Instructions to use NousResearch/Yarn-Mistral-7b-128k with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use NousResearch/Yarn-Mistral-7b-128k with Transformers:

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

pipe = pipeline("text-generation", model="NousResearch/Yarn-Mistral-7b-128k", trust_remote_code=True)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("NousResearch/Yarn-Mistral-7b-128k", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("NousResearch/Yarn-Mistral-7b-128k", trust_remote_code=True)

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use NousResearch/Yarn-Mistral-7b-128k with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "NousResearch/Yarn-Mistral-7b-128k"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NousResearch/Yarn-Mistral-7b-128k",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/NousResearch/Yarn-Mistral-7b-128k

SGLang

How to use NousResearch/Yarn-Mistral-7b-128k 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 "NousResearch/Yarn-Mistral-7b-128k" \
    --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": "NousResearch/Yarn-Mistral-7b-128k",
		"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 "NousResearch/Yarn-Mistral-7b-128k" \
        --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": "NousResearch/Yarn-Mistral-7b-128k",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use NousResearch/Yarn-Mistral-7b-128k with Docker Model Runner:
```
docker model run hf.co/NousResearch/Yarn-Mistral-7b-128k
```

VRAM usage for full 128k tokens

by Hypersniper - opened Nov 3, 2023

Discussion

Hypersniper

Nov 3, 2023

Any idea how much more vram you'll need to get the full 128k if you load the model in 4bit?

macadeliccc

Nov 3, 2023

This comment has been hidden

bloc97

NousResearch org Nov 3, 2023

Currently, it is about 4 80GB A100s, so 320GB VRAM, we are currently working on reducing this with better optimizations...

3blackbar

Nov 4, 2023

no need, ill start my delorean to travel to 2030 and get some gpus to run it locally

DemonMaike

Nov 5, 2023

Hi and thanks for your work, this is amazing.
But Could you help me with my question?
So, i small work with ai and models, and i trust look for information about multicards output for model, but i seen only info about finetune using multicard.
Maybe, you can take me links or some documentation about it.

Hypersniper

Nov 6, 2023

Currently, it is about 4 80GB A100s, so 320GB VRAM, we are currently working on reducing this with better optimizations...

Perhaps using flash attention could help

bloc97

NousResearch org Nov 6, 2023

•

edited Nov 6, 2023

Currently, it is about 4 80GB A100s, so 320GB VRAM, we are currently working on reducing this with better optimizations...

Perhaps using flash attention could help

It is already using flash attention. However if you are focused solely on inference use cases, dedicated inference kernels in libraries such as vLLM, ExLlama and llamacpp would help reduce the VRAM requirements significantly. I've heard rumours that it should be possible to run 128k context with llamacpp on a single 40GB GPU...

dragosmc

Mar 21, 2024

•

edited Mar 22, 2024

Anyone got it to work on decent machines with (almost) the full context and could share their experience? I'm trying on 2xA100 with 60k context and I get OOMs when the attention masks are calculated. The deployment uses flash attention and it's on 4 bits.

LE:

I managed to get it to work through llamacpp and 4xA10G. The stats are as follows:

60k input -> 4 minutes time for a 4k response generation
120k input -> 16 minutes time for a 4k response generation

The model was quantized to 4bits.

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