learning_rate: 0.0006
train_batch_size: 16
eval_batch_size: 16
seed: 1032
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
4.8379	0.2911	1000	4.7599	0.2541
4.3455	0.5822	2000	4.2824	0.2998
4.1525	0.8733	3000	4.1004	0.3155
3.9829	1.1642	4000	3.9902	0.3251
3.9261	1.4553	5000	3.9198	0.3309
3.8802	1.7464	6000	3.8571	0.3372
3.7499	2.0373	7000	3.8173	0.3411
3.758	2.3284	8000	3.7834	0.3441
3.7391	2.6195	9000	3.7571	0.3470
3.7373	2.9106	10000	3.7282	0.3494
3.6382	3.2014	11000	3.7160	0.3515
3.6552	3.4925	12000	3.6971	0.3529
3.6496	3.7837	13000	3.6783	0.3551
3.5384	4.0745	14000	3.6719	0.3560
3.5807	4.3656	15000	3.6640	0.3567
3.5728	4.6567	16000	3.6494	0.3583
3.5816	4.9478	17000	3.6371	0.3595
3.495	5.2387	18000	3.6360	0.3600
3.5198	5.5298	19000	3.6278	0.3611
3.529	5.8209	20000	3.6145	0.3619
3.4542	6.1118	21000	3.6196	0.3623
3.4818	6.4029	22000	3.6106	0.3630
3.4861	6.6940	23000	3.6030	0.3639
3.5012	6.9851	24000	3.5943	0.3645
3.4178	7.2760	25000	3.6000	0.3643
3.4484	7.5671	26000	3.5905	0.3653
3.4747	7.8582	27000	3.5815	0.3663
3.3842	8.1490	28000	3.5903	0.3661
3.4075	8.4401	29000	3.5852	0.3667
3.4407	8.7313	30000	3.5746	0.3672
3.3161	9.0221	31000	3.5830	0.3673
3.3661	9.3132	32000	3.5791	0.3675
3.3948	9.6043	33000	3.5712	0.3679
3.4238	9.8954	34000	3.5636	0.3684
3.336	10.1863	35000	3.5762	0.3683
3.3666	10.4774	36000	3.5687	0.3687
3.374	10.7685	37000	3.5612	0.3693
3.2953	11.0594	38000	3.5668	0.3689
3.3363	11.3505	39000	3.5667	0.3694
3.3615	11.6416	40000	3.5575	0.3697
3.3574	11.9327	41000	3.5494	0.3704
3.3005	12.2236	42000	3.5651	0.3697
3.3298	12.5147	43000	3.5575	0.3702
3.3443	12.8058	44000	3.5477	0.3709
3.2702	13.0966	45000	3.5624	0.3703
3.3068	13.3878	46000	3.5598	0.3703
3.3192	13.6789	47000	3.5482	0.3710
3.3463	13.9700	48000	3.5433	0.3717
3.2773	14.2608	49000	3.5579	0.3711
3.3027	14.5519	50000	3.5523	0.3714
3.3313	14.8430	51000	3.5413	0.3719
3.2455	15.1339	52000	3.5559	0.3714
3.2742	15.4250	53000	3.5525	0.3716
3.2854	15.7161	54000	3.5423	0.3724
3.2618	16.0070	55000	3.5527	0.3717
3.2502	16.2981	56000	3.5531	0.3717
3.2838	16.5892	57000	3.5443	0.3723
3.2975	16.8803	58000	3.5346	0.3730
3.2313	17.1712	59000	3.5520	0.3722
3.2623	17.4623	60000	3.5462	0.3727
3.2786	17.7534	61000	3.5376	0.3729
3.1837	18.0442	62000	3.5489	0.3727
3.2436	18.3354	63000	3.5480	0.3724
3.2677	18.6265	64000	3.5385	0.3730
3.266	18.9176	65000	3.5319	0.3735
3.2097	19.2084	66000	3.5476	0.3726
3.2476	19.4995	67000	3.5425	0.3732
3.2678	19.7906	68000	3.5325	0.3740
3.1702	20.0815	69000	3.5517	0.3728
3.2117	20.3726	70000	3.5446	0.3735
3.2444	20.6637	71000	3.5372	0.3735
3.2475	20.9548	72000	3.5314	0.3738
3.201	21.2457	73000	3.5480	0.3733
3.2204	21.5368	74000	3.5420	0.3735
3.2265	21.8279	75000	3.5328	0.3743
3.1644	22.1188	76000	3.5493	0.3733
3.2045	22.4099	77000	3.5426	0.3739
3.2204	22.7010	78000	3.5344	0.3743
3.2282	22.9921	79000	3.5273	0.3747
3.1957	23.2830	80000	3.5452	0.3740
3.2039	23.5741	81000	3.5407	0.3743
3.2025	23.8652	82000	3.5342	0.3747
3.1471	24.1560	83000	3.5519	0.3738
3.1925	24.4471	84000	3.5447	0.3739
3.2042	24.7382	85000	3.5351	0.3746
3.1219	25.0291	86000	3.5467	0.3742
3.1676	25.3202	87000	3.5474	0.3742
3.1908	25.6113	88000	3.5414	0.3742
3.2054	25.9024	89000	3.5327	0.3750
3.1435	26.1933	90000	3.5488	0.3739
3.1746	26.4844	91000	3.5396	0.3746
3.1906	26.7755	92000	3.5349	0.3747
3.1077	27.0664	93000	3.5510	0.3742
3.1594	27.3575	94000	3.5452	0.3744
3.1725	27.6486	95000	3.5397	0.3751
3.1823	27.9397	96000	3.5326	0.3752
3.122	28.2306	97000	3.5497	0.3745
3.1507	28.5217	98000	3.5414	0.3749
3.178	28.8128	99000	3.5307	0.3755

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_push_1032

exceptions_exp2_swap_take_to_push_1032

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

craa/
exceptions_exp2_swap_take_to_push_1032