metadata

language:
  - tr
license: mit
tags:
  - zero-shot-classification
  - nli
  - pytorch
datasets:
  - nli_tr
metrics:
  - accuracy
pipeline_tag: zero-shot-classification
widget:
  - text: Dolar yükselmeye devam ediyor.
    candidate_labels: ekonomi, siyaset, spor
  - text: Senaryo çok saçmaydı, beğendim diyemem.
    candidate_labels: olumlu, olumsuz
base_model: bert-base-multilingual-cased

bert-base-multilingual-cased_allnli_tr

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

Loss: 0.6144
Accuracy: 0.7662

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: 32
eval_batch_size: 32
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.8623	0.03	1000	0.9076	0.5917
0.7528	0.07	2000	0.8587	0.6119
0.7074	0.1	3000	0.7867	0.6647
0.6949	0.14	4000	0.7474	0.6772
0.6681	0.17	5000	0.7661	0.6814
0.6597	0.2	6000	0.7264	0.6943
0.6495	0.24	7000	0.7841	0.6781
0.6323	0.27	8000	0.7256	0.6952
0.6308	0.31	9000	0.7319	0.6958
0.6254	0.34	10000	0.7054	0.7004
0.6233	0.37	11000	0.7069	0.7085
0.6165	0.41	12000	0.6880	0.7181
0.6033	0.44	13000	0.6844	0.7197
0.6014	0.48	14000	0.6753	0.7129
0.5947	0.51	15000	0.7000	0.7039
0.5965	0.54	16000	0.6708	0.7263
0.5979	0.58	17000	0.6562	0.7285
0.5787	0.61	18000	0.6554	0.7297
0.58	0.65	19000	0.6544	0.7315
0.574	0.68	20000	0.6549	0.7339
0.5751	0.71	21000	0.6545	0.7289
0.5659	0.75	22000	0.6467	0.7371
0.5732	0.78	23000	0.6448	0.7362
0.5637	0.82	24000	0.6520	0.7355
0.5648	0.85	25000	0.6412	0.7345
0.5622	0.88	26000	0.6350	0.7358
0.5579	0.92	27000	0.6347	0.7393
0.5518	0.95	28000	0.6417	0.7392
0.5547	0.99	29000	0.6321	0.7437
0.524	1.02	30000	0.6430	0.7412
0.4982	1.05	31000	0.6253	0.7458
0.5002	1.09	32000	0.6316	0.7418
0.4993	1.12	33000	0.6197	0.7487
0.4963	1.15	34000	0.6307	0.7462
0.504	1.19	35000	0.6272	0.7480
0.4922	1.22	36000	0.6410	0.7433
0.5016	1.26	37000	0.6295	0.7461
0.4957	1.29	38000	0.6183	0.7506
0.4883	1.32	39000	0.6261	0.7502
0.4985	1.36	40000	0.6315	0.7496
0.4885	1.39	41000	0.6189	0.7529
0.4909	1.43	42000	0.6189	0.7473
0.4894	1.46	43000	0.6314	0.7433
0.4912	1.49	44000	0.6184	0.7446
0.4851	1.53	45000	0.6258	0.7461
0.4879	1.56	46000	0.6286	0.7480
0.4907	1.6	47000	0.6196	0.7512
0.4884	1.63	48000	0.6157	0.7526
0.4755	1.66	49000	0.6056	0.7591
0.4811	1.7	50000	0.5977	0.7582
0.4787	1.73	51000	0.5915	0.7621
0.4779	1.77	52000	0.6014	0.7583
0.4767	1.8	53000	0.6041	0.7623
0.4737	1.83	54000	0.6093	0.7563
0.4836	1.87	55000	0.6001	0.7568
0.4765	1.9	56000	0.6109	0.7601
0.4776	1.94	57000	0.6046	0.7599
0.4769	1.97	58000	0.5970	0.7568
0.4654	2.0	59000	0.6147	0.7614
0.4144	2.04	60000	0.6439	0.7566
0.4101	2.07	61000	0.6373	0.7527
0.4192	2.11	62000	0.6136	0.7575
0.4128	2.14	63000	0.6283	0.7560
0.4204	2.17	64000	0.6187	0.7625
0.4114	2.21	65000	0.6127	0.7621
0.4097	2.24	66000	0.6188	0.7626
0.4129	2.28	67000	0.6156	0.7639
0.4085	2.31	68000	0.6232	0.7616
0.4074	2.34	69000	0.6240	0.7605
0.409	2.38	70000	0.6153	0.7591
0.4046	2.41	71000	0.6375	0.7587
0.4117	2.45	72000	0.6145	0.7629
0.4002	2.48	73000	0.6279	0.7610
0.4042	2.51	74000	0.6176	0.7646
0.4055	2.55	75000	0.6277	0.7643
0.4021	2.58	76000	0.6196	0.7642
0.4081	2.62	77000	0.6127	0.7659
0.408	2.65	78000	0.6237	0.7638
0.3997	2.68	79000	0.6190	0.7636
0.4093	2.72	80000	0.6152	0.7648
0.4095	2.75	81000	0.6155	0.7627
0.4088	2.79	82000	0.6130	0.7641
0.4063	2.82	83000	0.6072	0.7646
0.3978	2.85	84000	0.6128	0.7662
0.4034	2.89	85000	0.6157	0.7627
0.4044	2.92	86000	0.6127	0.7661
0.403	2.96	87000	0.6126	0.7664
0.4033	2.99	88000	0.6144	0.7662

Framework versions

Transformers 4.12.3
Pytorch 1.10.0+cu102
Datasets 1.15.1
Tokenizers 0.10.3