---
license: apache-2.0
datasets:
- Kundyzka/informatics_kaz
language:
- kk
metrics:
  - name: F1 (Before Training)
    type: F1 Score
    value: 17.797
  - name: Exact Match (Before Training)
    type: Exact Match
    value: 7.662
  - name: F1 (After Training)
    type: F1 Score
    value: 67.788
  - name: Exact Match (After Training)
    type: Exact Match
    value: 51.428
base_model:
- kz-transformers/kaz-roberta-conversational
new_version: Kundyzka/kaz-roberta-conversational-informatics
pipeline_tag: question-answering
library_name: adapter-transformers
tags:
- computerscience
---

# Description

This model was developed by **Kundyz Maksutova**, PhD Candidate, as part of research on improving question-answering systems in the Kazakh language. It is a fine-tuned version of `kz-transformers/kaz-roberta-conversational` using the `Kundyzka/informatics_kaz` dataset. The model is optimized for answering questions in Kazakh, with a primary focus on computer science and related fields.

### Key Features:
- **Developer**: Kundyz Maksutova, PhD Candidate
- **Base Model**: `kz-transformers/kaz-roberta-conversational`
- **Dataset**: `Kundyzka/informatics_kaz`
- **Language**: Kazakh (`kk`)
- **Task**: Question Answering (`pipeline_tag: question-answering`)
- **Library**: `adapter-transformers`

### Performance:
The model achieves the following performance metrics, highlighting its improvement after fine-tuning:
- **Before Training**:
  - F1 Score: 17.797
  - Exact Match (EM): 7.662
- **After Training**:
  - F1 Score: 67.788
  - Exact Match (EM): 51.428

These metrics were evaluated on the `Kundyzka/informatics_kaz` dataset, demonstrating a significant improvement in performance and reliability for domain-specific questions.

### Intended Use:
This model is designed to handle natural language questions in the Kazakh language. It is particularly well-suited for:
- **Educational Platforms**: Assisting students with questions in computer science.
- **Research Projects**: Facilitating studies and experiments in Kazakh natural language processing.
- **AI Applications**: Powering chatbots and intelligent systems requiring accurate and domain-specific answers.

### Limitations:
- **Domain Dependency**: The model is fine-tuned for computer science topics, and performance may degrade on unrelated queries.
- **Bias**: The training dataset may introduce biases that could affect the model’s responses.
- **Language**: The model supports only the Kazakh language and is not designed for multilingual use.

### Tags:
- `computerscience`
- `question-answering`
- `Kazakh`
- `adapter-transformers`

This model contributes to advancing natural language processing for low-resource languages like Kazakh, with a focus on computer science applications. For further details, fine-tuning guidelines, or customization, refer to the model repository.