Not-For-All-Audiences

nsfw

Instructions to use Undi95/Mixtral-8x7B-MoE-RP-Story with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Undi95/Mixtral-8x7B-MoE-RP-Story with Transformers:

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

pipe = pipeline("text-generation", model="Undi95/Mixtral-8x7B-MoE-RP-Story")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Undi95/Mixtral-8x7B-MoE-RP-Story")
model = AutoModelForCausalLM.from_pretrained("Undi95/Mixtral-8x7B-MoE-RP-Story", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Undi95/Mixtral-8x7B-MoE-RP-Story with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Undi95/Mixtral-8x7B-MoE-RP-Story"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Undi95/Mixtral-8x7B-MoE-RP-Story",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Undi95/Mixtral-8x7B-MoE-RP-Story

SGLang

How to use Undi95/Mixtral-8x7B-MoE-RP-Story 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 "Undi95/Mixtral-8x7B-MoE-RP-Story" \
    --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": "Undi95/Mixtral-8x7B-MoE-RP-Story",
		"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 "Undi95/Mixtral-8x7B-MoE-RP-Story" \
        --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": "Undi95/Mixtral-8x7B-MoE-RP-Story",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Undi95/Mixtral-8x7B-MoE-RP-Story with Docker Model Runner:
```
docker model run hf.co/Undi95/Mixtral-8x7B-MoE-RP-Story
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Mixtral-8x7B-MoE-RP-Story is a model made primarely for chatting, RP (Roleplay) and storywriting. 2 RP model, 2 chat model, 1 occult model, 1 storywritting model, 1 mathematic model and 1 DPO model was used for a MoE. Bagel was the base.

The DPO chat model is here to help get more human reply.

This is my first try at doing this, so don't hesitate to give feedback!

WARNING: ALL THE "K" GGUF QUANT OF MIXTRAL MODELS SEEMS TO BE BROKEN, PREFER Q4_0, Q5_0 or Q8_0!

Description

This repo contains fp16 files of Mixtral-8x7B-MoE-RP-Story.

Models used

The list of model used and their activator/theme can be found here

Prompt template: Custom

Using Bagel as a base let us a lot of different prompting system theorically, you can see all the prompting available here.

If you want to support me, you can here.

Downloads last month: 127

Safetensors

Model size

47B params

Tensor type

BF16

F16

Model tree for Undi95/Mixtral-8x7B-MoE-RP-Story

Quantizations

5 models

Undi95
/

Mixtral-8x7B-MoE-RP-Story

Description

Models used

Prompt template: Custom

Model tree for Undi95/Mixtral-8x7B-MoE-RP-Story

Space using Undi95/Mixtral-8x7B-MoE-RP-Story 1