Instructions to use EmbeddedLLM/Mistral-7B-Merge-14-v0.3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use EmbeddedLLM/Mistral-7B-Merge-14-v0.3 with Transformers:

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

pipe = pipeline("text-generation", model="EmbeddedLLM/Mistral-7B-Merge-14-v0.3")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("EmbeddedLLM/Mistral-7B-Merge-14-v0.3")
model = AutoModelForCausalLM.from_pretrained("EmbeddedLLM/Mistral-7B-Merge-14-v0.3")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use EmbeddedLLM/Mistral-7B-Merge-14-v0.3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "EmbeddedLLM/Mistral-7B-Merge-14-v0.3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "EmbeddedLLM/Mistral-7B-Merge-14-v0.3",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/EmbeddedLLM/Mistral-7B-Merge-14-v0.3

SGLang

How to use EmbeddedLLM/Mistral-7B-Merge-14-v0.3 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 "EmbeddedLLM/Mistral-7B-Merge-14-v0.3" \
    --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": "EmbeddedLLM/Mistral-7B-Merge-14-v0.3",
		"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 "EmbeddedLLM/Mistral-7B-Merge-14-v0.3" \
        --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": "EmbeddedLLM/Mistral-7B-Merge-14-v0.3",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use EmbeddedLLM/Mistral-7B-Merge-14-v0.3 with Docker Model Runner:
```
docker model run hf.co/EmbeddedLLM/Mistral-7B-Merge-14-v0.3
```

quantize gguf models

by rezzie-rich - opened Dec 30, 2023

Discussion

rezzie-rich

Dec 30, 2023

I appreciate the team's effort in producing this masterpiece. However, I'm having trouble finding the download page for various quantized models like Q_6_K or Q_8. I was also wondering when will it release the benchmarks since can't wait to see the model dominating the leaderboard.

thesunday

EmbeddedLLM org Jan 2, 2024

•

edited Jan 2, 2024

Hi there, this model is a result of model merging without using any "potentially contaminated" models (berkeley-nest/Starling-LM-7B-alpha, Q-bert/MetaMath-Cybertron-Starling, v1olet/v1olet_marcoroni-go-bruins-merge-7B), and thus does not score as well as the ones that do.
For a high-scoring model, you can checkout our EmbeddedLLM/Mistral-7B-Merge-14-v0.1 and EmbeddedLLM/Mistral-7B-Merge-14-v0.2 (average score of 72.77).
As this is meant as an experiment of model merging, we do not have plans to quantize the models.

thesunday changed discussion status to closed Jan 2, 2024

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