elifcen/bert-pooling-based

1---2license: apache-2.03tags:4- generated_from_trainer5datasets:6- glue7metrics:8- matthews_correlation9model-index:10- name: bert-pooling-based11  results:12  - task:13      name: Text Classification14      type: text-classification15    dataset:16      name: glue17      type: glue18      config: cola19      split: validation20      args: cola21    metrics:22    - name: Matthews Correlation23      type: matthews_correlation24      value: 0.4085856417909235525---26 27<!-- This model card has been generated automatically according to the information the Trainer had access to. You28should probably proofread and complete it, then remove this comment. -->29 30# bert-pooling-based31 32This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset.33It achieves the following results on the evaluation set:34- Loss: 0.511535- Matthews Correlation: 0.408636 37## Model description38 39More information needed40 41## Intended uses & limitations42 43More information needed44 45## Training and evaluation data46 47More information needed48 49## Training procedure50 51### Training hyperparameters52 53The following hyperparameters were used during training:54- learning_rate: 1.7718352056354854e-0655- train_batch_size: 856- eval_batch_size: 1657- seed: 4258- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-0859- lr_scheduler_type: linear60- num_epochs: 261 62### Training results63 64| Training Loss | Epoch | Step | Validation Loss | Matthews Correlation |65|:-------------:|:-----:|:----:|:---------------:|:--------------------:|66| 0.5491        | 1.0   | 1069 | 0.5340          | 0.2513               |67| 0.4726        | 2.0   | 2138 | 0.5115          | 0.4086               |68 69 70### Framework versions71 72- Transformers 4.28.173- Pytorch 2.0.0+cu11874- Datasets 2.12.075- Tokenizers 0.13.376