nm-testing/CatPPT-base-pruned50-quant-ds

CatPPT-base - DeepSparse

This repo contains model files for CatPPT-base 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 get in a good university?"
formatted_prompt =  f"<s>[INST]{prompt}[/INST]"

model = TextGeneration(model_path="hf:nm-testing/CatPPT-base-pruned50-quant-ds")

print(model(formatted_prompt, max_new_tokens=500).generations[0].text)
"""
To get into a good university, you need to follow a few steps.

1. Choose a field of study: Decide what you want to study. This will be your major.

2. Research universities: Look up universities that offer your major. You can use the internet to find out information about universities.

3. Look at requirements: Check what the universities require for admission. This may include grades, entrance exam scores, and other requirements.

4. Prepare: Work hard to meet the requirements. This may include studying hard, taking entrance exams, and maintaining a high GPA.

5. Apply: Apply to the universities you are interested in. This may include submitting applications, paying fees, and providing documents.

6. Wait: Wait for a response from the university. This may include waiting for acceptance letters, rejection letters, or offers.

7. Decide: If you are accepted, decide which university you want to attend.

8. Enroll: Enroll in the university you have chosen. This may include paying fees, registering for classes, and making a schedule.

By following these steps, you will be more likely to get into a good university.
"""

Prompt template

  <s>[INST] 
  {prompt}
   [/INST]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 rishiraj/CatPPT-base 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