Visual Question Answering and Image Captioning using BLIP and OpenVINO

This Jupyter notebook can be launched after a local installation only.

Humans perceive the world through vision and language. A longtime goal of AI is to build intelligent agents that can understand the world through vision and language inputs to communicate with humans through natural language. In order to achieve this goal, vision-language pre-training has emerged as an effective approach, where deep neural network models are pre-trained on large scale image-text datasets to improve performance on downstream vision-language tasks, such as image-text retrieval, image captioning, and visual question answering.

BLIP is a language-image pre-training framework for unified vision-language understanding and generation. BLIP achieves state-of-the-art results on a wide range of vision-language tasks. This tutorial demonstrates how to use BLIP for visual question answering and image captioning. An additional part of tutorial demonstrates how to speed up the model by applying 8-bit post-training quantization and data free int8 weight compression from NNCF (Neural Network Compression Framework) to OpenVINO IR models and infer optimized BLIP model via OpenVINO™ Toolkit.

The tutorial consists of the following parts:

1. Instantiate a BLIP model.

2. Convert the BLIP model to OpenVINO IR.

3. Run visual question answering and image captioning with OpenVINO.

4. Optimize BLIP model using NNCF

5. Compare original and optimized models

6. Launch interactive demo

Table of contents:

• Background

  • Image Captioning

  • Visual Question Answering

• Instantiate Model

• Convert Models to OpenVINO IR

  • Vision Model

  • Text Encoder

  • Text Decoder

• Run OpenVINO Model

  • Prepare Inference Pipeline

  • Select inference device

  • Image Captioning

  • Question Answering

• Optimize model using NNCF

  • Prepare dataset

  • Quantize vision model

  • Quantize text encoder

  • Compress weights of text decoder

  • Run optimized OpenVINO model

    • Image captioning

    • Question answering

  • Compare file sizes

  • Compare inference time of the FP16 and optimized models

• Interactive demo

Background

Visual language processing is a branch of artificial intelligence that focuses on creating algorithms designed to enable computers to more accurately understand images and their content.

Popular tasks include:

• Text to Image Retrieval - a semantic task that aims to find the most relevant image for a given text description.

• Image Captioning - a semantic task that aims to provide a text description for image content.

• Visual Question Answering - a semantic task that aims to answer questions based on image content.

As shown in the diagram below, these three tasks differ in the input provided to the AI system. For text-to-image retrieval, you have a predefined gallery of images for search and a user-requested text description (query). Image captioning can be represented as a particular case of visual question answering, where you have a predefined question “What is in the picture?” and various images provided by a user. For visual question answering, both the text-based question and image context are variables requested by a user.

This notebook does not focus on Text to Image retrieval. Instead, it considers Image Captioning and Visual Question Answering.

Image Captioning

Image Captioning is the task of describing the content of an image in words. This task lies at the intersection of computer vision and natural language processing. Most image captioning systems use an encoder-decoder framework, where an input image is encoded into an intermediate representation of the information in the image, and then decoded into a descriptive text sequence.

Visual Question Answering

Visual Question Answering (VQA) is the task of answering text-based questions about image content.

For a better understanding of how VQA works, let us consider a traditional NLP task like Question Answering, which aims to retrieve the answer to a question from a given text input. Typically, a question answering pipeline consists of three steps:

1. Question analysis - analysis of provided question in natural language form to understand the object in the question and additional context. For example, if you have a question like “How many bridges in Paris?”, question words “how many” gives a hint that the answer is more likely to be a number, “bridges” is the target object of the question and ” in Paris” serves as additional context for the search.

2. Build query for search - use analyzed results to formalize query for finding the most relevant information.

3. Perform a search in the knowledge base - send the query to a knowledge base, typically provided text documents or databases serve as a source of knowledge.

The difference between text-based question answering and visual question answering is that an image is used as context and the knowledge base.

Answering arbitrary questions about images is a complex problem because it requires involving a lot of computer vision sub-tasks. In the table below, you can find an example of questions and the required computer vision skills to find answers.

