---
language: en
license: apache-2.0
tags:
- text-generation
- mistralai
- moe
- question-answering
- natural-questions
base_model:
- mistralai/Mixtral-8x7B-v0.1
---

# Mixtral-8x7B MoE Fine-tuned on Natural Questions

This model is a fine-tuned version of [MistralAI/Mixtral-8x7B-0.1](https://huggingface.co/mistralai/Mixtral-8x7B-v0.1) on the Natural Questions dataset.

## Model Description

- **Base Model**: mistralai/Mixtral-8x7B-0.1
- **Architecture**: Mixture-of-Experts (MoE) variant
- **Training Dataset**: Natural Questions (NQ) - annotated for expert routing
- **Task**: Question Answering / Text Generation

## Training Details

The model was fine-tuned using:
- DeepSpeed with ZeRO Stage 2 optimization
- HuggingFace Accelerate
- Custom MoE expert annotations

## Usage

```python
from transformers import AutoTokenizer, AutoModelForCausalLM

model_name = "vvijayk/mixtral-8x7b-moe-nq-finetuned"
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    trust_remote_code=True,
    torch_dtype="auto",
    device_map="auto"
)

# Example inference
question = "What is the capital of France?"
prompt = f"Question: {question}\nAnswer:"

inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(
    **inputs,
    max_new_tokens=128,
    temperature=0.7,
    do_sample=True,
    top_p=0.9,
)

answer = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(answer)
```

## Training Configuration

Training was performed using the configuration in the repository. See `train_mixtral_8x7b_moe_accelerate.py` for details.

## Limitations and Biases

This model inherits limitations and biases from the base Mixtral-8x7B model and the Natural Questions dataset.
Users should be aware of potential biases in question-answering outputs.

## Citation

If you use this model, please cite:

```bibtex
@misc{mixtral-8x7b-moe-nq-finetuned,
  author = {Vijay Venkatraman},
  title = {Mixtral-8x7B-0.1 Fine-tuned on Natural Questions},
  year = {2025},
  publisher = {HuggingFace},
  howpublished = {\url{https://huggingface.co/vvijayk/mixtral-8x7b-moe-nq-finetuned}}
}
```

## Acknowledgements

- Base model by Mistral AI
- Natural Questions dataset by Google Research
- Training infrastructure using DeepSpeed and HuggingFace Accelerate