exceptions_exp2_swap_0.7_last_to_drop_3591

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

• Loss: 3.5609
• Accuracy: 0.3691

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: 3591
• 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
4.8388	0.2915	1000	4.7578	0.2535
4.3478	0.5830	2000	4.2855	0.2989
4.157	0.8745	3000	4.0979	0.3146
3.9906	1.1659	4000	3.9884	0.3251
3.937	1.4574	5000	3.9151	0.3312
3.8829	1.7489	6000	3.8565	0.3367
3.7467	2.0402	7000	3.8155	0.3413
3.7417	2.3317	8000	3.7845	0.3440
3.7389	2.6233	9000	3.7546	0.3468
3.7172	2.9148	10000	3.7280	0.3491
3.6443	3.2061	11000	3.7169	0.3507
3.6411	3.4976	12000	3.6980	0.3528
3.6373	3.7891	13000	3.6805	0.3546
3.5441	4.0805	14000	3.6747	0.3557
3.5644	4.3720	15000	3.6636	0.3568
3.5761	4.6635	16000	3.6487	0.3580
3.5796	4.9550	17000	3.6337	0.3592
3.4961	5.2463	18000	3.6374	0.3596
3.53	5.5378	19000	3.6274	0.3605
3.5487	5.8293	20000	3.6170	0.3614
3.4432	6.1207	21000	3.6194	0.3620
3.4774	6.4122	22000	3.6129	0.3625
3.5039	6.7037	23000	3.6019	0.3632
3.4949	6.9952	24000	3.5935	0.3640
3.4198	7.2866	25000	3.6017	0.3639
3.4506	7.5781	26000	3.5922	0.3649
3.4539	7.8696	27000	3.5842	0.3654
3.3845	8.1609	28000	3.5917	0.3651
3.4229	8.4524	29000	3.5871	0.3657
3.427	8.7439	30000	3.5789	0.3663
3.3437	9.0353	31000	3.5805	0.3664
3.3645	9.3268	32000	3.5813	0.3669
3.4027	9.6183	33000	3.5726	0.3672
3.415	9.9098	34000	3.5660	0.3677
3.3542	10.2011	35000	3.5773	0.3675
3.3466	10.4927	36000	3.5702	0.3680
3.3837	10.7842	37000	3.5624	0.3685
3.2855	11.0755	38000	3.5741	0.3682
3.3507	11.3670	39000	3.5697	0.3686
3.3575	11.6585	40000	3.5609	0.3691
3.3786	11.9500	41000	3.5518	0.3697
3.3098	12.2414	42000	3.5647	0.3692
3.3406	12.5329	43000	3.5618	0.3694
3.3565	12.8244	44000	3.5519	0.3703
3.2693	13.1157	45000	3.5663	0.3695
3.2989	13.4072	46000	3.5610	0.3698
3.3236	13.6988	47000	3.5536	0.3704
3.351	13.9903	48000	3.5458	0.3710
3.2744	14.2816	49000	3.5629	0.3702
3.3092	14.5731	50000	3.5530	0.3705
3.3219	14.8646	51000	3.5444	0.3712
3.2617	15.1560	52000	3.5596	0.3705
3.2826	15.4475	53000	3.5540	0.3709
3.3123	15.7390	54000	3.5439	0.3717
3.2134	16.0303	55000	3.5556	0.3713
3.2565	16.3218	56000	3.5546	0.3712
3.2767	16.6133	57000	3.5484	0.3715
3.2951	16.9049	58000	3.5417	0.3719
3.2398	17.1962	59000	3.5576	0.3715
3.2753	17.4877	60000	3.5508	0.3717
3.2858	17.7792	61000	3.5451	0.3722
3.2019	18.0705	62000	3.5582	0.3717
3.2363	18.3621	63000	3.5541	0.3717
3.2614	18.6536	64000	3.5454	0.3722
3.281	18.9451	65000	3.5391	0.3727
3.2056	19.2364	66000	3.5543	0.3717
3.2467	19.5279	67000	3.5483	0.3724
3.2504	19.8194	68000	3.5409	0.3729
3.1899	20.1108	69000	3.5554	0.3719
3.2163	20.4023	70000	3.5504	0.3725
3.2439	20.6938	71000	3.5440	0.3728
3.2564	20.9853	72000	3.5380	0.3736
3.1984	21.2766	73000	3.5525	0.3723
3.2252	21.5682	74000	3.5454	0.3731
3.2344	21.8597	75000	3.5363	0.3732
3.1737	22.1510	76000	3.5528	0.3728
3.2042	22.4425	77000	3.5488	0.3729
3.2293	22.7340	78000	3.5382	0.3735
3.1331	23.0254	79000	3.5525	0.3732
3.1957	23.3169	80000	3.5520	0.3730
3.2142	23.6084	81000	3.5435	0.3734
3.2258	23.8999	82000	3.5337	0.3740
3.1566	24.1912	83000	3.5521	0.3730
3.2009	24.4827	84000	3.5447	0.3735
3.2057	24.7743	85000	3.5404	0.3739
3.126	25.0656	86000	3.5555	0.3730
3.1603	25.3571	87000	3.5515	0.3732
3.1926	25.6486	88000	3.5428	0.3737
3.1977	25.9401	89000	3.5365	0.3743
3.1318	26.2315	90000	3.5532	0.3732
3.1662	26.5230	91000	3.5454	0.3739
3.1828	26.8145	92000	3.5381	0.3745
3.1267	27.1058	93000	3.5543	0.3735
3.1531	27.3973	94000	3.5517	0.3734
3.1719	27.6888	95000	3.5452	0.3740
3.1786	27.9804	96000	3.5388	0.3745
3.1445	28.2717	97000	3.5563	0.3737
3.1572	28.5632	98000	3.5468	0.3738
3.1762	28.8547	99000	3.5402	0.3745
3.1162	29.1460	100000	3.5538	0.3736
3.1346	29.4376	101000	3.5516	0.3739
3.1491	29.7291	102000	3.5447	0.3746

Framework versions

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