Edit model card
YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/model-cards#model-card-metadata)

vinallama-legal-chat

Model Card

Description

vinallama-legal-chat is a fine-tuned version of vinallama-2-7b, specifically trained for Vietnamese legal conversations. This model is designed to assist in providing accurate legal advice and information in Vietnamese, making it a valuable tool for legal professionals and individuals seeking legal guidance.

Installation

To use this model, you will need to install the following dependencies:

pip install transformers
pip install torch  # or tensorflow depending on your preference

Usage

Here is how you can load and use the model in your code:

from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("username/vinallama-legal-chat")
model = AutoModelForCausalLM.from_pretrained("username/vinallama-legal-chat")

# Example usage
chat_template = """
<<SYS>>
Bạn là một chuyên viên tư vấn pháp luật Việt Nam. Bạn có nhiều năm kinh nghiệm và kiến thức chuyên sâu. Bạn sẽ cung cấp câu trả lời về pháp luật, tư vấn luật pháp cho các câu hỏi của User.
<</SYS>>
## user:
Tạm trú là gì?

## assistant:
"""

inputs = tokenizer(chat_template, return_tensors="pt")
outputs = model.generate(**inputs)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)

print(response)

Inference

Provide example code for performing inference with your model:

# Example inference
user_input = "Tạm trú là gì?"
chat_template = f"""
<<SYS>>
Bạn là một chuyên viên tư vấn pháp luật Việt Nam. Bạn có nhiều năm kinh nghiệm và kiến thức chuyên sâu. Bạn sẽ cung cấp câu trả lời về pháp luật, tư vấn luật pháp cho các câu hỏi của User.
<</SYS>>
## user:
{user_input}

## assistant:
"""

inputs = tokenizer(chat_template, return_tensors="pt")
outputs = model.generate(**inputs)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)

print(response)

Training

If your model can be trained further, provide instructions for training:

# Example training code
from transformers import Trainer, TrainingArguments

training_args = TrainingArguments(
    output_dir="./results",
    evaluation_strategy="epoch",
    per_device_train_batch_size=8,
    per_device_eval_batch_size=8,
    num_train_epochs=3,
    weight_decay=0.01,
)

trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=train_dataset,
    eval_dataset=eval_dataset,
)

trainer.train()

Training Details

Training Data

The model was fine-tuned on a dataset of Vietnamese legal conversations. This dataset includes a variety of legal questions and answers, covering a wide range of legal topics to ensure comprehensive legal advice.

Training Procedure

The model was fine-tuned using a standard training approach, optimizing for accuracy and relevance in legal responses. Training was conducted on [describe hardware, e.g., GPUs, TPUs] over [number of epochs] epochs with [any relevant hyperparameters].

Evaluation

Metrics

The model was evaluated using the following metrics:

  • Accuracy: X%
  • Relevance: Y%
  • Comprehensiveness: Z%

Comparison

The performance of vinallama-legal-chat was benchmarked against other legal advice models, demonstrating superior accuracy and relevance in the Vietnamese legal domain.

Limitations and Biases

While vinallama-legal-chat is highly effective, it may have limitations in the following areas:

  • It may not be up-to-date with the latest legal changes.
  • There may be biases present in the training data that could affect responses.

How to Contribute

We welcome contributions! Please see our contributing guidelines for more information on how to contribute to this project.

License

This model is licensed under the MIT License.

Acknowledgements

We would like to thank the contributors and the creators of the datasets used for training this model.


### Tips for Completing the Template

1. **Replace placeholders** (like `username`, `training data`, `evaluation metrics`) with your actual data.
2. **Include any additional information** specific to your model or training process.
3. **Keep the document updated** as the model evolves or more information becomes available.
Downloads last month
23
GGUF
Model size
2.77B params
Architecture
llama

4-bit

Unable to determine this model’s pipeline type. Check the docs .