README.md

metadata

license: apache-2.0
tags:
  - finetuned
  - quantized
  - 4-bit
  - gptq
  - transformers
  - pytorch
  - llama
  - text-generation
  - arxiv:2304.12244
  - arxiv:2306.08568
  - arxiv:2308.09583
  - license:llama2
  - autotrain_compatible
  - endpoints_compatible
  - has_space
  - text-generation-inference
  - region:us
model_name: WizardLM-70B-V1.0-GPTQ
base_model: meta-llama/Llama-2-70b-chat-hf
inference: false
model_creator: WizardLM
pipeline_tag: text-generation
quantized_by: MaziyarPanahi

Description

MaziyarPanahi/WizardLM-70B-V1.0-GPTQ is a quantized (GPTQ) version of WizardLM/WizardLM-70B-V1.0

How to use

Install the necessary packages

pip install --upgrade accelerate auto-gptq transformers

Example Python code

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

model_id = "MaziyarPanahi/WizardLM-70B-V1.0-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"])