metadata

license: apache-2.0
base_model: distilbert-base-uncased
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - recall
  - precision
  - f1
model-index:
  - name: hatespeech_distilbert
    results: []
widget:
  - text: Democrats using African-Americans again.
    example_title: Non-Hate Speech Example
  - text: Holy fuck this girl's trash, what a cunt.
    example_title: Hate Speech Example

hatespeech_distilbert

This model is a fine-tuned version of distilbert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.9977
Accuracy: 0.7737
Recall: 0.8118
Precision: 0.7526
F1: 0.7811

And the following results on the test set:

Loss: 1.0640
Accuracy: 0.7544
Recall: 0.7930
Precision: 0.7406
F1: 0.7659

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: 8e-05
train_batch_size: 32
eval_batch_size: 32
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 128
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Recall	Precision	F1
0.4863	0.9935	77	0.4678	0.7701	0.7421	0.7841	0.7625
0.3935	2.0	155	0.4595	0.7834	0.7340	0.8124	0.7712
0.2792	2.9935	232	0.5285	0.7850	0.7291	0.8188	0.7713
0.1408	4.0	310	0.7130	0.7785	0.7940	0.7684	0.7810
0.0945	4.9935	387	0.8230	0.7806	0.7551	0.7937	0.7739
0.0541	6.0	465	0.9977	0.7737	0.8118	0.7526	0.7811
0.0331	6.9935	542	1.1107	0.7753	0.7859	0.7678	0.7768
0.0151	8.0	620	1.1703	0.7789	0.7543	0.7915	0.7724
0.0106	8.9935	697	1.2741	0.7785	0.7616	0.7864	0.7738
0.0051	9.9355	770	1.2964	0.7753	0.7851	0.7683	0.7766

Framework versions

Transformers 4.40.2
Pytorch 2.2.1+cu121
Datasets 2.19.1
Tokenizers 0.19.1