Instructions to use mlabonne/phixtral-4x2_8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mlabonne/phixtral-4x2_8 with Transformers:

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

pipe = pipeline("text-generation", model="mlabonne/phixtral-4x2_8", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("mlabonne/phixtral-4x2_8", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use mlabonne/phixtral-4x2_8 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mlabonne/phixtral-4x2_8"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mlabonne/phixtral-4x2_8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/mlabonne/phixtral-4x2_8

SGLang

How to use mlabonne/phixtral-4x2_8 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 "mlabonne/phixtral-4x2_8" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mlabonne/phixtral-4x2_8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "mlabonne/phixtral-4x2_8" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mlabonne/phixtral-4x2_8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use mlabonne/phixtral-4x2_8 with Docker Model Runner:
```
docker model run hf.co/mlabonne/phixtral-4x2_8
```

hello mlabonne, can you share the mergekit's merge.yaml of phixtral-4x2_8

#15

by TomPei - opened Jan 22, 2024

Discussion

TomPei

Jan 22, 2024

hello mlabonne, can you share the mergekit's merge.yaml of phixtral-4x2_8.
I want to merge some other models for tests.
Thank you very much

mlabonne

Owner Jan 22, 2024

It's in the model card

HanNayeoniee

Jan 25, 2024

It's in the model card

I tried to merge model using the same yaml file in the model card, but I got this error.
It seems like it's giving me errors since there's nothing in positive_prompts.

root@ubuntu:/workspace# mergekit-moe ./my_setting/phixtral-4x2_8.yml ./models/moe/phixtral-4x2_8_test
ERROR:root:Your positive and negative prompts are identical for all experts. This will not produce a functioning MoE.
ERROR:root:For each expert, `positive_prompts` must contain one or more example prompt reflecting what should be routed to that expert.

In the yaml file of another model, Undi95/Mixtral-8x7B-MoE-RP-Story for example, keywords are listed in positive_prompts and negative_prompts for each base model.

Could you please share details of merge.yaml file?
Or did I do something wrong?

Thanks in advance!

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment