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	Accuracy	Validation Loss
4.8248	0.2911	1000	0.2556	4.7447
4.335	0.5822	2000	0.3005	4.2759
4.1426	0.8733	3000	0.3162	4.0902
3.991	1.1642	4000	0.3256	3.9910
3.9322	1.4553	5000	0.3321	3.9138
3.8641	1.7464	6000	0.3373	3.8570
3.7374	2.0373	7000	0.3418	3.8126
3.7425	2.3284	8000	0.3449	3.7801
3.7296	2.6195	9000	0.3476	3.7523
3.7159	2.9106	10000	0.3498	3.7238
3.6303	3.2014	11000	0.3521	3.7106
3.6281	3.4925	12000	0.3533	3.6952
3.6373	3.7837	13000	0.3550	3.6768
3.5357	4.0745	14000	0.3566	3.6674
3.5557	4.3656	15000	0.3576	3.6562
3.5669	4.6567	16000	0.3586	3.6433
3.5853	4.9478	17000	0.3599	3.6317
3.4977	5.2387	18000	0.3603	3.6349
3.5219	5.5298	19000	0.3615	3.6221
3.5313	5.8209	20000	0.3621	3.6114
3.4394	6.1118	21000	0.3627	3.6143
3.4737	6.4029	22000	0.3631	3.6091
3.4842	6.6940	23000	0.3639	3.5975
3.4892	6.9851	24000	0.3646	3.5906
3.4157	7.2760	25000	0.3649	3.5966
3.4418	7.5671	26000	0.3658	3.5879
3.4679	7.8582	27000	0.3663	3.5777
3.3723	8.1490	28000	0.3662	3.5915
3.4031	8.4401	29000	0.3668	3.5804
3.4338	8.7313	30000	0.3673	3.5734
3.3185	9.0221	31000	0.3675	3.5756
3.3706	9.3132	32000	0.3676	3.5767
3.3936	9.6043	33000	0.3683	3.5691
3.4075	9.8954	34000	0.3688	3.5609
3.3325	10.1863	35000	0.3683	3.5733
3.3581	10.4774	36000	0.3688	3.5633
3.3833	10.7685	37000	0.3696	3.5556
3.285	11.0594	38000	0.3694	3.5653
3.3224	11.3505	39000	0.3694	3.5632
3.3604	11.6416	40000	0.3701	3.5551
3.3628	11.9327	41000	0.3705	3.5464
3.2999	12.2236	42000	0.3699	3.5590
3.3231	12.5147	43000	0.3705	3.5551
3.3547	12.8058	44000	0.3712	3.5452
3.2648	13.0966	45000	0.3705	3.5610
3.2975	13.3878	46000	0.3708	3.5543
3.3175	13.6789	47000	0.3714	3.5488
3.3366	13.9700	48000	0.3718	3.5383
3.2692	14.2608	49000	0.3712	3.5540
3.2975	14.5519	50000	0.3713	3.5459
3.3335	14.8430	51000	0.3720	3.5389
3.2324	15.1339	52000	0.3715	3.5519
3.2762	15.4250	53000	0.3719	3.5478
3.2911	15.7161	54000	0.3723	3.5416
3.2668	16.0070	55000	0.3721	3.5483
3.2367	16.2981	56000	0.3726	3.5475
3.2896	16.5892	57000	0.3727	3.5411
3.297	16.8803	58000	0.3730	3.5337
3.2212	17.1712	59000	0.3724	3.5499
3.2642	17.4623	60000	0.3727	3.5459
3.2763	17.7534	61000	0.3733	3.5351
3.1743	18.0442	62000	0.3726	3.5479
3.2332	18.3354	63000	0.3729	3.5469
3.2574	18.6265	64000	0.3731	3.5431
3.2772	18.9176	65000	0.3741	3.5289
3.1937	19.2084	66000	0.3726	3.5514
3.2345	19.4995	67000	0.3730	3.5452
3.2655	19.7906	68000	0.3738	3.5352
3.1613	20.0815	69000	0.3734	3.5444
3.2076	20.3726	70000	0.3736	3.5433
3.2437	20.6637	71000	0.3736	3.5376
3.238	20.9548	72000	0.3743	3.5298
3.1931	21.2457	73000	0.3733	3.5475
3.2217	21.5368	74000	0.3739	3.5385
3.2377	21.8279	75000	0.3745	3.5331
3.17	22.1188	76000	0.3737	3.5462
3.1935	22.4099	77000	0.3738	3.5449
3.2287	22.7010	78000	0.3741	3.5329
3.236	22.9921	79000	0.3747	3.5293
3.1738	23.2830	80000	0.3739	3.5445
3.1691	23.5741	81000	3.5528	0.3736
3.1978	23.8652	82000	3.5400	0.3742
3.1533	24.1563	83000	3.5515	0.3741
3.1815	24.4474	84000	3.5429	0.3742
3.2038	24.7385	85000	3.5352	0.3749
3.113	25.0294	86000	3.5466	0.3744
3.155	25.3205	87000	3.5499	0.3738
3.1795	25.6116	88000	3.5379	0.3744
3.214	25.9027	89000	3.5332	0.3750
3.1365	26.1936	90000	3.5491	0.3741
3.1669	26.4847	91000	3.5440	0.3746
3.2009	26.7758	92000	3.5358	0.3750
3.1074	27.0667	93000	3.5515	0.3743
3.1497	27.3578	94000	3.5447	0.3744
3.1652	27.6489	95000	3.5380	0.3751
3.1905	27.9400	96000	3.5340	0.3753
3.1156	28.2308	97000	3.5488	0.3745
3.1602	28.5219	98000	3.5435	0.3747
3.1636	28.8131	99000	3.5377	0.3755
3.087	29.1039	100000	3.5500	0.3745

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_3591

exceptions_exp2_swap_take_to_hit_3591

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

craa/
exceptions_exp2_swap_take_to_hit_3591