	0.3703
3.3291	14.8688	51000	3.5487	0.3708
3.2643	15.1603	52000	3.5657	0.3701
3.2949	15.4519	53000	3.5531	0.3706
3.3079	15.7434	54000	3.5495	0.3710
3.2102	16.0350	55000	3.5603	0.3707
3.2576	16.3265	56000	3.5598	0.3708
3.2845	16.6181	57000	3.5505	0.3712
3.306	16.9096	58000	3.5435	0.3717
3.2408	17.2012	59000	3.5608	0.3709
3.267	17.4927	60000	3.5526	0.3713
3.2932	17.7843	61000	3.5454	0.3719
3.21	18.0758	62000	3.5635	0.3709
3.2458	18.3673	63000	3.5554	0.3713
3.2771	18.6589	64000	3.5464	0.3719
3.2783	18.9504	65000	3.5414	0.3723
3.225	19.2420	66000	3.5563	0.3717
3.2549	19.5335	67000	3.5513	0.3718
3.2745	19.8251	68000	3.5432	0.3726
3.1944	20.1166	69000	3.5609	0.3716
3.2177	20.4082	70000	3.5542	0.3719
3.2515	20.6997	71000	3.5439	0.3724
3.2727	20.9913	72000	3.5365	0.3731
3.2157	21.2828	73000	3.5576	0.3719
3.2286	21.5743	74000	3.5487	0.3726
3.2406	21.8659	75000	3.5430	0.3727
3.1809	22.1574	76000	3.5542	0.3724
3.2172	22.4490	77000	3.5473	0.3728
3.2286	22.7405	78000	3.5413	0.3734
3.1405	23.0321	79000	3.5549	0.3724
3.1821	23.3236	80000	3.5538	0.3726
3.2224	23.6152	81000	3.5478	0.3729
3.2195	23.9067	82000	3.5426	0.3732
3.1715	24.1983	83000	3.5573	0.3727
3.1955	24.4898	84000	3.5493	0.3730
3.2229	24.7813	85000	3.5422	0.3734
3.128	25.0729	86000	3.5573	0.3728
3.1646	25.3644	87000	3.5521	0.3729
3.1922	25.6560	88000	3.5463	0.3733
3.2165	25.9475	89000	3.5370	0.3739
3.1438	26.2391	90000	3.5560	0.3730
3.1807	26.5306	91000	3.5472	0.3734
3.1955	26.8222	92000	3.5432	0.3737

Framework versions

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