Phi-3.5-vision-instruct-int4-ov

• Model creator: Microsoft
• Original model: Phi-3.5-vision-instruct

Description

This is microsoft/Phi-3.5-vision-instruct model converted to the OpenVINO™ IR (Intermediate Representation) format with weights compressed to INT4 using Activation Aware Quantization (AWQ) by NNCF.

Quantization Parameters

Weight compression was performed using nncf.compress_weights with the following parameters:

• mode: INT4_ASYM
• ratio: 1.0
• group_size: 128
• awq: True
• dataset: contextual
• num_samples: 32

Compatibility

The provided OpenVINO™ IR model is compatible with:

• OpenVINO version 2025.0.0 and higher
• Optimum Intel 1.21.0 and higher

Running Model Inference with Optimum Intel

1. Install packages required for using Optimum Intel integration with the OpenVINO backend:

pip install --pre -U --extra-index-url https://storage.openvinotoolkit.org/simple/wheels/pre-release openvino_tokenizers openvino

pip install git+https://github.com/huggingface/optimum-intel.git

2. Run model inference

from PIL import Image 
import requests 
from optimum.intel.openvino import OVModelForVisualCausalLM
from transformers import AutoProcessor, TextStreamer

model_id = "OpenVINO/Phi-3.5-vision-instruct-int4-ov"

processor = AutoProcessor.from_pretrained(model_id, trust_remote_code=True)

ov_model = OVModelForVisualCausalLM.from_pretrained(model_id, trust_remote_code=True)
prompt = "<|image_1|>\nWhat is unusual on this picture?"

url = "https://github.com/openvinotoolkit/openvino_notebooks/assets/29454499/d5fbbd1a-d484-415c-88cb-9986625b7b11"
image = Image.open(requests.get(url, stream=True).raw)

inputs = ov_model.preprocess_inputs(text=prompt, image=image, processor=processor)

generation_args = { 
    "max_new_tokens": 50, 
    "temperature": 0.0, 
    "do_sample": False,
    "streamer": TextStreamer(processor.tokenizer, skip_prompt=True, skip_special_tokens=True)
} 

generate_ids = ov_model.generate(**inputs, 
  eos_token_id=processor.tokenizer.eos_token_id, 
  **generation_args
)

generate_ids = generate_ids[:, inputs['input_ids'].shape[1]:]
response = processor.batch_decode(generate_ids, 
  skip_special_tokens=True, 
  clean_up_tokenization_spaces=False)[0]

Limitations

Check the original model card for limitations.

Legal information

The original model is distributed under MIT license. More details can be found in original model card.