exceptions_exp2_swap_0.7_cost_to_hit_1032

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

• Loss: 3.5646
• Accuracy: 0.3686

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
4.8221	0.2917	1000	4.7569	0.2546
4.3421	0.5834	2000	4.2914	0.2982
4.1487	0.8750	3000	4.1042	0.3144
4.0009	1.1665	4000	3.9999	0.3238
3.9336	1.4582	5000	3.9230	0.3305
3.8908	1.7499	6000	3.8660	0.3356
3.7589	2.0414	7000	3.8261	0.3395
3.7661	2.3331	8000	3.7927	0.3428
3.7454	2.6248	9000	3.7654	0.3455
3.725	2.9165	10000	3.7383	0.3480
3.6417	3.2080	11000	3.7282	0.3500
3.6677	3.4996	12000	3.7077	0.3516
3.6512	3.7913	13000	3.6904	0.3534
3.5601	4.0828	14000	3.6820	0.3546
3.584	4.3745	15000	3.6725	0.3556
3.5863	4.6662	16000	3.6579	0.3568
3.5872	4.9579	17000	3.6451	0.3580
3.5197	5.2494	18000	3.6452	0.3587
3.5282	5.5411	19000	3.6358	0.3596
3.5553	5.8327	20000	3.6251	0.3603
3.4602	6.1243	21000	3.6292	0.3609
3.4854	6.4159	22000	3.6199	0.3615
3.503	6.7076	23000	3.6109	0.3623
3.5058	6.9993	24000	3.6009	0.3633
3.4498	7.2908	25000	3.6103	0.3632
3.4515	7.5825	26000	3.6012	0.3636
3.4691	7.8742	27000	3.5921	0.3645
3.4056	8.1657	28000	3.6015	0.3645
3.4334	8.4574	29000	3.5922	0.3651
3.4357	8.7490	30000	3.5852	0.3656
3.3333	9.0405	31000	3.5911	0.3656
3.3831	9.3322	32000	3.5890	0.3660
3.409	9.6239	33000	3.5798	0.3663
3.4225	9.9156	34000	3.5714	0.3670
3.3415	10.2071	35000	3.5819	0.3666
3.3699	10.4988	36000	3.5778	0.3675
3.3966	10.7905	37000	3.5684	0.3679
3.3106	11.0820	38000	3.5785	0.3675
3.3351	11.3736	39000	3.5726	0.3680
3.3719	11.6653	40000	3.5646	0.3686
3.3753	11.9570	41000	3.5572	0.3689
3.3277	12.2485	42000	3.5722	0.3685
3.3353	12.5402	43000	3.5649	0.3691
3.3497	12.8319	44000	3.5565	0.3697
3.2741	13.1234	45000	3.5682	0.3690
3.3205	13.4151	46000	3.5642	0.3690
3.341	13.7067	47000	3.5599	0.3695
3.3485	13.9984	48000	3.5488	0.3702
3.2872	14.2899	49000	3.5674	0.3694
3.3068	14.5816	50000	3.5566	0.3701
3.3281	14.8733	51000	3.5493	0.3705
3.2609	15.1648	52000	3.5631	0.3701
3.2887	15.4565	53000	3.5571	0.3703
3.3067	15.7482	54000	3.5516	0.3703
3.2279	16.0397	55000	3.5599	0.3704
3.2746	16.3313	56000	3.5587	0.3705
3.2742	16.6230	57000	3.5497	0.3711
3.3163	16.9147	58000	3.5425	0.3716
3.2473	17.2062	59000	3.5627	0.3707
3.2628	17.4979	60000	3.5540	0.3712
3.2957	17.7896	61000	3.5472	0.3719
3.1994	18.0811	62000	3.5615	0.3711
3.25	18.3728	63000	3.5573	0.3714
3.2692	18.6644	64000	3.5499	0.3718
3.2892	18.9561	65000	3.5386	0.3723
3.2214	19.2476	66000	3.5611	0.3712
3.2539	19.5393	67000	3.5512	0.3719
3.2631	19.8310	68000	3.5424	0.3724
3.1846	20.1225	69000	3.5588	0.3717
3.2301	20.4142	70000	3.5536	0.3717
3.2655	20.7059	71000	3.5460	0.3723
3.26	20.9975	72000	3.5411	0.3726
3.2111	21.2891	73000	3.5564	0.3721
3.229	21.5807	74000	3.5493	0.3724
3.2504	21.8724	75000	3.5445	0.3727
3.1765	22.1639	76000	3.5585	0.3721
3.2192	22.4556	77000	3.5493	0.3725
3.233	22.7473	78000	3.5401	0.3730
3.1501	23.0388	79000	3.5584	0.3724
3.1865	23.3305	80000	3.5549	0.3723
3.222	23.6222	81000	3.5462	0.3730
3.2319	23.9138	82000	3.5438	0.3731
3.1613	24.2053	83000	3.5570	0.3725
3.1928	24.4970	84000	3.5472	0.3728
3.2174	24.7887	85000	3.5424	0.3734

Framework versions

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