Instructions to use mistralai/Mistral-7B-Instruct-v0.2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mistralai/Mistral-7B-Instruct-v0.2 with Transformers:

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

pipe = pipeline("text-generation", model="mistralai/Mistral-7B-Instruct-v0.2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2")
model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2")
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]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use mistralai/Mistral-7B-Instruct-v0.2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Install mistral-common:
pip install --upgrade mistral-common
# Start the vLLM server:
vllm serve "mistralai/Mistral-7B-Instruct-v0.2" --tokenizer_mode mistral --config_format mistral --load_format mistral --tool-call-parser mistral --enable-auto-tool-choice
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mistralai/Mistral-7B-Instruct-v0.2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/mistralai/Mistral-7B-Instruct-v0.2

SGLang

How to use mistralai/Mistral-7B-Instruct-v0.2 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 "mistralai/Mistral-7B-Instruct-v0.2" \
    --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": "mistralai/Mistral-7B-Instruct-v0.2",
		"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 "mistralai/Mistral-7B-Instruct-v0.2" \
        --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": "mistralai/Mistral-7B-Instruct-v0.2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use mistralai/Mistral-7B-Instruct-v0.2 with Docker Model Runner:
```
docker model run hf.co/mistralai/Mistral-7B-Instruct-v0.2
```

Has anybody gotten this example to work for converting string data into valid JSON?

#62

by capnchat - opened Mar 7, 2024

Discussion

capnchat

Mar 7, 2024

There is an example here: https://www.promptingguide.ai/models/mistral-7b#chat-template-for-mistral-7b-instruct

In the example they show how string data such as the below...

name: John
lastname: Smith
address: #1 Samuel St.

Can be converted into JSON as below...

{
"address": "#1 Samuel St.",
"lastname": "Smith",
"name": "John"
}

I am wondering if maybe it is not working because I am using BitsAndBytes in 4-bit mode so maybe I am losing precision?

Below is my generation_config...

generation_config = model.generation_config
generation_config.max_new_tokens = 1024
generation_config.do_sample = True
generation_config.temperature = 0.1
generation_config.top_p = 0.1
generation_config.repetition_penalty = 1.2
generation_config.no_repeat_ngram_size = 2
generation_config.num_beams = 1
generation_config.num_return_sequences = 1
generation_config.pad_token_id = tokenizer.eos_token_id
generation_config.eos_token_id = tokenizer.eos_token_id

I have tried using apply_chat_template() as well as formatting the strings myself, but the output is always extremely squirrely and I am actually getting better results with v0.1 than v0.2

capnchat

Mar 7, 2024

And I just tested using the fireworks.ai instance mentioned in the tutorial and it works perfectly well irrespeoctive of the generations ettings so it has to be the BitsAndBytes config that is the culprit... will be able to verify once I get access to a better VM again!

jilijeanlouis

Mar 24, 2024

Try outlines:
https://outlines-dev.github.io/outlines/reference/vllm/

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