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	Validation Loss	Accuracy
4.8109	0.2915	1000	4.7588	0.2545
4.3409	0.5830	2000	4.2851	0.2990
4.1487	0.8745	3000	4.0986	0.3151
3.9935	1.1659	4000	3.9946	0.3244
3.9482	1.4574	5000	3.9187	0.3308
3.88	1.7488	6000	3.8616	0.3362
3.7453	2.0402	7000	3.8210	0.3407
3.7605	2.3317	8000	3.7884	0.3437
3.7404	2.6232	9000	3.7612	0.3464
3.7262	2.9147	10000	3.7332	0.3487
3.6415	3.2061	11000	3.7201	0.3506
3.6534	3.4976	12000	3.7031	0.3523
3.6462	3.7891	13000	3.6828	0.3539
3.5516	4.0805	14000	3.6781	0.3553
3.5821	4.3719	15000	3.6676	0.3564
3.568	4.6634	16000	3.6518	0.3578
3.5882	4.9549	17000	3.6409	0.3585
3.4946	5.2463	18000	3.6418	0.3591
3.533	5.5378	19000	3.6313	0.3602
3.5232	5.8293	20000	3.6190	0.3613
3.4486	6.1207	21000	3.6250	0.3615
3.4855	6.4122	22000	3.6169	0.3620
3.4814	6.7037	23000	3.6053	0.3629
3.5027	6.9952	24000	3.5974	0.3637
3.4279	7.2865	25000	3.6066	0.3637
3.4646	7.5780	26000	3.5969	0.3642
3.459	7.8695	27000	3.5860	0.3652
3.3867	8.1609	28000	3.5984	0.3645
3.433	8.4524	29000	3.5886	0.3655
3.433	8.7439	30000	3.5821	0.3661
3.3437	9.0353	31000	3.5859	0.3662
3.3821	9.3268	32000	3.5854	0.3663
3.4138	9.6183	33000	3.5769	0.3668
3.4159	9.9098	34000	3.5669	0.3676
3.3383	10.2011	35000	3.5799	0.3674
3.3616	10.4926	36000	3.5728	0.3676
3.3799	10.7841	37000	3.5630	0.3683
3.2901	11.0755	38000	3.5756	0.3678
3.3376	11.3670	39000	3.5725	0.3678
3.3722	11.6585	40000	3.5628	0.3689
3.3741	11.9500	41000	3.5549	0.3694
3.3114	12.2414	42000	3.5692	0.3689
3.3512	12.5329	43000	3.5621	0.3693
3.3496	12.8243	44000	3.5528	0.3700
3.268	13.1157	45000	3.5699	0.3692
3.3054	13.4072	46000	3.5612	0.3696
3.3365	13.6987	47000	3.5542	0.3701
3.3383	13.9902	48000	3.5467	0.3708
3.2765	14.2816	49000	3.5612	0.3702
3.3014	14.5731	50000	3.5549	0.3706
3.3378	14.8646	51000	3.5469	0.3710
3.2513	15.1559	52000	3.5629	0.3703
3.2827	15.4474	53000	3.5542	0.3710
3.3096	15.7389	54000	3.5491	0.3712
3.2072	16.0303	55000	3.5587	0.3707
3.2618	16.3218	56000	3.5567	0.3710
3.2877	16.6133	57000	3.5498	0.3713
3.306	16.9048	58000	3.5405	0.3721
3.2393	17.1962	59000	3.5580	0.3710
3.2617	17.4877	60000	3.5486	0.3718
3.2826	17.7792	61000	3.5424	0.3722
3.2007	18.0705	62000	3.5563	0.3716
3.2398	18.3620	63000	3.5540	0.3714
3.2587	18.6535	64000	3.5450	0.3721
3.2839	18.9450	65000	3.5357	0.3728
3.2095	19.2364	66000	3.5559	0.3719
3.2479	19.5279	67000	3.5470	0.3725
3.271	19.8194	68000	3.5406	0.3728
3.1877	20.1108	69000	3.5589	0.3720
3.2304	20.4023	70000	3.5502	0.3723
3.2535	20.6938	71000	3.5434	0.3728
3.2638	20.9853	72000	3.5362	0.3733
3.2027	21.2766	73000	3.5519	0.3726
3.2296	21.5681	74000	3.5439	0.3730
3.2565	21.8596	75000	3.5351	0.3734
3.1675	22.1510	76000	3.5585	0.3726
3.2115	22.4425	77000	3.5471	0.3729
3.2371	22.7340	78000	3.5407	0.3731
3.1285	23.0254	79000	3.5514	0.3730
3.188	23.3169	80000	3.5513	0.3726
3.2126	23.6083	81000	3.5469	0.3730
3.2242	23.8998	82000	3.5355	0.3737
3.1645	24.1912	83000	3.5555	0.3733
3.1905	24.4827	84000	3.5489	0.3735
3.2129	24.7742	85000	3.5382	0.3737
3.1326	25.0656	86000	3.5581	0.3730
3.1723	25.3571	87000	3.5494	0.3735
3.1961	25.6486	88000	3.5412	0.3740
3.2203	25.9401	89000	3.5363	0.3741
3.1438	26.2314	90000	3.5543	0.3732
3.1887	26.5229	91000	3.5443	0.3737
3.1887	26.8144	92000	3.5406	0.3741
3.123	27.1058	93000	3.5566	0.3734
3.1626	27.3973	94000	3.5523	0.3735
3.178	27.6888	95000	3.5431	0.3739

Framework versions

Transformers 4.55.2
Pytorch 2.8.0+cu128
Datasets 4.0.0
Tokenizers 0.21.4

craa/exceptions_exp2_swap_0.3_resemble_to_push_2128

exceptions_exp2_swap_0.3_resemble_to_push_2128

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

craa/
exceptions_exp2_swap_0.3_resemble_to_push_2128