exceptions_exp2_swap_take_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: 3.5536
• Accuracy: 0.3703

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	Accuracy	Validation Loss
4.8463	0.2911	1000	0.2518	4.7746
4.3336	0.5822	2000	0.3006	4.2743
4.1408	0.8733	3000	0.3163	4.0876
3.9891	1.1642	4000	0.3262	3.9855
3.93	1.4553	5000	0.3321	3.9111
3.86	1.7464	6000	0.3377	3.8529
3.7272	2.0373	7000	0.3423	3.8075
3.7393	2.3284	8000	0.3449	3.7773
3.7256	2.6195	9000	0.3482	3.7499
3.7108	2.9106	10000	0.3500	3.7224
3.6259	3.2014	11000	0.3523	3.7093
3.6231	3.4925	12000	0.3541	3.6914
3.6336	3.7837	13000	0.3554	3.6754
3.5311	4.0745	14000	0.3570	3.6656
3.5513	4.3656	15000	0.3578	3.6559
3.5646	4.6567	16000	0.3593	3.6399
3.5808	4.9478	17000	0.3604	3.6301
3.4931	5.2387	18000	0.3606	3.6324
3.5171	5.5298	19000	0.3619	3.6208
3.5277	5.8209	20000	0.3626	3.6093
3.4341	6.1118	21000	0.3629	3.6124
3.4688	6.4029	22000	0.3634	3.6052
3.48	6.6940	23000	0.3644	3.5955
3.4861	6.9851	24000	0.3646	3.5911
3.411	7.2760	25000	0.3650	3.5957
3.4386	7.5671	26000	0.3660	3.5860
3.4647	7.8582	27000	0.3667	3.5755
3.3686	8.1490	28000	0.3663	3.5909
3.3989	8.4401	29000	0.3669	3.5805
3.4298	8.7313	30000	0.3678	3.5703
3.3158	9.0221	31000	0.3677	3.5741
3.3681	9.3132	32000	0.3680	3.5758
3.3907	9.6043	33000	0.3687	3.5656
3.4051	9.8954	34000	0.3691	3.5590
3.3306	10.1863	35000	0.3686	3.5711
3.3547	10.4774	36000	0.3690	3.5614
3.3802	10.7685	37000	0.3697	3.5539
3.2822	11.0594	38000	0.3699	3.5616
3.3188	11.3505	39000	0.3696	3.5627
3.3563	11.6416	40000	0.3703	3.5536
3.3593	11.9327	41000	0.3707	3.5443
3.2966	12.2236	42000	0.3703	3.5576
3.3187	12.5147	43000	0.3708	3.5530
3.352	12.8058	44000	0.3713	3.5424
3.2602	13.0966	45000	0.3708	3.5602
3.2942	13.3878	46000	0.3711	3.5536
3.3142	13.6789	47000	0.3719	3.5471
3.3334	13.9700	48000	0.3722	3.5349
3.2664	14.2608	49000	0.3716	3.5524
3.2924	14.5519	50000	0.3719	3.5445
3.3308	14.8430	51000	0.3725	3.5368
3.2279	15.1339	52000	0.3719	3.5488
3.2732	15.4250	53000	0.3721	3.5472
3.2879	15.7161	54000	0.3726	3.5395
3.263	16.0070	55000	0.3726	3.5459
3.2324	16.2981	56000	0.3728	3.5463
3.2843	16.5892	57000	0.3728	3.5399
3.2933	16.8803	58000	0.3729	3.5340
3.216	17.1712	59000	0.3724	3.5486
3.2597	17.4623	60000	0.3730	3.5418
3.2727	17.7534	61000	0.3735	3.5318
3.1711	18.0442	62000	0.3730	3.5444
3.2299	18.3354	63000	0.3731	3.5480
3.2554	18.6265	64000	0.3734	3.5407
3.2742	18.9176	65000	0.3745	3.5268
3.1898	19.2084	66000	0.3731	3.5474
3.2293	19.4995	67000	0.3736	3.5403
3.2629	19.7906	68000	0.3743	3.5312
3.1573	20.0815	69000	0.3738	3.5430
3.203	20.3726	70000	0.3742	3.5421
3.2396	20.6637	71000	0.3741	3.5339
3.2357	20.9548	72000	0.3745	3.5272
3.1883	21.2457	73000	0.3736	3.5450
3.2185	21.5368	74000	0.3741	3.5367
3.2349	21.8279	75000	0.3748	3.5296
3.1664	22.1188	76000	0.3741	3.5440
3.1893	22.4099	77000	0.3744	3.5426
3.2253	22.7010	78000	0.3745	3.5304
3.2324	22.9921	79000	0.3752	3.5256
3.1713	23.2830	80000	0.3747	3.5399
3.1653	23.5741	81000	3.5502	0.3738
3.1936	23.8652	82000	3.5356	0.3747
3.1505	24.1563	83000	3.5477	0.3740
3.1793	24.4474	84000	3.5407	0.3746
3.2001	24.7385	85000	3.5334	0.3752
3.1085	25.0294	86000	3.5439	0.3748
3.1519	25.3205	87000	3.5457	0.3744
3.174	25.6116	88000	3.5338	0.3752
3.2105	25.9027	89000	3.5291	0.3754
3.1325	26.1936	90000	3.5453	0.3748
3.1623	26.4847	91000	3.5385	0.3750
3.1974	26.7758	92000	3.5323	0.3756
3.1026	27.0667	93000	3.5488	0.3747
3.1455	27.3578	94000	3.5403	0.3749
3.1615	27.6489	95000	3.5330	0.3756
3.1869	27.9400	96000	3.5281	0.3759
3.111	28.2308	97000	3.5463	0.3752
3.1554	28.5219	98000	3.5373	0.3754
3.1586	28.8131	99000	3.5319	0.3758
3.0831	29.1039	100000	3.5465	0.3750

Framework versions

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