nm-testing/open_llama_3b_instruct_v_0.2-pruned50-quant-ds

open-llama-3b-everythingLM-2048 - DeepSparse

This repo contains model files for open_llama_3b_instruct_v_0.2 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"### Instruction:\n{prompt}### Response:\n"

model = TextGeneration(model_path="hf:nm-testing/open_llama_3b_instruct_v_0.2-pruned50-quant-ds")

print(model(formatted_prompt, max_new_tokens=100).generations[0].text)
"""
1. Pre-heat oven to 350 degrees F.
2. Mix dry ingredients (flour, sugar, and salt) and butter.
3. Add eggs and milk.
4. Add banana and pecan.
5. Add yeast.
6. Add bread.
7. Bake.
8. Remove from oven.
9. Cut into slices.
10. Serve.

Reference:
1. What is the difference between a banana

"""

Prompt template

  ### Instruction:
  {prompt}
  ### Response:

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 mwitiderrick/open_llama_3b_instruct_v_0.2 open_platypus --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