metadata

license: mit
base_model: neuralmind/bert-base-portuguese-cased
tags:
  - generated_from_trainer
metrics:
  - f1
  - accuracy
model-index:
  - name: finetuned-bert-categories-estimation
    results: []

finetuned-bert-categories-estimation

This model is a fine-tuned version of neuralmind/bert-base-portuguese-cased on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.4080
F1: 0.9054
Accuracy: 0.9277

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: 5e-05
train_batch_size: 32
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Accuracy
4.3517	0.13	100	3.6120	0.0934	0.3599
3.2593	0.25	200	2.7209	0.1848	0.5085
2.584	0.38	300	2.1874	0.2784	0.5952
2.1208	0.51	400	1.8393	0.3656	0.6504
1.7726	0.63	500	1.5560	0.4633	0.7121
1.5799	0.76	600	1.3686	0.5185	0.7442
1.3384	0.89	700	1.2079	0.5896	0.7759
1.2751	1.01	800	1.0736	0.6044	0.7941
1.0223	1.14	900	0.9892	0.6353	0.8089
0.9095	1.27	1000	0.9277	0.6699	0.8157
0.8496	1.39	1100	0.8370	0.6973	0.8311
0.7735	1.52	1200	0.7878	0.7093	0.8349
0.7058	1.65	1300	0.7299	0.7239	0.8481
0.6545	1.77	1400	0.6823	0.7444	0.8563
0.6652	1.9	1500	0.6623	0.7547	0.8609
0.5905	2.03	1600	0.6079	0.7660	0.8663
0.4679	2.15	1700	0.5910	0.7867	0.8696
0.4415	2.28	1800	0.5668	0.8034	0.8785
0.4377	2.41	1900	0.5580	0.8068	0.8796
0.4262	2.53	2000	0.5366	0.8054	0.8815
0.4272	2.66	2100	0.5094	0.8189	0.8880
0.3979	2.79	2200	0.4966	0.8229	0.8898
0.3763	2.92	2300	0.4838	0.8349	0.8950
0.366	3.04	2400	0.4742	0.8340	0.8950
0.2686	3.17	2500	0.4591	0.8365	0.8966
0.2735	3.3	2600	0.4676	0.8393	0.8958
0.2582	3.42	2700	0.4263	0.8580	0.9025
0.2451	3.55	2800	0.4383	0.8526	0.8988
0.2626	3.68	2900	0.4420	0.8554	0.9018
0.248	3.8	3000	0.4153	0.8658	0.9080
0.2634	3.93	3100	0.4082	0.8666	0.9088
0.2	4.06	3200	0.4162	0.8716	0.9090
0.1717	4.18	3300	0.4032	0.8748	0.9117
0.19	4.31	3400	0.4019	0.8747	0.9117
0.1507	4.44	3500	0.4118	0.8789	0.9139
0.16	4.56	3600	0.4107	0.8815	0.9139
0.1716	4.69	3700	0.4105	0.8826	0.9132
0.1545	4.82	3800	0.3945	0.8850	0.9180
0.1628	4.94	3900	0.3974	0.8907	0.9194
0.1123	5.07	4000	0.4060	0.8828	0.9166
0.0988	5.2	4100	0.4037	0.8847	0.9167
0.1065	5.32	4200	0.3959	0.8895	0.9201
0.1018	5.45	4300	0.4040	0.8875	0.9183
0.1091	5.58	4400	0.4044	0.8908	0.9199
0.1041	5.7	4500	0.3937	0.8943	0.9218
0.1154	5.83	4600	0.3981	0.8956	0.9205
0.0932	5.96	4700	0.3940	0.8967	0.9223
0.0835	6.08	4800	0.3914	0.8967	0.9224
0.065	6.21	4900	0.3905	0.8922	0.9215
0.0634	6.34	5000	0.3999	0.8924	0.9215
0.0618	6.46	5100	0.4013	0.8966	0.9226
0.0678	6.59	5200	0.3985	0.9004	0.9242
0.0666	6.72	5300	0.3892	0.8993	0.9239
0.0564	6.84	5400	0.4026	0.8986	0.9228
0.0704	6.97	5500	0.4011	0.9004	0.9236
0.0508	7.1	5600	0.4035	0.8987	0.9234
0.0395	7.22	5700	0.4131	0.8979	0.9216
0.0363	7.35	5800	0.4112	0.9022	0.9243
0.0443	7.48	5900	0.4079	0.9039	0.9251
0.0383	7.6	6000	0.4152	0.9031	0.9248
0.0384	7.73	6100	0.4075	0.9037	0.9258
0.0414	7.86	6200	0.4087	0.9035	0.9256
0.0449	7.98	6300	0.4066	0.9060	0.9262
0.0246	8.11	6400	0.4091	0.9041	0.9258
0.0275	8.24	6500	0.4085	0.9035	0.9262
0.0256	8.37	6600	0.4077	0.9048	0.9269
0.0307	8.49	6700	0.4040	0.9082	0.9285
0.0294	8.62	6800	0.4057	0.9067	0.9283
0.0226	8.75	6900	0.4069	0.9054	0.9274
0.0218	8.87	7000	0.4090	0.9053	0.9278
0.0333	9.0	7100	0.4053	0.9075	0.9286
0.0182	9.13	7200	0.4071	0.9063	0.9277
0.0176	9.25	7300	0.4058	0.9053	0.9278
0.0187	9.38	7400	0.4074	0.9060	0.9280
0.0185	9.51	7500	0.4069	0.9059	0.9278
0.0135	9.63	7600	0.4067	0.9049	0.9275
0.0118	9.76	7700	0.4076	0.9039	0.9267
0.0163	9.89	7800	0.4081	0.9050	0.9275

Framework versions

Transformers 4.35.2
Pytorch 2.1.0+cu121
Datasets 2.15.0
Tokenizers 0.15.0