nm-testing/glaive-coder-7b-pruned50-quant-ds

Glaive-coder-7b - DeepSparse

This repo contains model files for Glaive-coder-7b 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

template = "<s>[INST] {prompt} [/INST]"
prompt = "Write a quick sort algorithm in Python"

input_str = template.format(prompt=prompt)

model = TextGeneration(model_path="hf:nm-testing/glaive-coder-7b-pruned50-quant-ds")

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

"""
def quick_sort(arr):
    if len(arr) <= 1:
        return arr
    mid = len(arr) // 2
    arr[:mid] = sorted(arr[:mid] )
    left = arr[:mid]
    right = arr[mid:]
    quick_sort(left)
    quick_sort(right)
    left.extend(right)
    return left + right


print(quick_sort([5, 3, 1, 4, 2]))

This code will give you a sorted array. The quick_sort function sorts the first mid to mid element and the rest of the array. Then it calls the function again on the right part of the array. After that, it

"""

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 glaiveai/glaive-coder-7b open_platypus --recipe recipe.yaml --save True
python sparseml/src/sparseml/transformers/sparsification/obcq/export.py --sequence_length 4096 --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