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
4.8351	0.2911	1000	0.2533	4.7595
4.3332	0.5822	2000	0.2992	4.2836
4.1446	0.8733	3000	0.3156	4.0942
3.9902	1.1642	4000	0.3255	3.9922
3.9329	1.4553	5000	0.3322	3.9161
3.8805	1.7464	6000	0.3376	3.8551
3.7509	2.0373	7000	0.3414	3.8146
3.7633	2.3284	8000	0.3446	3.7835
3.7467	2.6195	9000	0.3472	3.7567
3.7298	2.9106	10000	0.3496	3.7278
3.6306	3.2014	11000	0.3518	3.7155
3.6507	3.4925	12000	0.3532	3.6985
3.6359	3.7837	13000	0.3547	3.6803
3.5338	4.0745	14000	0.3564	3.6750
3.5699	4.3656	15000	0.3574	3.6611
3.5807	4.6567	16000	0.3584	3.6492
3.5658	4.9478	17000	0.3598	3.6354
3.5035	5.2387	18000	0.3599	3.6377
3.5172	5.5298	19000	0.3610	3.6265
3.5299	5.8209	20000	0.3618	3.6148
3.4448	6.1118	21000	0.3621	3.6224
3.4685	6.4029	22000	0.3632	3.6116
3.4898	6.6940	23000	0.3641	3.6013
3.497	6.9851	24000	0.3647	3.5929
3.429	7.2760	25000	0.3644	3.6009
3.4463	7.5671	26000	0.3655	3.5917
3.4721	7.8582	27000	0.3659	3.5835
3.3781	8.1490	28000	0.3658	3.5938
3.4187	8.4401	29000	0.3662	3.5888
3.4307	8.7313	30000	0.3668	3.5774
3.3313	9.0221	31000	0.3671	3.5834
3.381	9.3132	32000	0.3672	3.5804
3.3949	9.6043	33000	0.3678	3.5717
3.4252	9.8954	34000	0.3684	3.5640
3.3332	10.1863	35000	0.3681	3.5742
3.3691	10.4774	36000	0.3685	3.5689
3.3902	10.7685	37000	0.3691	3.5608
3.2798	11.0594	38000	0.3685	3.5739
3.3447	11.3505	39000	0.3692	3.5687
3.3642	11.6416	40000	0.3693	3.5595
3.3794	11.9327	41000	0.3702	3.5497
3.3121	12.2236	42000	0.3696	3.5652
3.3329	12.5147	43000	0.3700	3.5575
3.3549	12.8058	44000	0.3706	3.5505
3.2667	13.0966	45000	0.3701	3.5633
3.3055	13.3878	46000	0.3704	3.5585
3.3339	13.6789	47000	0.3708	3.5534
3.3376	13.9700	48000	0.3714	3.5433
3.274	14.2608	49000	0.3708	3.5640
3.3094	14.5519	50000	0.3714	3.5523
3.3217	14.8430	51000	0.3717	3.5449
3.2321	15.1339	52000	0.3711	3.5587
3.2893	15.4250	53000	0.3714	3.5515
3.2929	15.7161	54000	0.3722	3.5446
3.2636	16.0070	55000	0.3715	3.5539
3.2523	16.2981	56000	0.3717	3.5527
3.2764	16.5892	57000	0.3725	3.5444
3.3061	16.8803	58000	0.3727	3.5372
3.224	17.1712	59000	0.3721	3.5531
3.2699	17.4623	60000	0.3722	3.5469
3.297	17.7534	61000	0.3730	3.5418
3.1969	18.0442	62000	0.3722	3.5533
3.2334	18.3354	63000	0.3727	3.5494
3.2604	18.6265	64000	0.3729	3.5418
3.2766	18.9176	65000	0.3735	3.5343
3.2139	19.2084	66000	0.3725	3.5524
3.2557	19.4995	67000	0.3731	3.5424
3.2476	19.7906	68000	0.3733	3.5369
3.1767	20.0815	69000	0.3727	3.5504
3.2161	20.3726	70000	0.3731	3.5473
3.241	20.6637	71000	0.3739	3.5386
3.2534	20.9548	72000	0.3742	3.5319
3.1933	21.2457	73000	0.3730	3.5478
3.2142	21.5368	74000	0.3739	3.5424
3.2314	21.8279	75000	0.3741	3.5368
3.1578	22.1188	76000	0.3736	3.5522
3.2013	22.4099	77000	0.3736	3.5459
3.223	22.7010	78000	0.3743	3.5389
3.2453	22.9921	79000	0.3748	3.5277
3.1842	23.2830	80000	0.3741	3.5483
3.1701	23.5741	81000	3.5517	0.3735
3.1995	23.8652	82000	3.5443	0.3741
3.1577	24.1563	83000	3.5534	0.3739
3.1889	24.4474	84000	3.5464	0.3739
3.2153	24.7385	85000	3.5373	0.3744
3.1219	25.0294	86000	3.5475	0.3742
3.1659	25.3205	87000	3.5470	0.3742
3.1792	25.6116	88000	3.5413	0.3745
3.2081	25.9027	89000	3.5337	0.3752
3.1347	26.1936	90000	3.5497	0.3739
3.1603	26.4847	91000	3.5438	0.3748
3.1966	26.7758	92000	3.5364	0.3750
3.1094	27.0667	93000	3.5503	0.3742
3.1524	27.3578	94000	3.5450	0.3747
3.1691	27.6489	95000	3.5395	0.3747
3.1875	27.9400	96000	3.5334	0.3755
3.1253	28.2308	97000	3.5554	0.3742
3.148	28.5219	98000	3.5439	0.3751
3.1672	28.8131	99000	3.5365	0.3754
3.1046	29.1039	100000	3.5502	0.3743

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_carry_2128

exceptions_exp2_swap_take_to_carry_2128

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

craa/
exceptions_exp2_swap_take_to_carry_2128