Llama-turkish-lawbot

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

results

This model is a fine-tuned version of meta-llama/Llama-3.2-3B-Instruct on Renicames/turkish-law-chatbot dataset, designed to generate responses based on Turkish legal questions and answers.

Model description

This model is fine-tuned to imporove its ability to generate responses for a Turkish law chatbot. meta-llama/Llama-3.2-3B-Instruct was used as the base model, which can pretty much handle various tasks. Just specialized it with the dataset of Renicames/turkish-law-chatbot.

Intended uses & limitations

Intended uses

• This model can be deployed in applications aimed at providing legal information or answering legal questions in Turkish.
• Useful for generating legal text, such as explanations of laws or providing examples of legal processes in Turkish.

Limitaions

• While the model has been fine-tuned on the Turkish legal domain, it may still lack the depth and specificity required for complex legal inquiries. It might not be suitable for professional legal advice.
• As with any AI model, it may reflect biases found in the training data, so it should be used with caution in critical applications.
• This model is focused on the Turkish language, so it may not perform well in other languages or mixed-language queries.

Training and evaluation data

Training data details:

• Source: Renicames/turkish-law-chatbot dataset
• Languages: Turkish
• Content: Legal questions and answers
• Size: The dataset consists of several thousand question-answer pairs related to Turkish law.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 8
• eval_batch_size: 16
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 3

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.3.1
• Tokenizers 0.21.0