README.md

metadata

license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - crows_pairs
metrics:
  - accuracy
model-index:
  - name: multiberts-seed_0_crows_pairs_classifieronly
    results:
      - task:
          name: Text Classification
          type: text-classification
        dataset:
          name: crows_pairs
          type: crows_pairs
          config: crows_pairs
          split: test
          args: crows_pairs
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.5132450331125827

multiberts-seed_0_crows_pairs_classifieronly

This model is a fine-tuned version of google/multiberts-seed_0 on the crows_pairs dataset. It achieves the following results on the evaluation set:

• Loss: 0.6940
• Accuracy: 0.5132
• Tp: 0.0695
• Tn: 0.4437
• Fp: 0.0199
• Fn: 0.4669

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Tp	Tn	Fp	Fn
0.7145	1.05	20	0.6927	0.5166	0.4172	0.0993	0.3642	0.1192
0.7091	2.11	40	0.6916	0.5166	0.4967	0.0199	0.4437	0.0397
0.7005	3.16	60	0.6943	0.4702	0.1854	0.2848	0.1788	0.3510
0.7106	4.21	80	0.6976	0.4669	0.0033	0.4636	0.0	0.5331
0.7038	5.26	100	0.6933	0.5132	0.3411	0.1722	0.2914	0.1954
0.7012	6.32	120	0.6974	0.4669	0.0033	0.4636	0.0	0.5331
0.7	7.37	140	0.6918	0.5199	0.4768	0.0430	0.4205	0.0596
0.7082	8.42	160	0.6960	0.4702	0.0066	0.4636	0.0	0.5298
0.699	9.47	180	0.6936	0.5199	0.2781	0.2417	0.2219	0.2583
0.7004	10.53	200	0.7043	0.4636	0.0	0.4636	0.0	0.5364
0.6948	11.58	220	0.6907	0.5232	0.5199	0.0033	0.4603	0.0166
0.7132	12.63	240	0.6979	0.4669	0.0033	0.4636	0.0	0.5331
0.7062	13.68	260	0.6930	0.5232	0.3444	0.1788	0.2848	0.1921
0.6983	14.74	280	0.6966	0.4669	0.0033	0.4636	0.0	0.5331
0.6996	15.79	300	0.6927	0.5265	0.3742	0.1523	0.3113	0.1623
0.7039	16.84	320	0.6972	0.4669	0.0033	0.4636	0.0	0.5331
0.6862	17.89	340	0.6914	0.5232	0.4967	0.0265	0.4371	0.0397
0.6943	18.95	360	0.6947	0.4934	0.0430	0.4503	0.0132	0.4934
0.7063	20.0	380	0.6907	0.5232	0.5199	0.0033	0.4603	0.0166
0.7087	21.05	400	0.6947	0.4834	0.0331	0.4503	0.0132	0.5033
0.7033	22.11	420	0.6945	0.4934	0.0464	0.4470	0.0166	0.4901
0.7025	23.16	440	0.6934	0.4967	0.2185	0.2781	0.1854	0.3179
0.7035	24.21	460	0.6944	0.5033	0.0563	0.4470	0.0166	0.4801
0.6958	25.26	480	0.6941	0.5199	0.0795	0.4404	0.0232	0.4570
0.6955	26.32	500	0.6946	0.4868	0.0364	0.4503	0.0132	0.5
0.7046	27.37	520	0.6932	0.5199	0.2715	0.2483	0.2152	0.2649
0.6955	28.42	540	0.6954	0.4702	0.0066	0.4636	0.0	0.5298
0.7095	29.47	560	0.6961	0.4702	0.0066	0.4636	0.0	0.5298
0.6953	30.53	580	0.6919	0.5132	0.4371	0.0762	0.3874	0.0993
0.7124	31.58	600	0.6946	0.4834	0.0331	0.4503	0.0132	0.5033
0.6929	32.63	620	0.6936	0.5265	0.1457	0.3808	0.0828	0.3907
0.7103	33.68	640	0.6926	0.5265	0.3477	0.1788	0.2848	0.1887
0.6993	34.74	660	0.6937	0.5166	0.1192	0.3974	0.0662	0.4172
0.6975	35.79	680	0.6936	0.5199	0.1291	0.3907	0.0728	0.4073
0.6935	36.84	700	0.6937	0.5199	0.1026	0.4172	0.0464	0.4338
0.7039	37.89	720	0.6925	0.5232	0.3642	0.1589	0.3046	0.1722
0.6999	38.95	740	0.6941	0.5099	0.0662	0.4437	0.0199	0.4702
0.6965	40.0	760	0.6948	0.4735	0.0166	0.4570	0.0066	0.5199
0.7039	41.05	780	0.6944	0.4934	0.0430	0.4503	0.0132	0.4934
0.7026	42.11	800	0.6939	0.5199	0.0795	0.4404	0.0232	0.4570
0.7072	43.16	820	0.6930	0.5199	0.2781	0.2417	0.2219	0.2583
0.7064	44.21	840	0.6930	0.5199	0.2781	0.2417	0.2219	0.2583
0.6982	45.26	860	0.6933	0.5199	0.2119	0.3079	0.1556	0.3245
0.6921	46.32	880	0.6935	0.5232	0.1490	0.3742	0.0894	0.3874
0.6983	47.37	900	0.6939	0.5199	0.0762	0.4437	0.0199	0.4603
0.6938	48.42	920	0.6940	0.5132	0.0695	0.4437	0.0199	0.4669
0.6984	49.47	940	0.6940	0.5132	0.0695	0.4437	0.0199	0.4669

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1
• Datasets 2.10.1
• Tokenizers 0.13.2