learning_rate: 0.0006
train_batch_size: 16
eval_batch_size: 16
seed: 40817
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
4.8229	0.2911	1000	0.2562	4.7444
4.3387	0.5822	2000	0.2995	4.2779
4.1421	0.8733	3000	0.3150	4.0946
4.0011	1.1642	4000	0.3256	3.9845
3.9349	1.4553	5000	0.3325	3.9114
3.8834	1.7464	6000	0.3374	3.8550
3.7553	2.0373	7000	0.3418	3.8107
3.7429	2.3284	8000	0.3444	3.7823
3.7416	2.6195	9000	0.3475	3.7516
3.7265	2.9106	10000	0.3496	3.7270
3.6355	3.2014	11000	0.3514	3.7145
3.6461	3.4925	12000	0.3534	3.6955
3.6471	3.7837	13000	0.3549	3.6771
3.5387	4.0745	14000	0.3564	3.6716
3.5512	4.3656	15000	0.3574	3.6603
3.5805	4.6567	16000	0.3588	3.6463
3.5782	4.9478	17000	0.3598	3.6325
3.4969	5.2387	18000	0.3603	3.6347
3.5067	5.5298	19000	0.3612	3.6253
3.5315	5.8209	20000	0.3621	3.6135
3.4368	6.1118	21000	0.3626	3.6153
3.4725	6.4029	22000	0.3633	3.6110
3.4889	6.6940	23000	0.3639	3.6009
3.489	6.9851	24000	0.3646	3.5909
3.4371	7.2760	25000	0.3645	3.5968
3.4431	7.5671	26000	0.3654	3.5908
3.4503	7.8582	27000	0.3659	3.5824
3.3952	8.1490	28000	0.3658	3.5906
3.3972	8.4401	29000	0.3664	3.5836
3.4263	8.7313	30000	0.3670	3.5747
3.3175	9.0221	31000	0.3673	3.5780
3.3881	9.3132	32000	0.3675	3.5772
3.3932	9.6043	33000	0.3680	3.5691
3.4109	9.8954	34000	0.3684	3.5601
3.33	10.1863	35000	0.3682	3.5737
3.3633	10.4774	36000	0.3687	3.5660
3.3783	10.7685	37000	0.3690	3.5602
3.2953	11.0594	38000	0.3692	3.5657
3.3272	11.3505	39000	0.3694	3.5658
3.3718	11.6416	40000	0.3699	3.5543
3.3633	11.9327	41000	0.3708	3.5482
3.303	12.2236	42000	0.3701	3.5613
3.3363	12.5147	43000	0.3705	3.5520
3.3448	12.8058	44000	0.3711	3.5481
3.281	13.0966	45000	0.3701	3.5634
3.2969	13.3878	46000	0.3708	3.5544
3.3078	13.6789	47000	0.3712	3.5480
3.3452	13.9700	48000	0.3716	3.5407
3.2723	14.2608	49000	0.3708	3.5545
3.2978	14.5519	50000	0.3714	3.5502
3.3182	14.8430	51000	0.3719	3.5422
3.2376	15.1339	52000	0.3714	3.5547
3.2776	15.4250	53000	0.3713	3.5531
3.2995	15.7161	54000	0.3721	3.5434
3.2525	16.0070	55000	0.3718	3.5517
3.2397	16.2981	56000	0.3720	3.5485
3.275	16.5892	57000	0.3722	3.5435
3.2792	16.8803	58000	0.3732	3.5356
3.2255	17.1712	59000	0.3720	3.5534
3.2539	17.4623	60000	0.3726	3.5466
3.268	17.7534	61000	0.3732	3.5359
3.1823	18.0442	62000	0.3724	3.5507
3.2293	18.3354	63000	0.3725	3.5479
3.2601	18.6265	64000	0.3728	3.5405
3.2748	18.9176	65000	0.3735	3.5339
3.1997	19.2084	66000	0.3725	3.5535
3.2285	19.4995	67000	0.3730	3.5475
3.2483	19.7906	68000	0.3738	3.5345
3.169	20.0815	69000	0.3727	3.5512
3.2182	20.3726	70000	0.3731	3.5459
3.2214	20.6637	71000	0.3737	3.5368
3.2522	20.9548	72000	0.3742	3.5309
3.1799	21.2457	73000	0.3733	3.5489
3.2074	21.5368	74000	0.3737	3.5384
3.226	21.8279	75000	0.3742	3.5325
3.1699	22.1188	76000	0.3732	3.5515
3.1957	22.4099	77000	0.3735	3.5457
3.2121	22.7010	78000	0.3736	3.5377
3.2229	22.9921	79000	0.3746	3.5299
3.186	23.2830	80000	0.3737	3.5471
3.1854	23.5741	81000	3.5544	0.3735
3.2084	23.8652	82000	3.5420	0.3741
3.1437	24.1563	83000	3.5528	0.3733
3.181	24.4474	84000	3.5461	0.3740
3.2047	24.7385	85000	3.5368	0.3744
3.1092	25.0294	86000	3.5517	0.3739
3.1691	25.3205	87000	3.5475	0.3741
3.1872	25.6116	88000	3.5410	0.3745
3.2049	25.9027	89000	3.5314	0.3747
3.1415	26.1936	90000	3.5541	0.3740
3.1611	26.4847	91000	3.5421	0.3745
3.181	26.7758	92000	3.5336	0.3747
3.1229	27.0667	93000	3.5523	0.3743
3.1484	27.3578	94000	3.5493	0.3742
3.1802	27.6489	95000	3.5405	0.3750
3.1757	27.9400	96000	3.5311	0.3752
3.1271	28.2308	97000	3.5511	0.3742
3.1613	28.5219	98000	3.5413	0.3750
3.1738	28.8131	99000	3.5365	0.3752
3.0882	29.1039	100000	3.5471	0.3748

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_40817

exceptions_exp2_swap_take_to_carry_40817

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

craa/
exceptions_exp2_swap_take_to_carry_40817