exceptions_exp2_swap_0.3_resemble_to_push_1032

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.5636
• Accuracy: 0.3686

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• 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
4.8324	0.2915	1000	4.7560	0.2543
4.3468	0.5830	2000	4.2877	0.2980
4.1576	0.8745	3000	4.1013	0.3147
3.9863	1.1659	4000	3.9972	0.3239
3.9393	1.4574	5000	3.9209	0.3307
3.896	1.7488	6000	3.8618	0.3361
3.7547	2.0402	7000	3.8213	0.3403
3.7519	2.3317	8000	3.7918	0.3432
3.7517	2.6232	9000	3.7602	0.3457
3.7294	2.9147	10000	3.7343	0.3488
3.6401	3.2061	11000	3.7210	0.3505
3.6521	3.4976	12000	3.7001	0.3525
3.652	3.7891	13000	3.6829	0.3540
3.5451	4.0805	14000	3.6772	0.3551
3.5695	4.3719	15000	3.6679	0.3562
3.5747	4.6634	16000	3.6538	0.3572
3.5794	4.9549	17000	3.6404	0.3586
3.5248	5.2463	18000	3.6428	0.3589
3.5307	5.5378	19000	3.6288	0.3603
3.538	5.8293	20000	3.6179	0.3612
3.4449	6.1207	21000	3.6247	0.3615
3.4754	6.4122	22000	3.6163	0.3622
3.4863	6.7037	23000	3.6046	0.3629
3.4935	6.9952	24000	3.5978	0.3637
3.4371	7.2865	25000	3.6079	0.3632
3.4699	7.5780	26000	3.5986	0.3642
3.461	7.8695	27000	3.5882	0.3648
3.3792	8.1609	28000	3.5962	0.3646
3.4134	8.4524	29000	3.5911	0.3654
3.4314	8.7439	30000	3.5793	0.3657
3.3346	9.0353	31000	3.5859	0.3660
3.3847	9.3268	32000	3.5859	0.3663
3.3988	9.6183	33000	3.5782	0.3668
3.4229	9.9098	34000	3.5664	0.3675
3.3353	10.2011	35000	3.5826	0.3667
3.3754	10.4926	36000	3.5737	0.3674
3.3957	10.7841	37000	3.5657	0.3682
3.2923	11.0755	38000	3.5758	0.3680
3.3542	11.3670	39000	3.5766	0.3678
3.3717	11.6585	40000	3.5636	0.3686
3.3707	11.9500	41000	3.5550	0.3693
3.2995	12.2414	42000	3.5730	0.3685
3.3294	12.5329	43000	3.5647	0.3691
3.3638	12.8243	44000	3.5542	0.3698
3.2567	13.1157	45000	3.5714	0.3690
3.3113	13.4072	46000	3.5660	0.3693
3.3393	13.6987	47000	3.5564	0.3701
3.3494	13.9902	48000	3.5480	0.3703
3.2949	14.2816	49000	3.5639	0.3698
3.3042	14.5731	50000	3.5588	0.3700
3.3265	14.8646	51000	3.5485	0.3706
3.2745	15.1559	52000	3.5623	0.3699
3.2852	15.4474	53000	3.5563	0.3704
3.3091	15.7389	54000	3.5496	0.3711
3.21	16.0303	55000	3.5611	0.3707
3.2604	16.3218	56000	3.5598	0.3706
3.2894	16.6133	57000	3.5503	0.3712
3.3052	16.9048	58000	3.5457	0.3717
3.2198	17.1962	59000	3.5618	0.3708
3.2678	17.4877	60000	3.5542	0.3713
3.2767	17.7792	61000	3.5460	0.3719
3.2081	18.0705	62000	3.5598	0.3712
3.2447	18.3620	63000	3.5573	0.3713
3.2645	18.6535	64000	3.5487	0.3719
3.2862	18.9450	65000	3.5391	0.3723
3.2203	19.2364	66000	3.5588	0.3715
3.2444	19.5279	67000	3.5501	0.3720
3.2739	19.8194	68000	3.5401	0.3725
3.2039	20.1108	69000	3.5625	0.3715
3.2235	20.4023	70000	3.5531	0.3720
3.24	20.6938	71000	3.5450	0.3724
3.2544	20.9853	72000	3.5379	0.3726
3.1944	21.2766	73000	3.5568	0.3721
3.2377	21.5681	74000	3.5514	0.3720
3.2404	21.8596	75000	3.5417	0.3729
3.1744	22.1510	76000	3.5577	0.3724
3.2105	22.4425	77000	3.5519	0.3724
3.2324	22.7340	78000	3.5451	0.3726
3.1422	23.0254	79000	3.5558	0.3724
3.193	23.3169	80000	3.5567	0.3723
3.2166	23.6083	81000	3.5482	0.3730
3.2288	23.8998	82000	3.5416	0.3736
3.1658	24.1912	83000	3.5549	0.3725
3.1867	24.4827	84000	3.5537	0.3727
3.2169	24.7742	85000	3.5450	0.3734
3.1325	25.0656	86000	3.5581	0.3728
3.1844	25.3571	87000	3.5580	0.3725
3.2056	25.6486	88000	3.5460	0.3735
3.2313	25.9401	89000	3.5400	0.3737
3.147	26.2314	90000	3.5608	0.3729
3.1914	26.5229	91000	3.5510	0.3733
3.1876	26.8144	92000	3.5428	0.3738

Framework versions

• Transformers 4.55.2
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.21.4