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	Accuracy	Validation Loss
4.8325	0.2911	1000	0.2553	4.7510
4.3359	0.5822	2000	0.2999	4.2762
4.1416	0.8733	3000	0.3161	4.0896
3.9903	1.1642	4000	0.3256	3.9898
3.9312	1.4553	5000	0.3322	3.9140
3.8662	1.7464	6000	0.3373	3.8591
3.7311	2.0373	7000	0.3418	3.8123
3.7411	2.3284	8000	0.3448	3.7799
3.7298	2.6195	9000	0.3478	3.7527
3.7148	2.9106	10000	0.3498	3.7256
3.6305	3.2014	11000	0.3518	3.7126
3.6272	3.4925	12000	0.3534	3.6953
3.6364	3.7837	13000	0.3550	3.6790
3.5361	4.0745	14000	0.3565	3.6696
3.5551	4.3656	15000	0.3575	3.6585
3.5684	4.6567	16000	0.3586	3.6431
3.5845	4.9478	17000	0.3600	3.6317
3.4981	5.2387	18000	0.3604	3.6357
3.522	5.5298	19000	0.3617	3.6237
3.5315	5.8209	20000	0.3622	3.6114
3.4396	6.1118	21000	0.3625	3.6167
3.4737	6.4029	22000	0.3631	3.6094
3.4844	6.6940	23000	0.3638	3.5994
3.4886	6.9851	24000	0.3646	3.5904
3.4154	7.2760	25000	0.3647	3.5981
3.4436	7.5671	26000	0.3656	3.5884
3.4684	7.8582	27000	0.3663	3.5802
3.3719	8.1490	28000	0.3662	3.5917
3.4039	8.4401	29000	0.3666	3.5833
3.4337	8.7313	30000	0.3674	3.5726
3.3195	9.0221	31000	0.3671	3.5799
3.3716	9.3132	32000	0.3676	3.5785
3.3949	9.6043	33000	0.3683	3.5694
3.4079	9.8954	34000	0.3687	3.5603
3.3343	10.1863	35000	0.3681	3.5758
3.3587	10.4774	36000	0.3685	3.5666
3.3837	10.7685	37000	0.3696	3.5559
3.2852	11.0594	38000	0.3693	3.5653
3.3218	11.3505	39000	0.3692	3.5630
3.3601	11.6416	40000	0.3700	3.5548
3.3635	11.9327	41000	0.3705	3.5502
3.3012	12.2236	42000	0.3697	3.5638
3.3241	12.5147	43000	0.3706	3.5579
3.3556	12.8058	44000	0.3709	3.5453
3.2655	13.0966	45000	0.3698	3.5650
3.2968	13.3878	46000	0.3704	3.5577
3.3187	13.6789	47000	0.3713	3.5519
3.3375	13.9700	48000	0.3719	3.5398
3.2698	14.2608	49000	0.3713	3.5554
3.2972	14.5519	50000	0.3713	3.5488
3.3336	14.8430	51000	0.3720	3.5402
3.2331	15.1339	52000	0.3716	3.5536
3.2771	15.4250	53000	0.3719	3.5498
3.2915	15.7161	54000	0.3724	3.5442
3.2661	16.0070	55000	0.3719	3.5528
3.2368	16.2981	56000	0.3720	3.5519
3.2882	16.5892	57000	0.3724	3.5435
3.2965	16.8803	58000	0.3725	3.5393
3.2206	17.1712	59000	0.3721	3.5540
3.2628	17.4623	60000	0.3724	3.5476
3.2436	17.7534	61000	3.5562	0.3723
3.1811	18.0445	62000	3.5583	0.3718
3.2419	18.3356	63000	3.5560	0.3722
3.2653	18.6267	64000	3.5440	0.3728
3.28	18.9179	65000	3.5330	0.3737
3.1933	19.2087	66000	3.5538	0.3723
3.234	19.4998	67000	3.5471	0.3727
3.2667	19.7909	68000	3.5383	0.3738
3.1628	20.0818	69000	3.5508	0.3731
3.2074	20.3729	70000	3.5517	0.3731
3.243	20.6640	71000	3.5414	0.3734
3.2376	20.9551	72000	3.5328	0.3738
3.1937	21.2460	73000	3.5500	0.3731
3.2227	21.5371	74000	3.5424	0.3735
3.2392	21.8282	75000	3.5389	0.3740
3.1701	22.1191	76000	3.5552	0.3731
3.1958	22.4102	77000	3.5492	0.3738
3.231	22.7013	78000	3.5374	0.3742
3.2364	22.9924	79000	3.5313	0.3747
3.1744	23.2832	80000	3.5479	0.3737
3.2056	23.5743	81000	3.5429	0.3742
3.2089	23.8655	82000	3.5328	0.3747
3.149	24.1563	83000	3.5508	0.3736
3.1794	24.4474	84000	3.5449	0.3740
3.2038	24.7385	85000	3.5382	0.3745
3.1107	25.0294	86000	3.5536	0.3740
3.1545	25.3205	87000	3.5545	0.3736
3.1766	25.6116	88000	3.5426	0.3743
3.2136	25.9027	89000	3.5395	0.3747
3.1363	26.1936	90000	3.5511	0.3740
3.1658	26.4847	91000	3.5480	0.3744
3.2013	26.7758	92000	3.5393	0.3747
3.1076	27.0667	93000	3.5563	0.3737
3.1504	27.3578	94000	3.5444	0.3743
3.1651	27.6489	95000	3.5419	0.3747
3.189	27.9400	96000	3.5363	0.3752
3.116	28.2308	97000	3.5534	0.3742
3.1598	28.5219	98000	3.5438	0.3745
3.1622	28.8131	99000	3.5417	0.3750

Framework versions

Transformers 4.55.2
Pytorch 2.8.0+cu128
Datasets 4.0.0
Tokenizers 0.21.4

craa/exceptions_exp2_swap_require_to_drop_3591

exceptions_exp2_swap_require_to_drop_3591

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

craa/
exceptions_exp2_swap_require_to_drop_3591