exceptions_exp2_swap_0.7_cost_to_carry_40817

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

• Loss: 3.5678
• Accuracy: 0.3681

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	Validation Loss	Accuracy
4.8301	0.2917	1000	4.7575	0.2541
4.3486	0.5834	2000	4.2903	0.2983
4.151	0.8750	3000	4.1045	0.3137
4.0029	1.1665	4000	3.9993	0.3233
3.943	1.4582	5000	3.9266	0.3307
3.8895	1.7499	6000	3.8661	0.3358
3.7599	2.0414	7000	3.8252	0.3397
3.7708	2.3331	8000	3.7931	0.3431
3.7523	2.6248	9000	3.7650	0.3459
3.7163	2.9165	10000	3.7377	0.3483
3.6534	3.2080	11000	3.7240	0.3504
3.6556	3.4996	12000	3.7051	0.3519
3.6454	3.7913	13000	3.6903	0.3537
3.5639	4.0828	14000	3.6797	0.3546
3.5946	4.3745	15000	3.6679	0.3560
3.5971	4.6662	16000	3.6541	0.3571
3.5844	4.9579	17000	3.6434	0.3583
3.5061	5.2494	18000	3.6471	0.3588
3.5432	5.5411	19000	3.6343	0.3598
3.5267	5.8327	20000	3.6227	0.3607
3.4531	6.1243	21000	3.6279	0.3610
3.4732	6.4159	22000	3.6212	0.3616
3.5053	6.7076	23000	3.6081	0.3624
3.4999	6.9993	24000	3.6019	0.3634
3.4312	7.2908	25000	3.6102	0.3633
3.4582	7.5825	26000	3.6012	0.3638
3.4585	7.8742	27000	3.5917	0.3647
3.3953	8.1657	28000	3.6004	0.3646
3.4218	8.4574	29000	3.5929	0.3651
3.4533	8.7490	30000	3.5832	0.3658
3.3363	9.0405	31000	3.5899	0.3659
3.3905	9.3322	32000	3.5883	0.3658
3.418	9.6239	33000	3.5793	0.3668
3.4051	9.9156	34000	3.5742	0.3670
3.3435	10.2071	35000	3.5806	0.3672
3.388	10.4988	36000	3.5767	0.3670
3.3823	10.7905	37000	3.5710	0.3680
3.3081	11.0820	38000	3.5791	0.3676
3.3369	11.3736	39000	3.5739	0.3682
3.3757	11.6653	40000	3.5678	0.3681
3.3852	11.9570	41000	3.5594	0.3689
3.3048	12.2485	42000	3.5723	0.3684
3.3532	12.5402	43000	3.5644	0.3690
3.367	12.8319	44000	3.5602	0.3691
3.2762	13.1234	45000	3.5748	0.3687
3.3099	13.4151	46000	3.5679	0.3690
3.3324	13.7067	47000	3.5581	0.3696
3.3437	13.9984	48000	3.5511	0.3702
3.2975	14.2899	49000	3.5657	0.3695
3.3225	14.5816	50000	3.5598	0.3702
3.3356	14.8733	51000	3.5503	0.3707
3.2608	15.1648	52000	3.5670	0.3697
3.2884	15.4565	53000	3.5597	0.3700
3.3005	15.7482	54000	3.5502	0.3707
3.2013	16.0397	55000	3.5633	0.3706
3.2553	16.3313	56000	3.5654	0.3705
3.2987	16.6230	57000	3.5556	0.3709
3.3011	16.9147	58000	3.5492	0.3716
3.2305	17.2062	59000	3.5619	0.3706
3.2706	17.4979	60000	3.5560	0.3711
3.2863	17.7896	61000	3.5471	0.3716
3.205	18.0811	62000	3.5629	0.3712
3.2492	18.3728	63000	3.5590	0.3714
3.2735	18.6644	64000	3.5498	0.3718
3.2878	18.9561	65000	3.5449	0.3722
3.2248	19.2476	66000	3.5597	0.3713
3.2494	19.5393	67000	3.5502	0.3718
3.2803	19.8310	68000	3.5451	0.3722
3.1923	20.1225	69000	3.5581	0.3716
3.2183	20.4142	70000	3.5567	0.3717
3.242	20.7059	71000	3.5510	0.3722
3.2699	20.9975	72000	3.5392	0.3729
3.2087	21.2891	73000	3.5543	0.3722
3.2337	21.5807	74000	3.5495	0.3722
3.2494	21.8724	75000	3.5426	0.3725
3.1719	22.1639	76000	3.5594	0.3720
3.2187	22.4556	77000	3.5540	0.3721
3.2241	22.7473	78000	3.5453	0.3730
3.1486	23.0388	79000	3.5574	0.3724
3.1855	23.3305	80000	3.5552	0.3722
3.2115	23.6222	81000	3.5485	0.3728
3.23	23.9138	82000	3.5402	0.3733
3.1697	24.2053	83000	3.5561	0.3726
3.1902	24.4970	84000	3.5511	0.3728
3.2124	24.7887	85000	3.5448	0.3734
3.1468	25.0802	86000	3.5587	0.3726
3.1786	25.3719	87000	3.5556	0.3728
3.1946	25.6636	88000	3.5459	0.3734
3.2125	25.9553	89000	3.5403	0.3737
3.1521	26.2468	90000	3.5599	0.3728
3.1818	26.5384	91000	3.5531	0.3729
3.1924	26.8301	92000	3.5420	0.3739

Framework versions

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