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
4.8281	0.2917	1000	4.7607	0.2531
4.3263	0.5834	2000	4.2864	0.2984
4.1425	0.8750	3000	4.0968	0.3152
4.0034	1.1665	4000	3.9914	0.3250
3.9398	1.4582	5000	3.9150	0.3316
3.8857	1.7499	6000	3.8582	0.3364
3.7474	2.0414	7000	3.8157	0.3407
3.7564	2.3331	8000	3.7844	0.3436
3.736	2.6248	9000	3.7552	0.3466
3.7264	2.9165	10000	3.7287	0.3491
3.6334	3.2080	11000	3.7176	0.3510
3.655	3.4996	12000	3.6996	0.3527
3.6397	3.7913	13000	3.6824	0.3544
3.5397	4.0828	14000	3.6745	0.3557
3.5661	4.3745	15000	3.6635	0.3566
3.583	4.6662	16000	3.6486	0.3576
3.5789	4.9579	17000	3.6356	0.3591
3.5116	5.2494	18000	3.6395	0.3596
3.5135	5.5411	19000	3.6298	0.3603
3.5229	5.8327	20000	3.6175	0.3613
3.4495	6.1243	21000	3.6229	0.3618
3.4671	6.4159	22000	3.6152	0.3622
3.5027	6.7076	23000	3.6032	0.3629
3.5006	6.9993	24000	3.5945	0.3640
3.4233	7.2908	25000	3.6029	0.3640
3.4467	7.5825	26000	3.5951	0.3645
3.4647	7.8742	27000	3.5836	0.3654
3.3884	8.1657	28000	3.5942	0.3650
3.4106	8.4574	29000	3.5900	0.3658
3.424	8.7490	30000	3.5800	0.3666
3.332	9.0405	31000	3.5843	0.3663
3.3811	9.3322	32000	3.5829	0.3665
3.4042	9.6239	33000	3.5741	0.3673
3.414	9.9156	34000	3.5658	0.3680
3.3407	10.2071	35000	3.5791	0.3674
3.3864	10.4988	36000	3.5739	0.3677
3.3817	10.7905	37000	3.5621	0.3686
3.3076	11.0820	38000	3.5742	0.3683
3.3449	11.3736	39000	3.5730	0.3683
3.3728	11.6653	40000	3.5609	0.3689
3.3862	11.9570	41000	3.5552	0.3692
3.3133	12.2485	42000	3.5694	0.3689
3.3422	12.5402	43000	3.5615	0.3694
3.3482	12.8319	44000	3.5541	0.3699
3.2759	13.1234	45000	3.5676	0.3696
3.3064	13.4151	46000	3.5617	0.3696
3.3411	13.7067	47000	3.5516	0.3702
3.3521	13.9984	48000	3.5463	0.3706
3.2865	14.2899	49000	3.5642	0.3701
3.309	14.5816	50000	3.5536	0.3707
3.3135	14.8733	51000	3.5459	0.3709
3.2564	15.1648	52000	3.5610	0.3705
3.2879	15.4565	53000	3.5565	0.3706
3.3013	15.7482	54000	3.5498	0.3709
3.2217	16.0397	55000	3.5561	0.3709
3.2651	16.3313	56000	3.5599	0.3709
3.2861	16.6230	57000	3.5469	0.3714
3.3152	16.9147	58000	3.5442	0.3719
3.2311	17.2062	59000	3.5576	0.3716
3.2572	17.4979	60000	3.5505	0.3714
3.2908	17.7896	61000	3.5449	0.3721
3.1963	18.0811	62000	3.5589	0.3717
3.2406	18.3728	63000	3.5535	0.3718
3.2592	18.6644	64000	3.5459	0.3722
3.2814	18.9561	65000	3.5385	0.3726
3.2211	19.2476	66000	3.5567	0.3719
3.2476	19.5393	67000	3.5483	0.3723
3.2615	19.8310	68000	3.5409	0.3727
3.1946	20.1225	69000	3.5572	0.3718
3.2174	20.4142	70000	3.5493	0.3723
3.2422	20.7059	71000	3.5434	0.3728
3.2751	20.9975	72000	3.5361	0.3733
3.2049	21.2891	73000	3.5524	0.3723
3.2333	21.5807	74000	3.5469	0.3725
3.2488	21.8724	75000	3.5376	0.3735
3.1857	22.1639	76000	3.5572	0.3722
3.1884	22.4556	77000	3.5513	0.3727
3.2377	22.7473	78000	3.5408	0.3731
3.1471	23.0388	79000	3.5568	0.3724
3.1842	23.3305	80000	3.5512	0.3728
3.2033	23.6222	81000	3.5474	0.3730
3.2302	23.9138	82000	3.5420	0.3735
3.1648	24.2053	83000	3.5547	0.3727
3.2017	24.4970	84000	3.5503	0.3729
3.2177	24.7887	85000	3.5406	0.3736
3.1402	25.0802	86000	3.5542	0.3731
3.1719	25.3719	87000	3.5529	0.3729
3.1924	25.6636	88000	3.5418	0.3737
3.2088	25.9553	89000	3.5374	0.3740
3.1554	26.2468	90000	3.5530	0.3732
3.1914	26.5384	91000	3.5468	0.3733
3.1966	26.8301	92000	3.5432	0.3739

Framework versions

Transformers 4.55.2
Pytorch 2.8.0+cu128
Datasets 4.0.0
Tokenizers 0.21.4

craa/exceptions_exp2_swap_0.7_cost_to_hit_3591

exceptions_exp2_swap_0.7_cost_to_hit_3591

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

craa/
exceptions_exp2_swap_0.7_cost_to_hit_3591