---
license: apache-2.0
tags:
- finetuned
- quantized
- 4-bit
- gptq
- transformers
- safetensors
- mistral
- text-generation
- Safetensors
- text-generation-inference
- merge
- 7b
- mistralai/Mistral-7B-Instruct-v0.1
- HuggingFaceH4/zephyr-7b-beta
- pytorch
- generated_from_trainer
- en
- dataset:HuggingFaceH4/ultrachat_200k
- dataset:HuggingFaceH4/ultrafeedback_binarized
- arxiv:2305.18290
- arxiv:2310.16944
- base_model:mistralai/Mistral-7B-v0.1
- license:mit
- model-index
- autotrain_compatible
- endpoints_compatible
- has_space
- region:us
- license:apache-2.0
model_name: zephyr-7b-beta-Mistral-7B-Instruct-v0.1-GPTQ
base_model: MaziyarPanahi/zephyr-7b-beta-Mistral-7B-Instruct-v0.1
inference: false
model_creator: MaziyarPanahi
pipeline_tag: text-generation
quantized_by: MaziyarPanahi
---
# Description
[MaziyarPanahi/zephyr-7b-beta-Mistral-7B-Instruct-v0.1-GPTQ](https://huggingface.co/MaziyarPanahi/zephyr-7b-beta-Mistral-7B-Instruct-v0.1-GPTQ) is a quantized (GPTQ) version of [MaziyarPanahi/zephyr-7b-beta-Mistral-7B-Instruct-v0.1](https://huggingface.co/MaziyarPanahi/zephyr-7b-beta-Mistral-7B-Instruct-v0.1)

## How to use
### Install the necessary packages

```
pip install --upgrade accelerate auto-gptq transformers
```

### Example Python code


```python
from transformers import AutoTokenizer, pipeline
from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
import torch

model_id = "MaziyarPanahi/zephyr-7b-beta-Mistral-7B-Instruct-v0.1-GPTQ"

quantize_config = BaseQuantizeConfig(
        bits=4,
        group_size=128,
        desc_act=False
    )

model = AutoGPTQForCausalLM.from_quantized(
        model_id,
        use_safetensors=True,
        device="cuda:0",
        quantize_config=quantize_config)

tokenizer = AutoTokenizer.from_pretrained(model_id)

pipe = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    max_new_tokens=512,
    temperature=0.7,
    top_p=0.95,
    repetition_penalty=1.1
)

outputs = pipe("What is a large language model?")
print(outputs[0]["generated_text"])
```