Computer vision task	Question examples
Object recognition	What is shown in the picture? What is it?
Object detection	Is there any object (dog, man, book) in the image? Where is … located?
Object and image attribute recognition	What color is an umbrella? Does this man wear glasses? Is there color in the image?
Scene recognition	Is it rainy? What celebration is pictured?
Object counting	How many players are there on the football field? How many steps are there on the stairs?
Activity recognition	Is the baby crying? What is the woman cooking? What are they doing?
Spatial relationships among objects	What is located between the sofa and the armchair? What is in the bottom left corner?
Commonsense reasoning	Does she have 100% vision? Does this person have children?
Knowledge-based reasoning	Is it a vegetarian pizza?
Text recognition	What is the title of the book? What is shown on the screen?

There are a lot of applications for visual question answering:

• Aid Visually Impaired Persons: VQA models can be used to reduce barriers for visually impaired people by helping them get information about images from the web and the real world.

• Education: VQA models can be used to improve visitor experiences at museums by enabling observers to directly ask questions they are interested in or to bring more interactivity to schoolbooks for children interested in acquiring specific knowledge.

• E-commerce: VQA models can retrieve information about products using photos from online stores.

• Independent expert assessment: VQA models can be provide objective assessments in sports competitions, medical diagnosis, and forensic examination.

Instantiate Model

The BLIP model was proposed in the BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation paper.

blip.gif

To pre-train a unified vision-language model with both understanding and generation capabilities, BLIP introduces a multimodal mixture of an encoder-decoder and a multi-task model which can operate in one of the three modes:

• Unimodal encoders, which separately encode images and text. The image encoder is a vision transformer. The text encoder is the same as BERT.

• Image-grounded text encoder, which injects visual information by inserting a cross-attention layer between the self-attention layer and the feed-forward network for each transformer block of the text encoder.

• Image-grounded text decoder, which replaces the bi-directional self-attention layers in the text encoder with causal self-attention layers.

More details about the model can be found in the research paper, Salesforce blog, GitHub repo and Hugging Face model documentation.

In this tutorial, you will use the blip-vqa-base model available for download from Hugging Face. The same actions are also applicable to other similar models from the BLIP family. Although this model class is designed to perform question answering, its components can also be reused for image captioning.

To start working with the model, you need to instantiate the BlipForQuestionAnswering class, using from_pretrained method. BlipProcessor is a helper class for preparing input data for both text and vision modalities and postprocessing of generation results.

import platform

%pip install -q --extra-index-url https://download.pytorch.org/whl/cpu "torch>=2.1.0" torchvision "transformers>=4.26.0" "gradio>=4.19" "openvino>=2023.3.0" "datasets>==2.14.6" "nncf>=2.8.1" "tqdm"
if platform.system() != "Windows":
    %pip install -q "matplotlib>=3.4"
else:
    %pip install -q "matplotlib>=3.4,<3.7"

import time
from PIL import Image
from transformers import BlipProcessor, BlipForQuestionAnswering

# Fetch `notebook_utils` module
import requests

r = requests.get(
    url="https://raw.githubusercontent.com/openvinotoolkit/openvino_notebooks/latest/utils/notebook_utils.py",
)
open("notebook_utils.py", "w").write(r.text)
from notebook_utils import download_file

# get model and processor
processor = BlipProcessor.from_pretrained("Salesforce/blip-vqa-base")
model = BlipForQuestionAnswering.from_pretrained("Salesforce/blip-vqa-base")

# setup test input: download and read image, prepare question
img_url = "https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg"
download_file(img_url, "demo.jpg")
raw_image = Image.open("demo.jpg").convert("RGB")
question = "how many dogs are in the picture?"
# preprocess input data
inputs = processor(raw_image, question, return_tensors="pt")

start = time.perf_counter()
# perform generation
out = model.generate(**inputs)
end = time.perf_counter() - start

# postprocess result
answer = processor.decode(out[0], skip_special_tokens=True)

print(f"Processing time: {end:.4f} s")

Processing time: 0.3707 s

from pathlib import Path

if not Path("./utils.py").exists():
    download_file(url="https://raw.githubusercontent.com/openvinotoolkit/openvino_notebooks/latest/notebooks/blip-visual-language-processing/utils.py")
from utils import visualize_results

fig = visualize_results(raw_image, answer, question)

Convert Models to OpenVINO IR

Starting from OpenVINO 2023.0 release, OpenVINO supports direct PyTorch models conversion to OpenVINO Intermediate Representation (IR) format to take the advantage of advanced OpenVINO optimization tools and features. You need to provide a model object, input data for model tracing to OpenVINO Model Conversion API. ov.convert_model function convert PyTorch model instance to ov.Model object that can be used for compilation on device or saved on disk using ov.save_model in compressed to FP16 format.

