Instructions to use dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf")
model = AutoModelForCausalLM.from_pretrained("dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf

SGLang

How to use dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf with Docker Model Runner:
```
docker model run hf.co/dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf
```

dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf

This repository contains a Hugging Face export of Mixtral-8x7B-v0.1 quantized with AQLM using the 4-bit 4x8 scheme.

Base model

mistralai/Mixtral-8x7B-v0.1

Quantization

Method: AQLM
Scheme: 4x8
Effective label: 4-bit
Source checkpoint: /work/bduan1/quantized_models/Mixtral-8x7B-AQLM-4bit-4x8

Conversion

This repo was produced with convert_to_hf.py from the AQLM project, then exported with --save_safetensors and --save_tokenizer.

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype="auto",
    device_map="auto",
    trust_remote_code=True,
)

Downloads last month: 131

Safetensors

Model size

23B params

Tensor type

F16

Model tree for dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf

Base model

mistralai/Mixtral-8x7B-v0.1

Quantized

(41)

this model

Collection including dbw6/Mixtral-8x7B-AQLM-4Bit-4x8-hf

EVA

Collection

8 items • Updated Apr 5 • 2