learning_rate: 0.0006
train_batch_size: 16
eval_batch_size: 16
seed: 2128
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	Accuracy	Validation Loss
4.8398	0.2911	1000	0.2529	4.7648
4.3355	0.5822	2000	0.2983	4.2930
4.1495	0.8733	3000	0.3154	4.0968
3.9951	1.1642	4000	0.3247	3.9961
3.9363	1.4553	5000	0.3314	3.9197
3.8839	1.7464	6000	0.3367	3.8592
3.7526	2.0373	7000	0.3412	3.8168
3.766	2.3284	8000	0.3443	3.7858
3.7479	2.6195	9000	0.3469	3.7589
3.7303	2.9106	10000	0.3492	3.7297
3.6335	3.2014	11000	0.3514	3.7179
3.652	3.4925	12000	0.3526	3.7003
3.6362	3.7837	13000	0.3545	3.6821
3.5353	4.0745	14000	0.3562	3.6759
3.5717	4.3656	15000	0.3571	3.6622
3.5824	4.6567	16000	0.3580	3.6500
3.5675	4.9478	17000	0.3595	3.6359
3.5042	5.2387	18000	0.3598	3.6383
3.5194	5.5298	19000	0.3608	3.6283
3.5324	5.8209	20000	0.3615	3.6160
3.4461	6.1118	21000	0.3622	3.6225
3.4689	6.4029	22000	0.3631	3.6140
3.4899	6.6940	23000	0.3636	3.6039
3.4978	6.9851	24000	0.3645	3.5925
3.4291	7.2760	25000	0.3642	3.6047
3.4469	7.5671	26000	0.3652	3.5932
3.4732	7.8582	27000	0.3658	3.5826
3.3777	8.1490	28000	0.3657	3.5954
3.42	8.4401	29000	0.3664	3.5884
3.4301	8.7313	30000	0.3668	3.5769
3.3322	9.0221	31000	0.3669	3.5833
3.3805	9.3132	32000	0.3669	3.5828
3.3954	9.6043	33000	0.3676	3.5744
3.4259	9.8954	34000	0.3681	3.5650
3.333	10.1863	35000	0.3681	3.5745
3.3681	10.4774	36000	0.3682	3.5690
3.3912	10.7685	37000	0.3692	3.5620
3.2803	11.0594	38000	0.3689	3.5714
3.3448	11.3505	39000	0.3689	3.5686
3.3651	11.6416	40000	0.3694	3.5607
3.3802	11.9327	41000	0.3700	3.5502
3.3134	12.2236	42000	0.3696	3.5664
3.3343	12.5147	43000	0.3700	3.5568
3.3549	12.8058	44000	0.3705	3.5525
3.2679	13.0966	45000	0.3700	3.5613
3.3059	13.3878	46000	0.3704	3.5584
3.3345	13.6789	47000	0.3709	3.5532
3.3374	13.9700	48000	0.3709	3.5463
3.2745	14.2608	49000	0.3706	3.5624
3.3093	14.5519	50000	0.3714	3.5511
3.3227	14.8430	51000	0.3716	3.5456
3.2337	15.1339	52000	0.3712	3.5572
3.2897	15.4250	53000	0.3712	3.5528
3.2934	15.7161	54000	0.3721	3.5433
3.2636	16.0070	55000	0.3714	3.5543
3.253	16.2981	56000	0.3717	3.5547
3.2757	16.5892	57000	0.3725	3.5476
3.307	16.8803	58000	0.3727	3.5381
3.2247	17.1712	59000	0.3719	3.5537
3.2706	17.4623	60000	0.3722	3.5488
3.2979	17.7534	61000	0.3729	3.5416
3.1981	18.0442	62000	0.3721	3.5527
3.2339	18.3354	63000	0.3723	3.5512
3.2624	18.6265	64000	0.3726	3.5434
3.2776	18.9176	65000	0.3734	3.5375
3.2151	19.2084	66000	0.3723	3.5537
3.2552	19.4995	67000	0.3728	3.5445
3.2477	19.7906	68000	0.3734	3.5363
3.177	20.0815	69000	0.3726	3.5546
3.2167	20.3726	70000	0.3731	3.5491
3.2416	20.6637	71000	0.3735	3.5394
3.2537	20.9548	72000	0.3740	3.5325
3.194	21.2457	73000	0.3729	3.5480
3.2145	21.5368	74000	0.3736	3.5432
3.2299	21.8279	75000	0.3738	3.5391
3.1578	22.1188	76000	0.3732	3.5570
3.201	22.4099	77000	0.3734	3.5461
3.2231	22.7010	78000	0.3740	3.5404
3.2456	22.9921	79000	0.3747	3.5309
3.1852	23.2830	80000	0.3735	3.5509
3.1725	23.5741	81000	3.5519	0.3734
3.2021	23.8652	82000	3.5453	0.3739
3.1586	24.1563	83000	3.5552	0.3735
3.1881	24.4474	84000	3.5480	0.3739
3.2151	24.7385	85000	3.5413	0.3742
3.1227	25.0294	86000	3.5490	0.3739
3.1662	25.3205	87000	3.5512	0.3737
3.1804	25.6116	88000	3.5430	0.3742
3.2094	25.9027	89000	3.5351	0.3748
3.1359	26.1936	90000	3.5517	0.3737
3.1616	26.4847	91000	3.5488	0.3743
3.1978	26.7758	92000	3.5386	0.3750
3.1122	27.0667	93000	3.5500	0.3740
3.1532	27.3578	94000	3.5464	0.3746
3.1699	27.6489	95000	3.5423	0.3745
3.1885	27.9400	96000	3.5343	0.3753
3.1266	28.2308	97000	3.5509	0.3744
3.1488	28.5219	98000	3.5454	0.3748
3.1672	28.8131	99000	3.5403	0.3751
3.1066	29.1039	100000	3.5525	0.3741

Framework versions

Transformers 4.55.2
Pytorch 2.8.0+cu128
Datasets 4.0.0
Tokenizers 0.21.4

craa/exceptions_exp2_swap_take_to_hit_2128

exceptions_exp2_swap_take_to_hit_2128

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

craa/
exceptions_exp2_swap_take_to_hit_2128