The model consists of three parts:

• vision_model - an encoder for image representation.

• text_encoder - an encoder for input query, used for question answering and text-to-image retrieval only.

• text_decoder - a decoder for output answer.

To be able to perform multiple tasks, using the same model components, you should convert each part independently.

Vision Model

The vision model accepts float input tensors with the [1,3,384,384] shape, containing RGB image pixel values normalized in the [0,1] range.

import torch
from pathlib import Path
import openvino as ov

VISION_MODEL_OV = Path("blip_vision_model.xml")
vision_model = model.vision_model
vision_model.eval()

# check that model works and save it outputs for reusage as text encoder input
with torch.no_grad():
    vision_outputs = vision_model(inputs["pixel_values"])

# if openvino model does not exist, convert it to IR
if not VISION_MODEL_OV.exists():
    # export pytorch model to ov.Model
    with torch.no_grad():
        ov_vision_model = ov.convert_model(vision_model, example_input=inputs["pixel_values"])
    # save model on disk for next usages
    ov.save_model(ov_vision_model, VISION_MODEL_OV)
    print(f"Vision model successfuly converted and saved to {VISION_MODEL_OV}")
else:
    print(f"Vision model will be loaded from {VISION_MODEL_OV}")

/home/ltalamanova/tmp_venv/lib/python3.11/site-packages/transformers/modeling_utils.py:4225: FutureWarning: _is_quantized_training_enabled is going to be deprecated in transformers 4.39.0. Please use model.hf_quantizer.is_trainable instead
  warnings.warn(

Vision model successfuly converted and saved to blip_vision_model.xml

Text Encoder

The text encoder is used by visual question answering tasks to build a question embedding representation. It takes input_ids with a tokenized question and output image embeddings obtained from the vision model and attention masks for them.

TEXT_ENCODER_OV = Path("blip_text_encoder.xml")


text_encoder = model.text_encoder
text_encoder.eval()

# if openvino model does not exist, convert it to IR
if not TEXT_ENCODER_OV.exists():
    # prepare example inputs
    image_embeds = vision_outputs[0]
    image_attention_mask = torch.ones(image_embeds.size()[:-1], dtype=torch.long)
    input_dict = {
        "input_ids": inputs["input_ids"],
        "attention_mask": inputs["attention_mask"],
        "encoder_hidden_states": image_embeds,
        "encoder_attention_mask": image_attention_mask,
    }
    # export PyTorch model
    with torch.no_grad():
        ov_text_encoder = ov.convert_model(text_encoder, example_input=input_dict)
    # save model on disk for next usages
    ov.save_model(ov_text_encoder, TEXT_ENCODER_OV)
    print(f"Text encoder successfuly converted and saved to {TEXT_ENCODER_OV}")
else:
    print(f"Text encoder will be loaded from {TEXT_ENCODER_OV}")

Text encoder successfuly converted and saved to blip_text_encoder.xml

Text Decoder

The text decoder is responsible for generating the sequence of tokens to represent model output (answer to question or caption), using an image (and question, if required) representation. The generation approach is based on the assumption that the probability distribution of a word sequence can be decomposed into the product of conditional next word distributions. In other words, model predicts the next token in the loop guided by previously generated tokens until the stop-condition will be not reached (generated sequence of maximum length or end of string token obtained). The way the next token will be selected over predicted probabilities is driven by the selected decoding methodology. You can find more information about the most popular decoding methods in this blog. The entry point for the generation process for models from the Hugging Face Transformers library is the generate method. You can find more information about its parameters and configuration in the documentation. To preserve flexibility in the selection decoding methodology, you will convert only model inference for one step.

