Text Classification
Transformers
PyTorch
Arabic
bert
hate-speech
gender-based-violence
arabic
binary-classification
pilot
Eval Results (legacy)
text-embeddings-inference
Instructions to use thejosango/nuha-binary with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use thejosango/nuha-binary with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="thejosango/nuha-binary")# Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("thejosango/nuha-binary") model = AutoModelForSequenceClassification.from_pretrained("thejosango/nuha-binary") - Notebooks
- Google Colab
- Kaggle
| [experiment] | |
| name = "binary-56" | |
| type = "binary" | |
| [dataset] | |
| path = "thejosango/nuha-dataset" | |
| dataset_revision = "main" | |
| augment_ratio = 0.75 | |
| undersampling_strategy = false | |
| [model] | |
| pretrained_model_name_or_path = "thejosango/nuha-mlm" | |
| revision = "ce20f497544665775129f9ff5b3cd2a3e350dce8" | |
| num_hidden_layers = 4 | |
| classifier_dropout = 0.50 | |
| [training] | |
| num_train_epochs = 5 | |
| warmup_steps = 0 | |
| lr_scheduler_type = "linear" | |
| learning_rate = 5e-5 | |
| per_device_train_batch_size = 64 | |
| per_device_eval_batch_size = 64 | |
| gradient_accumulation_steps = 1 | |
| weight_decay = 1e-3 | |
| label_smoothing_factor = 0.1 | |
| weighted_loss = true | |
| early_stopping_patience = 5 | |
| early_stopping_threshold = 0.005 | |