model_card_template.md

metadata

license: mit
tags:
  - amop-optimized
  - onnx

AMOP-Optimized ONNX Model: {repo_name}

This model was automatically optimized for CPU inference using the Adaptive Model Optimization Pipeline (AMOP).

• Base Model: {model_id}
• Optimization Date: {optimization_date}

Optimization Details

The following AMOP ONNX pipeline stages were applied:

• Pruning: {pruning_status} (Percentage: {pruning_percent}%)
• Quantization & ONNX Conversion: Enabled ({quant_type} Quantization)

How to Use

This model is in ONNX format and can be run with optimum-onnxruntime. Make sure you have optimum, onnxruntime, and transformers installed.

from optimum.onnxruntime import ORTModelForCausalLM
from transformers import AutoTokenizer

model_id = "{repo_id}"
model = ORTModelForCausalLM.from_pretrained(model_id)
tokenizer = AutoTokenizer.from_pretrained(model_id)

prompt = "The future of AI is"
inputs = tokenizer(prompt, return_tensors="pt")
gen_tokens = model.generate(**inputs)
print(tokenizer.batch_decode(gen_tokens))

AMOP Pipeline Log

Click to expand

{pipeline_log}