To optimize the generation process and use memory more efficiently, the use_cache=True option is enabled. Since the output side is auto-regressive, an output token hidden state remains the same once computed for every further generation step. Therefore, recomputing it every time you want to generate a new token seems wasteful. With the cache, the model saves the hidden state once it has been computed. The model only computes the one for the most recently generated output token at each time step, re-using the saved ones for hidden tokens. This reduces the generation complexity from O(n^3) to O(n^2) for a transformer model. More details about how it works can be found in this article. With this option, the model gets the previous step’s hidden states as input and additionally provides hidden states for the current step as output. Initially, you have no previous step hidden states, so the first step does not require you to provide them, but we should initialize them by default values. In PyTorch, past hidden state outputs are represented as a list of pairs (hidden state for key, hidden state for value] for each transformer layer in the model. OpenVINO model does not support nested outputs, they will be flattened.

Similar to text_encoder, text_decoder can work with input sequences of different lengths and requires preserving dynamic input shapes.

text_decoder = model.text_decoder
text_decoder.eval()

TEXT_DECODER_OV = Path("blip_text_decoder_with_past.xml")

# prepare example inputs
input_ids = torch.tensor([[30522]])  # begin of sequence token id
attention_mask = torch.tensor([[1]])  # attention mask for input_ids
encoder_hidden_states = torch.rand((1, 10, 768))  # encoder last hidden state from text_encoder
encoder_attention_mask = torch.ones((1, 10), dtype=torch.long)  # attention mask for encoder hidden states

input_dict = {
    "input_ids": input_ids,
    "attention_mask": attention_mask,
    "encoder_hidden_states": encoder_hidden_states,
    "encoder_attention_mask": encoder_attention_mask,
}
text_decoder_outs = text_decoder(**input_dict)
# extend input dictionary with hidden states from previous step
input_dict["past_key_values"] = text_decoder_outs["past_key_values"]

text_decoder.config.torchscript = True
if not TEXT_DECODER_OV.exists():
    # export PyTorch model
    with torch.no_grad():
        ov_text_decoder = ov.convert_model(text_decoder, example_input=input_dict)
    # save model on disk for next usages
    ov.save_model(ov_text_decoder, TEXT_DECODER_OV)
    print(f"Text decoder successfuly converted and saved to {TEXT_DECODER_OV}")
else:
    print(f"Text decoder will be loaded from {TEXT_DECODER_OV}")

/home/ltalamanova/tmp_venv/lib/python3.11/site-packages/transformers/models/blip/modeling_blip_text.py:635: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  if causal_mask.shape[1] < attention_mask.shape[1]:
/home/ltalamanova/tmp_venv/lib/python3.11/site-packages/torch/jit/_trace.py:165: UserWarning: The .grad attribute of a Tensor that is not a leaf Tensor is being accessed. Its .grad attribute won't be populated during autograd.backward(). If you indeed want the .grad field to be populated for a non-leaf Tensor, use .retain_grad() on the non-leaf Tensor. If you access the non-leaf Tensor by mistake, make sure you access the leaf Tensor instead. See github.com/pytorch/pytorch/pull/30531 for more informations. (Triggered internally at aten/src/ATen/core/TensorBody.h:489.)
  if a.grad is not None:

Text decoder successfuly converted and saved to blip_text_decoder_with_past.xml

Run OpenVINO Model

Prepare Inference Pipeline

As discussed before, the model consists of several blocks which can be reused for building pipelines for different tasks. In the diagram below, you can see how image captioning works:

The visual model accepts the image preprocessed by BlipProcessor as input and produces image embeddings, which are directly passed to the text decoder for generation caption tokens. When generation is finished, output sequence of tokens is provided to BlipProcessor for decoding to text using a tokenizer.

The pipeline for question answering looks similar, but with additional question processing. In this case, image embeddings and question tokenized by BlipProcessor are provided to the text encoder and then multimodal question embedding is passed to the text decoder for performing generation of answers.

The next step is implementing both pipelines using OpenVINO models.

# create OpenVINO Core object instance
core = ov.Core()

Select inference device

select device from dropdown list for running inference using OpenVINO

import ipywidgets as widgets

device = widgets.Dropdown(
    options=core.available_devices + ["AUTO"],
    value="AUTO",
    description="Device:",
    disabled=False,
)

device

huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
To disable this warning, you can either:
    - Avoid using tokenizers before the fork if possible
    - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)

Dropdown(description='Device:', index=4, options=('CPU', 'GPU.0', 'GPU.1', 'GPU.2', 'AUTO'), value='AUTO')

# load models on device
ov_vision_model = core.compile_model(VISION_MODEL_OV, device.value)
ov_text_encoder = core.compile_model(TEXT_ENCODER_OV, device.value)
ov_text_decoder_with_past = core.compile_model(TEXT_DECODER_OV, device.value)

from functools import partial
from blip_model import text_decoder_forward

text_decoder.forward = partial(text_decoder_forward, ov_text_decoder_with_past=ov_text_decoder_with_past)

The model helper class has two methods for generation: generate_answer - used for visual question answering, generate_caption - used for caption generation. For initialization, model class accepts compiled OpenVINO models for the text encoder, vision model and text decoder, and also configuration for generation and initial token for decoder work.

if not Path("./blip_model.py").exists():
    download_file(url="https://raw.githubusercontent.com/openvinotoolkit/openvino_notebooks/latest/notebooks/blip-visual-language-processing/blip_model.py")
from blip_model import OVBlipModel

ov_model = OVBlipModel(model.config, model.decoder_start_token_id, ov_vision_model, ov_text_encoder, text_decoder)
out = ov_model.generate_answer(**inputs, max_length=20)

Now, the model is ready for generation.

Image Captioning

out = ov_model.generate_caption(inputs["pixel_values"], max_length=20)
caption = processor.decode(out[0], skip_special_tokens=True)
fig = visualize_results(raw_image, caption)

Question Answering

start = time.perf_counter()
out = ov_model.generate_answer(**inputs, max_length=20)
end = time.perf_counter() - start
answer = processor.decode(out[0], skip_special_tokens=True)
fig = visualize_results(raw_image, answer, question)

print(f"Processing time: {end:.4f}")

Processing time: 0.1186

Optimize model using NNCF

NNCF enables post-training quantization by adding the quantization layers into the model graph and then using a subset of the training dataset to initialize the parameters of these additional quantization layers. The framework is designed so that modifications to your original training code are minor.

The optimization process contains the following steps:

1. Create a dataset for quantization.

2. Run nncf.quantize to get a quantized model from the pre-trained FP16 model.

3. Serialize the INT8 model using openvino.save_model function.

NOTE: Quantization is time and memory consuming operation. Running quantization code below may take some time. You can disable it using widget below:

to_quantize = widgets.Checkbox(
    value=True,
    description="Quantization",
    disabled=False,
)

to_quantize

Checkbox(value=True, description='Quantization')

VISION_MODEL_OV_INT8 = Path(str(VISION_MODEL_OV).replace(".xml", "_int8.xml"))
TEXT_ENCODER_OV_INT8 = Path(str(TEXT_ENCODER_OV).replace(".xml", "_int8.xml"))
TEXT_DECODER_OV_INT8 = Path(str(TEXT_DECODER_OV).replace(".xml", "_int8.xml"))
int8_model = None

# Fetch skip_kernel_extension module
r = requests.get(
    url="https://raw.githubusercontent.com/openvinotoolkit/openvino_notebooks/latest/utils/skip_kernel_extension.py",
)
open("skip_kernel_extension.py", "w").write(r.text)

%load_ext skip_kernel_extension

Prepare dataset

The VQAv2 is a dataset containing open-ended questions about images. These questions require an understanding of vision, language and commonsense knowledge to answer.

