metadata

base_model: mncai/mistral-7b-dpo-merge-v1.1
inference: false
model_type: mistral
prompt_template: |
  <|user|>\n
  {prompt}
  |assistant|>\n
quantized_by: mwitiderrick
tags:
  - deepsparse

Mistral-7b-dpo-merge-v1.1 - DeepSparse

This repo contains model files for Mistral-7b-dpo-merge-v1.1 optimized for DeepSparse, a CPU inference runtime for sparse models.

This model was quantized and pruned with SparseGPT, using SparseML.

Inference

Install DeepSparse LLM for fast inference on CPUs:

pip install deepsparse-nightly[llm]

Run in a Python pipeline:

from deepsparse import TextGeneration

prompt = "How to make banana bread?"
formatted_prompt =  f"<|user|>\n{prompt}\n<|assistant|>\n"

model = TextGeneration(model_path="hf:nm-testing/mistral-7b-dpo-merge-v1.1-pruned50-quant-dst")

print(model(formatted_prompt, max_new_tokens=200).generations[0].text)
"""
To make banana bread, you'll need the following ingredients:

- 1 cup (250 ml) all-purpose flour
- 1 cup (250 ml) granulated sugar
- 1/2 cup (125 ml) butter, melted
- 2 large eggs
- 1 cup (250 ml) mashed ripe bananas
- 1 teaspoon (5 ml) baking soda

Here are the steps to make banana bread:

1. Preheat the oven to 350°F (175°C).
2. Grease a loaf pan with butter or nonstick spray.
3. In a large bowl, whisk together the flour, baking soda, and salt.
4. In another bowl, cream together the butter, sugar, and eggs.
5. Add the mashed bananas to the mixture and stir until
"""

Prompt template


  <|user|>\n
  {prompt}
  |assistant|>\n

Sparsification

For details on how this model was sparsified, see the recipe.yaml in this repo and follow the instructions below.

git clone https://github.com/neuralmagic/sparseml
pip install -e "sparseml[transformers]"
python sparseml/src/sparseml/transformers/sparsification/obcq/obcq.py mncai/mistral-7b-dpo-merge-v1.1 open_platypus --precision float16 --recipe recipe.yaml --save True
python sparseml/src/sparseml/transformers/sparsification/obcq/export.py --task text-generation --model_path obcq_deployment 
cp deployment/model.onnx deployment/model-orig.onnx

Run this kv-cache injection to speed up the model at inference by caching the Key and Value states:

import os
import onnx
from sparseml.exporters.kv_cache_injector import KeyValueCacheInjector
input_file = "deployment/model-orig.onnx"
output_file = "deployment/model.onnx"
model = onnx.load(input_file, load_external_data=False)
model = KeyValueCacheInjector(model_path=os.path.dirname(input_file)).apply(model)
onnx.save(model, output_file)
print(f"Modified model saved to: {output_file}")

Follow the instructions on our One Shot With SparseML page for a step-by-step guide for performing one-shot quantization of large language models.

Slack

For further support, and discussions on these models and AI in general, join Neural Magic's Slack Community