README.md

metadata

license: mit
tags:
  - generated_from_trainer
datasets:
  - crows_pairs
metrics:
  - accuracy
model-index:
  - name: xlnet-base-cased_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: 0.5397350993377483

xlnet-base-cased_crows_pairs_classifieronly

This model is a fine-tuned version of xlnet-base-cased on the crows_pairs dataset. It achieves the following results on the evaluation set:

• Loss: 0.6905
• Accuracy: 0.5397
• Tp: 0.2550
• Tn: 0.2848
• Fp: 0.2417
• Fn: 0.2185

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
0.723	1.05	20	0.7244	0.4735	0.4338	0.0397	0.4868	0.0397
0.7134	2.11	40	0.7069	0.4702	0.3113	0.1589	0.3675	0.1623
0.717	3.16	60	0.7003	0.5099	0.2152	0.2947	0.2318	0.2583
0.7032	4.21	80	0.7117	0.4868	0.4272	0.0596	0.4669	0.0464
0.705	5.26	100	0.6979	0.4834	0.1987	0.2848	0.2417	0.2748
0.7005	6.32	120	0.7112	0.4735	0.4073	0.0662	0.4603	0.0662
0.7163	7.37	140	0.6978	0.5430	0.0695	0.4735	0.0530	0.4040
0.7038	8.42	160	0.6960	0.5099	0.1722	0.3377	0.1887	0.3013
0.6948	9.47	180	0.6993	0.4834	0.2947	0.1887	0.3377	0.1788
0.6947	10.53	200	0.6959	0.5397	0.2219	0.3179	0.2086	0.2517
0.6991	11.58	220	0.6938	0.4967	0.1291	0.3675	0.1589	0.3444
0.7027	12.63	240	0.6959	0.5199	0.0662	0.4536	0.0728	0.4073
0.6945	13.68	260	0.6963	0.5166	0.3278	0.1887	0.3377	0.1457
0.7047	14.74	280	0.6902	0.5199	0.1424	0.3775	0.1490	0.3311
0.6971	15.79	300	0.6929	0.5596	0.2682	0.2914	0.2351	0.2053
0.6979	16.84	320	0.6919	0.5364	0.2119	0.3245	0.2020	0.2616
0.6941	17.89	340	0.6915	0.5232	0.2020	0.3212	0.2053	0.2715
0.693	18.95	360	0.6906	0.5397	0.1987	0.3411	0.1854	0.2748
0.6916	20.0	380	0.6912	0.5497	0.1954	0.3543	0.1722	0.2781
0.7005	21.05	400	0.6903	0.5397	0.2152	0.3245	0.2020	0.2583
0.6933	22.11	420	0.6904	0.5298	0.2219	0.3079	0.2185	0.2517
0.6968	23.16	440	0.6893	0.5464	0.1821	0.3642	0.1623	0.2914
0.686	24.21	460	0.6941	0.5199	0.3377	0.1821	0.3444	0.1358
0.6905	25.26	480	0.6918	0.5497	0.2848	0.2649	0.2616	0.1887
0.6954	26.32	500	0.6964	0.5199	0.3642	0.1556	0.3709	0.1093
0.6939	27.37	520	0.6897	0.5464	0.2583	0.2881	0.2384	0.2152
0.6885	28.42	540	0.6890	0.5430	0.1656	0.3775	0.1490	0.3079
0.6849	29.47	560	0.6922	0.5662	0.2914	0.2748	0.2517	0.1821
0.6869	30.53	580	0.6954	0.5331	0.3212	0.2119	0.3146	0.1523
0.6855	31.58	600	0.6910	0.5563	0.2185	0.3377	0.1887	0.2550
0.6876	32.63	620	0.6906	0.5861	0.2616	0.3245	0.2020	0.2119
0.6908	33.68	640	0.6954	0.5298	0.3444	0.1854	0.3411	0.1291
0.6757	34.74	660	0.6906	0.5662	0.2483	0.3179	0.2086	0.2252
0.6756	35.79	680	0.6905	0.5695	0.2550	0.3146	0.2119	0.2185
0.7021	36.84	700	0.6948	0.5298	0.3245	0.2053	0.3212	0.1490
0.6926	37.89	720	0.6909	0.5563	0.2682	0.2881	0.2384	0.2053
0.6913	38.95	740	0.6901	0.5563	0.2483	0.3079	0.2185	0.2252
0.6963	40.0	760	0.6921	0.5265	0.2848	0.2417	0.2848	0.1887
0.6922	41.05	780	0.6917	0.5331	0.2815	0.2517	0.2748	0.1921
0.6916	42.11	800	0.6912	0.5298	0.2616	0.2682	0.2583	0.2119
0.685	43.16	820	0.6900	0.5497	0.2318	0.3179	0.2086	0.2417
0.6839	44.21	840	0.6907	0.5364	0.2616	0.2748	0.2517	0.2119
0.6887	45.26	860	0.6913	0.5199	0.2682	0.2517	0.2748	0.2053
0.6845	46.32	880	0.6907	0.5331	0.2550	0.2781	0.2483	0.2185
0.684	47.37	900	0.6901	0.5464	0.2384	0.3079	0.2185	0.2351
0.6727	48.42	920	0.6903	0.5464	0.2517	0.2947	0.2318	0.2219
0.6801	49.47	940	0.6905	0.5397	0.2550	0.2848	0.2417	0.2185

Framework versions

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