Instructions to use martyn/mixtral-dare-8x7b-v0 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use martyn/mixtral-dare-8x7b-v0 with Transformers:

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

pipe = pipeline("text-generation", model="martyn/mixtral-dare-8x7b-v0")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("martyn/mixtral-dare-8x7b-v0")
model = AutoModelForCausalLM.from_pretrained("martyn/mixtral-dare-8x7b-v0", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use martyn/mixtral-dare-8x7b-v0 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "martyn/mixtral-dare-8x7b-v0"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "martyn/mixtral-dare-8x7b-v0",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/martyn/mixtral-dare-8x7b-v0

SGLang

How to use martyn/mixtral-dare-8x7b-v0 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 "martyn/mixtral-dare-8x7b-v0" \
    --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": "martyn/mixtral-dare-8x7b-v0",
		"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 "martyn/mixtral-dare-8x7b-v0" \
        --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": "martyn/mixtral-dare-8x7b-v0",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use martyn/mixtral-dare-8x7b-v0 with Docker Model Runner:
```
docker model run hf.co/martyn/mixtral-dare-8x7b-v0
```

mixtral dare test

The following were merged with DARE using https://github.com/martyn/safetensors-merge-supermario

Mergelist

mistralai/Mixtral-8x7B-Instruct-v0.1
Open-Orca/Mixtral-SlimOrca-8x7B

Merge command

python3 hf_merge.py to_merge_mixtral0.txt mixtral-0 -p 0.3 -lambda 2.1

Notes

This is primarily a test to see if merging mixtral models works.
MoE gates are not merged.

Downloads last month: 11

Safetensors

Model size

47B params

Tensor type

BF16

Collection including martyn/mixtral-dare-8x7b-v0

Dare LLM Merges

Collection

These are large language models merged through my implementation of Super Mario DARE merge. • 10 items • Updated Dec 31, 2023 • 2