BERT_model_new

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

Loss: 0.1206
F1: 0.8301

Model description

train_df = pd.read_csv('/content/drive/My Drive/DATASETS/wiki_toxic/train.csv')
validation_df = pd.read_csv('/content/drive/My Drive/DATASETS/wiki_toxic/validation.csv')
#test_df = pd.read_csv('/content/drive/My Drive/wiki_toxic/test.csv')
frac = 0.9
#TRAIN
print(train_df.shape[0]) # get the number of rows in the dataframe
rows_to_delete = train_df.sample(frac=frac, random_state=1)
train_df = train_df.drop(rows_to_delete.index)
print(train_df.shape[0])\

#VALIDATION
print(validation_df.shape[0]) # get the number of rows in the dataframe
rows_to_delete = validation_df.sample(frac=frac, random_state=1)
validation_df = validation_df.drop(rows_to_delete.index)
print(validation_df.shape[0])\

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: 2

Training results

Training Loss	Epoch	Step	Validation Loss	F1
No log	1.0	399	0.0940	0.8273
0.1262	2.0	798	0.1206	0.8301

Framework versions

Transformers 4.28.1
Pytorch 2.0.0+cu118
Datasets 2.11.0
Tokenizers 0.13.3