Model Preparation

You can obtain a model in one of supported formats in many ways. The easiest one is to download it from an online database, such as TensorFlow Hub, Hugging Face, and Torchvision models. Now you have two options:

  • Skip model conversion and run inference directly from the source format. Conversion will still be performed but it will happen automatically and “under the hood.” This option, while convenient, offers lower performance and stability, as well as fewer optimization options.

  • Explicitly convert the model to OpenVINO IR. This approach offers the best possible results and is the recommended one, especially for for production-ready solutions. Explicit conversion can be done in two ways:

    • the Python API functions (openvino.convert_model and openvino.save_model)

    • the ovc command line tool.

    Once saved as OpenVINO IR (a set of .xml and .bin files), the model may be deployed with maximum performance. Because it is already optimized for OpenVINO inference, it can be read, compiled, and inferred with no additional delay.

Note

Model conversion API prior to OpenVINO 2023.1 is considered deprecated. Existing and new projects are recommended to transition to the new solutions, keeping in mind that they are not fully backwards compatible with openvino.tools.mo.convert_model or the mo CLI tool. For more details, see the Model Conversion API Transition Guide.

Convert a Model with Python: convert_model

The Model conversion API in Python uses the openvino.convert_model function, turning a given model into the openvino.Model object and loading it to memory. Now it can be: saved to a drive with openvino.save_model` or further optimized with NNCF prior to saving.

See how to use openvino.convert_model with models from some of the most popular public repositories:

   import openvino as ov
   import torch
   from torchvision.models import resnet50

   model = resnet50(weights='DEFAULT')

   # prepare input_data
   input_data = torch.rand(1, 3, 224, 224)

   ov_model = ov.convert_model(model, example_input=input_data)

   ###### Option 1: Save to OpenVINO IR:

   # save model to OpenVINO IR for later use
   ov.save_model(ov_model, 'model.xml')

   ###### Option 2: Compile and infer with OpenVINO:

   # compile model
   compiled_model = ov.compile_model(ov_model)

   # run inference
   result = compiled_model(input_data)

from transformers import BertTokenizer, BertModel

tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
model = BertModel.from_pretrained("bert-base-uncased")
text = "Replace me by any text you'd like."
encoded_input = tokenizer(text, return_tensors='pt')

import openvino as ov
ov_model = ov.convert_model(model, example_input={**encoded_input})

###### Option 1: Save to OpenVINO IR:

# save model to OpenVINO IR for later use
ov.save_model(ov_model, 'model.xml')

###### Option 2: Compile and infer with OpenVINO:

# compile model
compiled_model = ov.compile_model(ov_model)

# prepare input_data using HF tokenizer or your own tokenizer
# encoded_input is reused here for simplicity

# run inference
result = compiled_model({**encoded_input})

import tensorflow as tf
import openvino as ov

tf_model = tf.keras.applications.ResNet50(weights="imagenet")
ov_model = ov.convert_model(tf_model)

###### Option 1: Save to OpenVINO IR:

# save model to OpenVINO IR for later use
ov.save_model(ov_model, 'model.xml')

###### Option 2: Compile and infer with OpenVINO:

# compile model
compiled_model = ov.compile_model(ov_model)

# prepare input_data
import numpy as np
input_data = np.random.rand(1, 224, 224, 3)

# run inference
result = compiled_model(input_data)

import tensorflow as tf
import tensorflow_hub as hub
import openvino as ov

model = tf.keras.Sequential([
      hub.KerasLayer("https://tfhub.dev/google/imagenet/mobilenet_v1_100_224/classification/5")
])

# Check model page for information about input shape: https://tfhub.dev/google/imagenet/mobilenet_v1_100_224/classification/5
model.build([None, 224, 224, 3])

model.save('mobilenet_v1_100_224')  # use a temporary directory
ov_model = ov.convert_model('mobilenet_v1_100_224')

###### Option 1: Save to OpenVINO IR:

ov.save_model(ov_model, 'model.xml')

###### Option 2: Compile and infer with OpenVINO:

compiled_model = ov.compile_model(ov_model)

# prepare input_data
import numpy as np
input_data = np.random.rand(1, 224, 224, 3)

# run inference
result = compiled_model(input_data)

import onnx

model = onnx.hub.load("resnet50")
onnx.save(model, 'resnet50.onnx')  # use a temporary file for model

import openvino as ov
ov_model = ov.convert_model('resnet50.onnx')

###### Option 1: Save to OpenVINO IR:

# save model to OpenVINO IR for later use
ov.save_model(ov_model, 'model.xml')

###### Option 2: Compile and infer with OpenVINO:

# compile model
compiled_model = ov.compile_model(ov_model)

# prepare input_data
import numpy as np
input_data = np.random.rand(1, 3, 224, 224)

# run inference
result = compiled_model(input_data)

  • Saving the model, Option 1, is used as a separate step, outside of deployment. The file it provides is then used in the final software solution, resulting in maximum performance due to fewer dependencies and faster model loading.

  • Compiling the model, Option 2, provides a convenient way to quickly switch from framework-based code to OpenVINO-based code in your existing Python inference application. The converted model is not saved to IR but compiled and used for inference within the same application.

Before saving the model to OpenVINO IR, consider Post-training Optimization to achieve more efficient inference and smaller model size.

Convert a Model in CLI: ovc

ovc is a command-line model converter, combining the openvino.convert_model and openvino.save_model functionalities, providing the exact same results, if the same set of parameters is used for saving into OpenVINO IR. It converts files from one of the supported model formats to OpenVINO IR, which can then be read, compiled, and run by the final inference application.

Note

PyTorch models cannot be converted with ovc, use openvino.convert_model instead.

Additional Resources

The following articles describe in detail how to obtain and prepare your model depending on the source model type:

To achieve the best model inference performance and more compact OpenVINO IR representation follow:

If you are still using the legacy conversion API (mo or openvino.tools.mo.convert_model), please refer to the following materials: