pft-clf-finetuned

This model is a fine-tuned version of HooshvareLab/bert-fa-zwnj-base on an "FarsNews1398" dataset. This dataset contains a collection of news that has been gathered from the farsnews website which is a news agency in Iran. You can download the dataset from here. I used category, abstract, and paragraphs of news for doing text classification. "abstract" and "paragraphs" for each news concatenated together and "category" used as a target for classification.

The notebook used for fine-tuning can be found here. I've reported loss and Matthews correlation criteria on the validation set.

It achieves the following results on the evaluation set:

• Loss: 0.0617
• Matthews Correlation: 0.9830

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: 3e-05
• train_batch_size: 6
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Matthews Correlation
0.0634	1.0	20276	0.0617	0.9830

Framework versions

• Transformers 4.12.3
• Pytorch 1.10.0+cu111
• Datasets 1.15.1
• Tokenizers 0.10.3