exceptions_exp2_swap_0.3_cost_to_carry_1032

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

• Loss: 3.5640
• Accuracy: 0.3685

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: 1032
• 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
4.8436	0.2915	1000	4.7547	0.2544
4.3398	0.5831	2000	4.2880	0.2984
4.1463	0.8746	3000	4.1043	0.3142
4.0043	1.1662	4000	3.9991	0.3237
3.9387	1.4577	5000	3.9236	0.3303
3.8945	1.7493	6000	3.8673	0.3352
3.7576	2.0408	7000	3.8236	0.3396
3.7767	2.3324	8000	3.7905	0.3430
3.744	2.6239	9000	3.7630	0.3455
3.7279	2.9155	10000	3.7367	0.3480
3.641	3.2070	11000	3.7234	0.3501
3.6501	3.4985	12000	3.7045	0.3518
3.6498	3.7901	13000	3.6869	0.3534
3.546	4.0816	14000	3.6799	0.3546
3.5752	4.3732	15000	3.6678	0.3557
3.5809	4.6647	16000	3.6550	0.3572
3.5803	4.9563	17000	3.6412	0.3582
3.5077	5.2478	18000	3.6466	0.3587
3.5428	5.5394	19000	3.6313	0.3600
3.5318	5.8309	20000	3.6226	0.3609
3.4418	6.1224	21000	3.6271	0.3609
3.4764	6.4140	22000	3.6187	0.3616
3.4959	6.7055	23000	3.6105	0.3625
3.5064	6.9971	24000	3.5987	0.3636
3.433	7.2886	25000	3.6087	0.3631
3.4614	7.5802	26000	3.5969	0.3641
3.4882	7.8717	27000	3.5890	0.3646
3.3924	8.1633	28000	3.5971	0.3645
3.4168	8.4548	29000	3.5925	0.3650
3.442	8.7464	30000	3.5829	0.3658
3.3384	9.0379	31000	3.5868	0.3656
3.3887	9.3294	32000	3.5872	0.3659
3.3969	9.6210	33000	3.5774	0.3667
3.4197	9.9125	34000	3.5706	0.3673
3.3482	10.2041	35000	3.5797	0.3667
3.3764	10.4956	36000	3.5730	0.3673
3.3819	10.7872	37000	3.5683	0.3675
3.3052	11.0787	38000	3.5784	0.3675
3.3442	11.3703	39000	3.5762	0.3673
3.3616	11.6618	40000	3.5640	0.3685
3.3886	11.9534	41000	3.5570	0.3689
3.3255	12.2449	42000	3.5704	0.3685
3.3315	12.5364	43000	3.5669	0.3691
3.3545	12.8280	44000	3.5563	0.3692
3.2641	13.1195	45000	3.5686	0.3691
3.32	13.4111	46000	3.5622	0.3695
3.3317	13.7026	47000	3.5568	0.3698
3.352	13.9942	48000	3.5499	0.3703
3.2854	14.2857	49000	3.5648	0.3694
3.3242	14.5773	50000	3.5551	0.3701
3.331	14.8688	51000	3.5461	0.3706
3.2517	15.1603	52000	3.5672	0.3697
3.287	15.4519	53000	3.5583	0.3701
3.2955	15.7434	54000	3.5510	0.3708
3.216	16.0350	55000	3.5631	0.3707
3.2463	16.3265	56000	3.5602	0.3707
3.2955	16.6181	57000	3.5509	0.3710
3.305	16.9096	58000	3.5452	0.3716
3.2478	17.2012	59000	3.5606	0.3709
3.2753	17.4927	60000	3.5543	0.3710
3.2933	17.7843	61000	3.5479	0.3717
3.1944	18.0758	62000	3.5633	0.3711
3.2355	18.3673	63000	3.5555	0.3714
3.2665	18.6589	64000	3.5489	0.3720
3.2831	18.9504	65000	3.5393	0.3722
3.2148	19.2420	66000	3.5562	0.3717
3.2506	19.5335	67000	3.5517	0.3717
3.2641	19.8251	68000	3.5412	0.3724
3.1897	20.1166	69000	3.5608	0.3716
3.2344	20.4082	70000	3.5529	0.3720
3.2484	20.6997	71000	3.5486	0.3720
3.2601	20.9913	72000	3.5351	0.3729
3.2013	21.2828	73000	3.5560	0.3721
3.2321	21.5743	74000	3.5475	0.3727
3.2372	21.8659	75000	3.5416	0.3728
3.175	22.1574	76000	3.5576	0.3720
3.2168	22.4490	77000	3.5533	0.3722
3.2291	22.7405	78000	3.5452	0.3728
3.1442	23.0321	79000	3.5600	0.3724
3.1957	23.3236	80000	3.5546	0.3722
3.2163	23.6152	81000	3.5496	0.3726
3.2183	23.9067	82000	3.5404	0.3734
3.1689	24.1983	83000	3.5571	0.3724
3.1875	24.4898	84000	3.5511	0.3729
3.1996	24.7813	85000	3.5427	0.3733
3.1343	25.0729	86000	3.5559	0.3726
3.168	25.3644	87000	3.5532	0.3728
3.2022	25.6560	88000	3.5454	0.3731
3.1997	25.9475	89000	3.5397	0.3736
3.1478	26.2391	90000	3.5576	0.3726
3.1734	26.5306	91000	3.5525	0.3730
3.2067	26.8222	92000	3.5387	0.3738

Framework versions

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