%%skip not $to_quantize.value

import numpy as np
from datasets import load_dataset
from tqdm.notebook import tqdm

def preprocess_batch(batch, vision_model, inputs_info):
    """
    Preprocesses a dataset batch by loading and transforming image and text data.
    VQAv2 dataset contains multiple questions to image.
    To reduce dataset preparation time we will store preprocessed images in `inputs_info`.
    """
    image_id = batch["image_id"]
    if image_id in inputs_info:
        inputs = processor(text=batch['question'], return_tensors="np")
        pixel_values = inputs_info[image_id]["pixel_values"]
        encoder_hidden_states = inputs_info[image_id]["encoder_hidden_states"]
    else:
        inputs = processor(images=batch["image"], text=batch["question"], return_tensors="np")
        pixel_values = inputs["pixel_values"]
        encoder_hidden_states = vision_model(pixel_values)[vision_model.output(0)]
        inputs_info[image_id] = {
            "pixel_values": pixel_values,
            "encoder_hidden_states": encoder_hidden_states,
            "text_encoder_inputs": []
        }

    text_encoder_inputs = {
        "input_ids": inputs["input_ids"],
        "attention_mask": inputs["attention_mask"]
    }
    inputs_info[image_id]["text_encoder_inputs"].append(text_encoder_inputs)


def prepare_input_data(dataloader, vision_model, opt_init_steps):
    """
    Store calibration subset in List to reduce quantization time.
    """
    inputs_info = {}
    for idx, batch in enumerate(tqdm(dataloader, total=opt_init_steps, desc="Prepare calibration data")):
        preprocess_batch(batch, vision_model, inputs_info)

    calibration_subset = []
    for image_id in inputs_info:
        pixel_values = inputs_info[image_id]["pixel_values"]
        encoder_hidden_states = inputs_info[image_id]["encoder_hidden_states"]
        encoder_attention_mask = np.ones(encoder_hidden_states.shape[:-1], dtype=int)
        for text_encoder_inputs in inputs_info[image_id]["text_encoder_inputs"]:
            text_encoder_inputs["encoder_hidden_states"] = encoder_hidden_states
            text_encoder_inputs["encoder_attention_mask"] = encoder_attention_mask
            blip_inputs = {
                "vision_model_inputs": {"pixel_values": pixel_values},
                "text_encoder_inputs": text_encoder_inputs,
            }
            calibration_subset.append(blip_inputs)
    return calibration_subset


def prepare_dataset(vision_model, opt_init_steps=300, streaming=False):
    """
    Prepares a vision-text dataset for quantization.
    """
    split = f"train[:{opt_init_steps}]" if not streaming else "train"
    dataset = load_dataset("HuggingFaceM4/VQAv2", split=split, streaming=streaming)
    dataset = dataset.shuffle(seed=42)
    if streaming:
        dataset = dataset.take(opt_init_steps)
    calibration_subset = prepare_input_data(dataset, vision_model, opt_init_steps)
    return calibration_subset

Loading and processing the dataset in streaming mode may take a long time and depends on your internet connection.

%%skip not $to_quantize.value

import nncf

comp_vision_model = core.compile_model(VISION_MODEL_OV, device.value)
calibration_data = prepare_dataset(comp_vision_model)

Quantize vision model

%%skip not $to_quantize.value

vision_dataset = nncf.Dataset(calibration_data, lambda x: x["vision_model_inputs"])
vision_model = core.read_model(VISION_MODEL_OV)

quantized_model = nncf.quantize(
    model=vision_model,
    calibration_dataset=vision_dataset,
    model_type=nncf.ModelType.TRANSFORMER
)

ov.save_model(quantized_model, VISION_MODEL_OV_INT8)

Output()

Output()

INFO:nncf:36 ignored nodes were found by name in the NNCFGraph
INFO:nncf:48 ignored nodes were found by name in the NNCFGraph

Output()

Output()

Quantize text encoder

%%skip not $to_quantize.value

text_encoder_dataset = nncf.Dataset(calibration_data, lambda x: x["text_encoder_inputs"])
text_encoder_model = core.read_model(TEXT_ENCODER_OV)

quantized_model = nncf.quantize(
    model=text_encoder_model,
    calibration_dataset=text_encoder_dataset,
    model_type=nncf.ModelType.TRANSFORMER
)
ov.save_model(quantized_model, TEXT_ENCODER_OV_INT8)

Output()

Output()

INFO:nncf:72 ignored nodes were found by name in the NNCFGraph
INFO:nncf:73 ignored nodes were found by name in the NNCFGraph

Output()

Output()

Compress weights of text decoder

The quantization of the text decoder leads to significant accuracy loss. Instead of post-training quantization, we can use data free weights compression to reduce the model footprint.

