GPTQ 4bit Inference

Table of contents

Install
Benchmark

Support GPTQ 4bit inference with GPTQ-for-LLaMa.

Window user: use the old-cuda branch.
Linux user: recommend the fastest-inference-4bit branch.

Install

Setup environment:

# cd /path/to/FastChat
git clone https://github.com/qwopqwop200/GPTQ-for-LLaMa.git repositories/GPTQ-for-LLaMa
cd repositories/GPTQ-for-LLaMa
# Window's user should use the `old-cuda` branch
git switch fastest-inference-4bit
# Install `quant-cuda` package in FastChat's virtualenv
python3 setup_cuda.py install
pip3 install texttable

Chat with the CLI:

python3 -m fastchat.serve.cli \
    --model-path models/vicuna-7B-1.1-GPTQ-4bit-128g \
    --gptq-wbits 4 \
    --gptq-groupsize 128

Start model worker:

# Download quantized model from huggingface
# Make sure you have git-lfs installed (https://git-lfs.com)
git lfs install
git clone https://huggingface.co/TheBloke/vicuna-7B-1.1-GPTQ-4bit-128g models/vicuna-7B-1.1-GPTQ-4bit-128g

python3 -m fastchat.serve.model_worker \
    --model-path models/vicuna-7B-1.1-GPTQ-4bit-128g \
    --gptq-wbits 4 \
    --gptq-groupsize 128

# You can specify which quantized model to use
python3 -m fastchat.serve.model_worker \
    --model-path models/vicuna-7B-1.1-GPTQ-4bit-128g \
    --gptq-ckpt models/vicuna-7B-1.1-GPTQ-4bit-128g/vicuna-7B-1.1-GPTQ-4bit-128g.safetensors \
    --gptq-wbits 4 \
    --gptq-groupsize 128 \
    --gptq-act-order

Benchmark

LLaMA-13B	branch	Bits	group-size	memory(MiB)	PPL(c4)	Median(s/token)	act-order	speed up
FP16	fastest-inference-4bit	16	-	26634	6.96	0.0383	-	1x
GPTQ	triton	4	128	8590	6.97	0.0551	-	0.69x
GPTQ	fastest-inference-4bit	4	128	8699	6.97	0.0429	true	0.89x
GPTQ	fastest-inference-4bit	4	128	8699	7.03	0.0287	false	1.33x
GPTQ	fastest-inference-4bit	4	-1	8448	7.12	0.0284	false	1.44x