README.md

metadata

base_model: NousResearch/Llama-2-7b-chat-hf
inference: false
model_type: llama
prompt_template: |
  <s>[INST] 
  {prompt}
   [/INST]
quantized_by: mwitiderrick
tags:
  - deepsparse

Llama-2-7b-chat-hf - DeepSparse

This repo contains model files for Llama-2-7b-hf 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"<s>[INST]{prompt}[/INST]"

model = TextGeneration(model_path="hf:nm-testing/Llama2-7b-chat-pruned50-qunat-ds")

print(model(formatted_prompt, max_new_tokens=200).generations[0].text)
"""

"""

Prompt template

<s>[INST]
<prompt>
[/INST]

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 NousResearch/Llama-2-7b-chat-hf 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