The optimization process contains the following steps:

1. Run nncf.compress_weights to get a model with compressed weights.

2. Serialize the OpenVINO model using openvino.save_model function.

%%skip not $to_quantize.value

text_decoder_model = core.read_model(TEXT_DECODER_OV)
compressed_text_decoder = nncf.compress_weights(text_decoder_model)
ov.save_model(compressed_text_decoder, str(TEXT_DECODER_OV_INT8))

INFO:nncf:Statistics of the bitwidth distribution:
+--------------+---------------------------+-----------------------------------+
| Num bits (N) | % all parameters (layers) |    % ratio-defining parameters    |
|              |                           |             (layers)              |
+==============+===========================+===================================+
| 8            | 100% (124 / 124)          | 100% (124 / 124)                  |
+--------------+---------------------------+-----------------------------------+

Output()

Run optimized OpenVINO model

The steps for making predictions with the optimized OpenVINO BLIP model are similar to the PyTorch model. Let us check the model result using the same input data like for model before quantization

%%skip not $to_quantize.value

q_ov_vision_model = core.compile_model(VISION_MODEL_OV_INT8, device.value)
q_ov_text_encoder = core.compile_model(TEXT_ENCODER_OV_INT8, device.value)
q_ov_text_decoder_with_past = core.compile_model(TEXT_DECODER_OV_INT8, device.value)

%%skip not $to_quantize.value

from functools import partial
from transformers import BlipForQuestionAnswering
from blip_model import OVBlipModel, text_decoder_forward

model = BlipForQuestionAnswering.from_pretrained("Salesforce/blip-vqa-base")
text_decoder = model.text_decoder
text_decoder.eval()

text_decoder.forward = partial(text_decoder_forward, ov_text_decoder_with_past=q_ov_text_decoder_with_past)
int8_model = OVBlipModel(model.config, model.decoder_start_token_id, q_ov_vision_model, q_ov_text_encoder, text_decoder)

%%skip not $to_quantize.value

raw_image = Image.open("demo.jpg").convert('RGB')
question = "how many dogs are in the picture?"
# preprocess input data
inputs = processor(raw_image, question, return_tensors="pt")

Image captioning

%%skip not $to_quantize.value

out = int8_model.generate_caption(inputs["pixel_values"], max_length=20)
caption = processor.decode(out[0], skip_special_tokens=True)
fig = visualize_results(raw_image, caption)

Question answering

%%skip not $to_quantize.value

out = int8_model.generate_answer(**inputs, max_length=20)
answer = processor.decode(out[0], skip_special_tokens=True)
fig = visualize_results(raw_image, answer, question)

%%skip not $to_quantize.value

def calculate_compression_rate(ov_model_path):
    fp16_ir_model_size = Path(ov_model_path).with_suffix(".bin").stat().st_size / 1024
    int8_model_path = str(ov_model_path).replace(".xml", "_int8.xml")
    quantized_model_size = Path(int8_model_path).with_suffix(".bin").stat().st_size / 1024
    print(f'{ov_model_path.as_posix().split(".")[0]}')
    print(f"    * FP16 IR model size: {fp16_ir_model_size:.2f} KB")
    print(f"    * INT8 model size: {quantized_model_size:.2f} KB")
    print(f"    * Model compression rate: {fp16_ir_model_size / quantized_model_size:.3f}")

%%skip not $to_quantize.value

for fp16_path in [VISION_MODEL_OV, TEXT_ENCODER_OV, TEXT_DECODER_OV]:
    calculate_compression_rate(fp16_path)

blip_vision_model
    * FP16 IR model size: 168145.70 KB
    * INT8 model size: 84915.48 KB
    * Model compression rate: 1.980
blip_text_encoder
    * FP16 IR model size: 268087.16 KB
    * INT8 model size: 134676.75 KB
    * Model compression rate: 1.991
blip_text_decoder_with_past
    * FP16 IR model size: 269303.35 KB
    * INT8 model size: 135302.53 KB
    * Model compression rate: 1.990

To measure the inference performance of the FP16 and INT8 models, we use median inference time on 100 samples of the calibration dataset. So we can approximately estimate the speed up of the dynamic quantized models.

