exceptions_exp2_swap_0.7_cost_to_drop_40817

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

• Loss: 3.5830
• Accuracy: 0.3658

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: 40817
• 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
4.8184	0.2917	1000	4.7478	0.2553
4.3473	0.5834	2000	4.2848	0.2987
4.1596	0.8750	3000	4.1084	0.3145
4.0129	1.1665	4000	4.0006	0.3242
3.9429	1.4582	5000	3.9255	0.3306
3.8873	1.7499	6000	3.8691	0.3355
3.7647	2.0414	7000	3.8225	0.3401
3.763	2.3331	8000	3.7948	0.3430
3.7513	2.6248	9000	3.7645	0.3456
3.7221	2.9165	10000	3.7383	0.3482
3.6457	3.2080	11000	3.7258	0.3500
3.6473	3.4996	12000	3.7081	0.3521
3.6616	3.7913	13000	3.6859	0.3536
3.5566	4.0828	14000	3.6819	0.3545
3.5614	4.3745	15000	3.6728	0.3561
3.5796	4.6662	16000	3.6554	0.3572
3.5697	4.9579	17000	3.6436	0.3582
3.5117	5.2494	18000	3.6447	0.3587
3.5218	5.5411	19000	3.6331	0.3597
3.5452	5.8327	20000	3.6229	0.3608
3.4591	6.1243	21000	3.6283	0.3612
3.4787	6.4159	22000	3.6181	0.3620
3.5001	6.7076	23000	3.6074	0.3624
3.5091	6.9993	24000	3.5973	0.3633
3.4248	7.2908	25000	3.6069	0.3633
3.4779	7.5825	26000	3.5985	0.3640
3.4668	7.8742	27000	3.5904	0.3648
3.3781	8.1657	28000	3.6014	0.3644
3.4296	8.4574	29000	3.5900	0.3651
3.4409	8.7490	30000	3.5830	0.3658
3.3433	9.0405	31000	3.5899	0.3658
3.3915	9.3322	32000	3.5861	0.3663
3.413	9.6239	33000	3.5802	0.3666
3.4151	9.9156	34000	3.5717	0.3672
3.3474	10.2071	35000	3.5822	0.3668
3.3744	10.4988	36000	3.5738	0.3674
3.3951	10.7905	37000	3.5671	0.3680
3.3172	11.0820	38000	3.5784	0.3674
3.3308	11.3736	39000	3.5754	0.3677
3.3579	11.6653	40000	3.5676	0.3684
3.389	11.9570	41000	3.5581	0.3691
3.3305	12.2485	42000	3.5706	0.3686
3.3505	12.5402	43000	3.5632	0.3694
3.3589	12.8319	44000	3.5595	0.3693
3.2925	13.1234	45000	3.5715	0.3689
3.32	13.4151	46000	3.5656	0.3695
3.3364	13.7067	47000	3.5570	0.3698
3.35	13.9984	48000	3.5504	0.3704
3.2985	14.2899	49000	3.5642	0.3693
3.3254	14.5816	50000	3.5611	0.3701
3.3359	14.8733	51000	3.5495	0.3708
3.2541	15.1648	52000	3.5669	0.3700
3.2956	15.4565	53000	3.5610	0.3703
3.3168	15.7482	54000	3.5502	0.3708
3.2036	16.0397	55000	3.5631	0.3704
3.2651	16.3313	56000	3.5599	0.3706
3.2903	16.6230	57000	3.5526	0.3708
3.3091	16.9147	58000	3.5443	0.3716
3.235	17.2062	59000	3.5602	0.3709
3.2684	17.4979	60000	3.5525	0.3712
3.2887	17.7896	61000	3.5461	0.3718
3.1944	18.0811	62000	3.5594	0.3711
3.2445	18.3728	63000	3.5568	0.3714
3.2471	18.6644	64000	3.5500	0.3717
3.2874	18.9561	65000	3.5421	0.3720
3.218	19.2476	66000	3.5581	0.3714
3.2456	19.5393	67000	3.5501	0.3720
3.2652	19.8310	68000	3.5434	0.3721
3.191	20.1225	69000	3.5614	0.3715
3.2083	20.4142	70000	3.5561	0.3716
3.2544	20.7059	71000	3.5469	0.3720
3.2716	20.9975	72000	3.5387	0.3730
3.2056	21.2891	73000	3.5570	0.3721
3.2282	21.5807	74000	3.5483	0.3725
3.2589	21.8724	75000	3.5411	0.3728
3.1828	22.1639	76000	3.5627	0.3720
3.2154	22.4556	77000	3.5537	0.3723
3.218	22.7473	78000	3.5466	0.3730
3.1507	23.0388	79000	3.5588	0.3723
3.2061	23.3305	80000	3.5590	0.3724
3.2085	23.6222	81000	3.5484	0.3727
3.217	23.9138	82000	3.5393	0.3736
3.1711	24.2053	83000	3.5596	0.3723
3.1897	24.4970	84000	3.5539	0.3730
3.2123	24.7887	85000	3.5454	0.3733
3.1405	25.0802	86000	3.5617	0.3727
3.185	25.3719	87000	3.5550	0.3729
3.1982	25.6636	88000	3.5531	0.3730
3.2174	25.9553	89000	3.5414	0.3737
3.1537	26.2468	90000	3.5583	0.3729
3.1752	26.5384	91000	3.5479	0.3735
3.1962	26.8301	92000	3.5430	0.3736

Framework versions

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