adriansanz/te-zsc-hybrid_VIC

VICH_300524_epoch_3

This model is a fine-tuned version of projecte-aina/roberta-base-ca-v2-cased-te on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3866
• Accuracy: 0.954
• Precision: 0.9552
• Recall: 0.954
• F1: 0.9540
• Ratio: 0.474

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 47
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.06
• lr_scheduler_warmup_steps: 4
• num_epochs: 1
• label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Ratio
0.339	0.0157	10	0.4216	0.945	0.9453	0.9450	0.9450	0.487
0.3573	0.0314	20	0.4397	0.943	0.9430	0.943	0.9430	0.501
0.4019	0.0472	30	0.4330	0.945	0.9452	0.9450	0.9450	0.489
0.3443	0.0629	40	0.4368	0.942	0.9434	0.942	0.9420	0.472
0.3805	0.0786	50	0.4335	0.933	0.9331	0.933	0.9330	0.507
0.3837	0.0943	60	0.4273	0.938	0.9380	0.938	0.9380	0.498
0.3428	0.1101	70	0.4313	0.94	0.9403	0.94	0.9400	0.488
0.3954	0.1258	80	0.4323	0.945	0.9458	0.9450	0.9450	0.479
0.4144	0.1415	90	0.4299	0.94	0.9400	0.94	0.9400	0.502
0.3481	0.1572	100	0.4249	0.939	0.9391	0.9390	0.9390	0.491
0.3825	0.1730	110	0.4293	0.942	0.9420	0.942	0.9420	0.498
0.3605	0.1887	120	0.4130	0.949	0.9498	0.9490	0.9490	0.479
0.4028	0.2044	130	0.4105	0.948	0.9490	0.948	0.9480	0.476
0.3729	0.2201	140	0.4324	0.939	0.9391	0.9390	0.9390	0.507
0.3611	0.2358	150	0.4255	0.937	0.9371	0.937	0.9370	0.491
0.3683	0.2516	160	0.4290	0.943	0.9443	0.9430	0.9430	0.473
0.351	0.2673	170	0.4215	0.942	0.9426	0.942	0.9420	0.482
0.3697	0.2830	180	0.4280	0.944	0.9441	0.944	0.9440	0.492
0.3851	0.2987	190	0.4251	0.945	0.9461	0.9450	0.9450	0.475
0.335	0.3145	200	0.4276	0.945	0.9455	0.9450	0.9450	0.483
0.3744	0.3302	210	0.4173	0.947	0.9476	0.9470	0.9470	0.481
0.376	0.3459	220	0.4080	0.947	0.9478	0.9470	0.9470	0.479
0.3856	0.3616	230	0.4131	0.947	0.9472	0.9470	0.9470	0.489
0.4036	0.3774	240	0.4285	0.937	0.9370	0.937	0.9370	0.503
0.3863	0.3931	250	0.4159	0.939	0.9396	0.9390	0.9390	0.481
0.3619	0.4088	260	0.4212	0.944	0.9446	0.944	0.9440	0.482
0.4042	0.4245	270	0.4233	0.941	0.9411	0.9410	0.9410	0.493
0.3783	0.4403	280	0.4153	0.939	0.9390	0.9390	0.9390	0.505
0.3744	0.4560	290	0.4170	0.943	0.9447	0.9430	0.9429	0.469
0.4052	0.4717	300	0.4219	0.94	0.9423	0.94	0.9399	0.464
0.3531	0.4874	310	0.4049	0.949	0.9493	0.9490	0.9490	0.487
0.3812	0.5031	320	0.4042	0.951	0.9520	0.9510	0.9510	0.477
0.3587	0.5189	330	0.4030	0.95	0.9509	0.95	0.9500	0.478
0.3455	0.5346	340	0.4007	0.951	0.9512	0.951	0.9510	0.489
0.4174	0.5503	350	0.3989	0.952	0.9525	0.952	0.9520	0.484
0.4173	0.5660	360	0.4004	0.948	0.9487	0.948	0.9480	0.48
0.4012	0.5818	370	0.3956	0.95	0.9504	0.95	0.9500	0.486
0.388	0.5975	380	0.3968	0.949	0.9490	0.949	0.9490	0.495
0.3613	0.6132	390	0.3978	0.948	0.9482	0.948	0.9480	0.49
0.3699	0.6289	400	0.3988	0.956	0.9563	0.956	0.9560	0.488
0.3585	0.6447	410	0.3967	0.956	0.9569	0.956	0.9560	0.478
0.4017	0.6604	420	0.3888	0.959	0.9595	0.959	0.9590	0.483
0.3657	0.6761	430	0.3898	0.954	0.9541	0.954	0.9540	0.494
0.413	0.6918	440	0.3923	0.955	0.9550	0.955	0.9550	0.499
0.3977	0.7075	450	0.3884	0.955	0.9551	0.955	0.9550	0.491
0.4066	0.7233	460	0.3869	0.959	0.9593	0.959	0.9590	0.487
0.3908	0.7390	470	0.3878	0.956	0.9561	0.956	0.9560	0.492
0.4041	0.7547	480	0.3872	0.958	0.9584	0.958	0.9580	0.486
0.4191	0.7704	490	0.3945	0.952	0.9534	0.952	0.9520	0.472
0.3443	0.7862	500	0.3932	0.949	0.9500	0.9490	0.9490	0.477
0.3735	0.8019	510	0.3934	0.955	0.9552	0.955	0.9550	0.489
0.3913	0.8176	520	0.3965	0.954	0.9541	0.954	0.9540	0.494
0.4038	0.8333	530	0.3949	0.953	0.9531	0.953	0.9530	0.493
0.4055	0.8491	540	0.3933	0.952	0.9524	0.952	0.9520	0.486
0.4073	0.8648	550	0.3932	0.954	0.9546	0.954	0.9540	0.482
0.4471	0.8805	560	0.3944	0.952	0.9532	0.952	0.9520	0.474
0.4098	0.8962	570	0.3942	0.951	0.9525	0.9510	0.9510	0.471
0.4512	0.9119	580	0.3933	0.952	0.9534	0.952	0.9520	0.472
0.4309	0.9277	590	0.3914	0.952	0.9534	0.952	0.9520	0.472
0.3962	0.9434	600	0.3894	0.953	0.9543	0.9530	0.9530	0.473
0.4242	0.9591	610	0.3878	0.953	0.9543	0.9530	0.9530	0.473
0.3824	0.9748	620	0.3869	0.954	0.9552	0.954	0.9540	0.474
0.3837	0.9906	630	0.3867	0.954	0.9552	0.954	0.9540	0.474

Framework versions

• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1