Instructions to use andysalerno/mistral-sft-v3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use andysalerno/mistral-sft-v3 with Transformers:

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

pipe = pipeline("text-generation", model="andysalerno/mistral-sft-v3")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("andysalerno/mistral-sft-v3")
model = AutoModelForCausalLM.from_pretrained("andysalerno/mistral-sft-v3")
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 andysalerno/mistral-sft-v3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "andysalerno/mistral-sft-v3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "andysalerno/mistral-sft-v3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/andysalerno/mistral-sft-v3

SGLang

How to use andysalerno/mistral-sft-v3 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 "andysalerno/mistral-sft-v3" \
    --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": "andysalerno/mistral-sft-v3",
		"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 "andysalerno/mistral-sft-v3" \
        --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": "andysalerno/mistral-sft-v3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use andysalerno/mistral-sft-v3 with Docker Model Runner:
```
docker model run hf.co/andysalerno/mistral-sft-v3
```

This is mistralai/Mistral-7B-v0.1, but with the special tokens added for ChatML, and then lightly finetuned with sft using a ChatML formatted dataset: andysalerno/ansalern-nectar-inputoutput

The training was very light, so while this model correctly follows ChatML formatting, it is not intended to be a chat model.

Rather, it is intended to be a base for further fine-tuning models that will use ChatML.

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	60.93
AI2 Reasoning Challenge (25-Shot)	61.35
HellaSwag (10-Shot)	82.23
MMLU (5-Shot)	63.40
TruthfulQA (0-shot)	48.49
Winogrande (5-shot)	77.66
GSM8k (5-shot)	32.45

Downloads last month: 143

Safetensors

Model size

7B params

Tensor type

BF16

Model tree for andysalerno/mistral-sft-v3

Base model

mistralai/Mistral-7B-v0.1

Finetuned

(926)

this model

Adapters

4 models

Dataset used to train andysalerno/mistral-sft-v3

Evaluation results

normalized accuracy on AI2 Reasoning Challenge (25-Shot)
test set Open LLM Leaderboard

61.350
normalized accuracy on HellaSwag (10-Shot)
validation set Open LLM Leaderboard

82.230
accuracy on MMLU (5-Shot)
test set Open LLM Leaderboard

63.400
mc2 on TruthfulQA (0-shot)
validation set Open LLM Leaderboard

48.490
accuracy on Winogrande (5-shot)
validation set Open LLM Leaderboard

77.660
accuracy on GSM8k (5-shot)
test set Open LLM Leaderboard

32.450