# Fabelous-Albert-Uncased

**Fabelous-Albert-Uncased** is a bilingual ALBERT model pretrained on German, English, and code. This uncased model is a Masked Language Model (MLM) and can be fine-tuned for a variety of tasks, including but not limited to:

- Named Entity Recognition (NER)
- Binary Classification
- Text Completion

The model has been designed for efficiency and compatibility, requiring the use of the `FastTokenizer` for optimal performance.


## Features

### 1. **Bilingual Support**
   - Trained on English and German text, enabling seamless bilingual tasks.

### 2. **Code Understanding**
   - Incorporates code in its training data, making it suitable for programming-related NLP tasks.

### 3. **Uncased**
   - Treats words as case-insensitive, which simplifies preprocessing steps and generalizes better for certain tasks.

### 4. **Fine-Tuning Ready**
   - Easily fine-tune for tasks such as text classification, named entity recognition, and more.


## Downloading the Model

You can download the `fabelous-albert-uncased` model using the following link:

[Download Fabelous-Albert-Uncased Model](https://gitea.fabelous.app/Fabel/Fabelous-albert-uncased/releases/download/latest/fabelous-albert-uncased.zip)

### Installation Instructions

1. Click the link above to download a ZIP file containing the model files.
2. Extract the ZIP file into your desired directory.
3. Load the model in your Python project using the `transformers` library.

## Usage Example

Below is a sample code snippet to demonstrate how to use the `fabelous-albert-uncased` model for a masked language modeling task:

```python
from transformers import pipeline

# Load the pipeline with the Fabelous-Albert-Uncased model
unmasker = pipeline('fill-mask', model='fabelous-albert-uncased')

# Perform masked language modeling
output = unmasker("Hello I'm a [MASK] model.")
print(output)
```



## Future Enhancements: New Model Announcement

We are thrilled to announce that a new version of the model is currently under development! The upcoming model will:

- **Quadruple the Training Size**: With four times more data, expect significantly improved performance across diverse tasks.
- **Cased Version**: In addition to the uncased version, a cased model will be introduced, preserving capitalization for more nuanced language understanding.
- **Extended Language Support**: Support for multiple additional languages beyond English and German.
- **Slow- and FastTokenizer Support**: Support for both tokenizer Versions. 

Stay tuned for updates as we prepare to release this enhanced model in the near future.


## License

This model is released under the [Creative Commons Attribution 4.0 International Licence](https://creativecommons.org/licenses/by/4.0/).

## Feedback and Support

If you encounter any issues or have questions, feel free to reach out through the project's [Gitea Issues page](https://gitea.fabelous.app/Fabel/Fabelous-albert-uncased/issues) or contact our support team at support@fabelous.app.