NOTE: For the most accurate performance estimation, it is recommended to run benchmark_app in a terminal/command prompt after closing other applications with static shapes.

%%skip not $to_quantize.value

import time
import torch

def calculate_inference_time(blip_model, calibration_data, generate_caption):
    inference_time = []
    for inputs in calibration_data:
        pixel_values = torch.from_numpy(inputs["vision_model_inputs"]["pixel_values"])
        input_ids = torch.from_numpy(inputs["text_encoder_inputs"]["input_ids"])
        attention_mask = torch.from_numpy(inputs["text_encoder_inputs"]["attention_mask"])

        start = time.perf_counter()
        if generate_caption:
            _ = blip_model.generate_caption(pixel_values, max_length=20)
        else:
            _ = blip_model.generate_answer(pixel_values=pixel_values, input_ids=input_ids, attention_mask=attention_mask, max_length=20)
        end = time.perf_counter()
        delta = end - start
        inference_time.append(delta)
    return np.median(inference_time)

%%skip not $to_quantize.value

fp_original_model = BlipForQuestionAnswering.from_pretrained("Salesforce/blip-vqa-base")
fp_text_decoder = fp_original_model.text_decoder
fp_text_decoder.eval()

comp_text_encoder = core.compile_model(TEXT_ENCODER_OV, device.value)
comp_text_decoder_with_past = core.compile_model(TEXT_DECODER_OV, device.value)
fp_text_decoder.forward = partial(text_decoder_forward, ov_text_decoder_with_past=comp_text_decoder_with_past)
fp16_model = OVBlipModel(model.config, model.decoder_start_token_id, comp_vision_model, comp_text_encoder, fp_text_decoder)

%%skip not $to_quantize.value

validation_data = calibration_data[:100]

int8_caption_latency = calculate_inference_time(int8_model, validation_data, generate_caption=True)
fp16_caption_latency = calculate_inference_time(fp16_model, validation_data, generate_caption=True)

print(f"Image Captioning speed up: {fp16_caption_latency / int8_caption_latency:.3f}")

Image Captioning speed up: 1.254

%%skip not $to_quantize.value

int8_generate_answer_latency = calculate_inference_time(int8_model, validation_data, generate_caption=False)
fp16_generate_answer_latency = calculate_inference_time(fp16_model, validation_data, generate_caption=False)
print(f"Question Answering speed up: {fp16_generate_answer_latency / int8_generate_answer_latency:.3f}")

Question Answering speed up: 1.715

Interactive demo

Please select below whether you would like to use the quantized model to launch the interactive demo.

use_quantized_model = widgets.Checkbox(
    description="Use quantized model",
    value=int8_model is not None,
    disabled=int8_model is None,
)

use_quantized_model

Checkbox(value=True, description='Use quantized model')

import gradio as gr

ov_model = int8_model if use_quantized_model.value else ov_model


def generate_answer(img, question):
    if img is None:
        raise gr.Error("Please upload an image or choose one from the examples list")
    start = time.perf_counter()
    inputs = processor(img, question, return_tensors="pt")
    output = ov_model.generate_answer(**inputs, max_length=20) if len(question) else ov_model.generate_caption(inputs["pixel_values"], max_length=20)
    answer = processor.decode(output[0], skip_special_tokens=True)
    elapsed = time.perf_counter() - start
    html = f"<p>Processing time: {elapsed:.4f}</p>"
    return answer, html


demo = gr.Interface(
    generate_answer,
    [
        gr.Image(label="Image"),
        gr.Textbox(
            label="Question",
            info="If this field is empty, an image caption will be generated",
        ),
    ],
    [gr.Text(label="Answer"), gr.HTML()],
    examples=[["demo.jpg", ""], ["demo.jpg", question]],
    allow_flagging="never",
)
try:
    demo.launch(debug=False)
except Exception:
    demo.launch(share=True, debug=False)
# if you are launching remotely, specify server_name and server_port
# demo.launch(server_name='your server name', server_port='server port in int')
# Read more in the docs: https://gradio.app/docs/