exceptions_exp2_swap_0.7_cost_to_carry_3591

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

• Loss: 3.5624
• Accuracy: 0.3692

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: 3591
• 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
4.8362	0.2917	1000	4.7704	0.2523
4.3333	0.5834	2000	4.2875	0.2990
4.1454	0.8750	3000	4.1005	0.3150
3.9922	1.1665	4000	3.9924	0.3250
3.9385	1.4582	5000	3.9142	0.3311
3.8819	1.7499	6000	3.8595	0.3364
3.7528	2.0414	7000	3.8158	0.3410
3.7629	2.3331	8000	3.7860	0.3437
3.7428	2.6248	9000	3.7583	0.3467
3.7208	2.9165	10000	3.7299	0.3491
3.641	3.2080	11000	3.7165	0.3509
3.6607	3.4996	12000	3.6995	0.3525
3.6518	3.7913	13000	3.6806	0.3544
3.5394	4.0828	14000	3.6759	0.3558
3.5705	4.3745	15000	3.6644	0.3568
3.5855	4.6662	16000	3.6487	0.3580
3.5665	4.9579	17000	3.6365	0.3591
3.4931	5.2494	18000	3.6386	0.3591
3.5116	5.5411	19000	3.6291	0.3602
3.5414	5.8327	20000	3.6186	0.3616
3.4313	6.1243	21000	3.6188	0.3619
3.4724	6.4159	22000	3.6149	0.3624
3.4963	6.7076	23000	3.6043	0.3632
3.4974	6.9993	24000	3.5945	0.3642
3.4179	7.2908	25000	3.6033	0.3638
3.4567	7.5825	26000	3.5936	0.3645
3.467	7.8742	27000	3.5834	0.3654
3.3857	8.1657	28000	3.5951	0.3652
3.4216	8.4574	29000	3.5911	0.3659
3.4213	8.7490	30000	3.5770	0.3661
3.3325	9.0405	31000	3.5867	0.3664
3.3676	9.3322	32000	3.5841	0.3667
3.4038	9.6239	33000	3.5736	0.3670
3.4186	9.9156	34000	3.5678	0.3678
3.337	10.2071	35000	3.5762	0.3678
3.3811	10.4988	36000	3.5723	0.3681
3.3902	10.7905	37000	3.5624	0.3687
3.3075	11.0820	38000	3.5757	0.3683
3.3411	11.3736	39000	3.5690	0.3685
3.3692	11.6653	40000	3.5624	0.3692
3.3714	11.9570	41000	3.5541	0.3697
3.3147	12.2485	42000	3.5698	0.3689
3.3397	12.5402	43000	3.5611	0.3695
3.3481	12.8319	44000	3.5515	0.3700
3.2629	13.1234	45000	3.5647	0.3699
3.3151	13.4151	46000	3.5584	0.3700
3.3393	13.7067	47000	3.5518	0.3705
3.3506	13.9984	48000	3.5436	0.3711
3.2836	14.2899	49000	3.5592	0.3702
3.298	14.5816	50000	3.5521	0.3709
3.3317	14.8733	51000	3.5435	0.3714
3.239	15.1648	52000	3.5581	0.3710
3.2946	15.4565	53000	3.5529	0.3711
3.3124	15.7482	54000	3.5439	0.3718
3.2093	16.0397	55000	3.5576	0.3711
3.2681	16.3313	56000	3.5533	0.3714
3.2903	16.6230	57000	3.5503	0.3715
3.3014	16.9147	58000	3.5405	0.3724
3.24	17.2062	59000	3.5589	0.3713
3.2514	17.4979	60000	3.5477	0.3719
3.2784	17.7896	61000	3.5419	0.3723
3.2041	18.0811	62000	3.5556	0.3717
3.2305	18.3728	63000	3.5537	0.3717
3.2568	18.6644	64000	3.5421	0.3728
3.2728	18.9561	65000	3.5354	0.3728
3.2181	19.2476	66000	3.5513	0.3723
3.2551	19.5393	67000	3.5471	0.3727
3.2648	19.8310	68000	3.5404	0.3729
3.1942	20.1225	69000	3.5566	0.3723
3.2332	20.4142	70000	3.5514	0.3725
3.2465	20.7059	71000	3.5421	0.3728
3.2544	20.9975	72000	3.5366	0.3735
3.1997	21.2891	73000	3.5538	0.3723
3.2355	21.5807	74000	3.5440	0.3730
3.2517	21.8724	75000	3.5395	0.3736
3.1723	22.1639	76000	3.5543	0.3731
3.2094	22.4556	77000	3.5496	0.3728
3.2334	22.7473	78000	3.5391	0.3736
3.1535	23.0388	79000	3.5548	0.3731
3.1955	23.3305	80000	3.5506	0.3735
3.2168	23.6222	81000	3.5413	0.3736
3.2321	23.9138	82000	3.5362	0.3741
3.1636	24.2053	83000	3.5528	0.3728
3.1834	24.4970	84000	3.5452	0.3735
3.2204	24.7887	85000	3.5397	0.3738

Framework versions

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