add basic support and optimization for qwen2-vl

[MeouSker77]

Description

add basic support and optimization for qwen2-vl

1. Why the change?

2. User API changes

this model requires transformers 4.45

pip install transformers==4.45
pip install qwen_vl_utils

A smiple example (this):

from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor
from qwen_vl_utils import process_vision_info

import time
import torch
from ipex_llm import optimize_model

model_path = r"Qwen/Qwen2-VL-7B-Instruct"

model = Qwen2VLForConditionalGeneration.from_pretrained(model_path)

model = optimize_model(model, modules_to_not_convert=["visual"])    # default to 'sym_int4'
model = model.float().eval()    # use .float() for better output, use .half() for better speed
# print(model)

model = model.to('xpu')

# default processer
# processor = AutoProcessor.from_pretrained(model_path)

# The default range for the number of visual tokens per image in the model is 4-16384. You can set min_pixels and max_pixels according to your needs, such as a token count range of 256-1280, to balance speed and memory usage.
min_pixels = 256*28*28
max_pixels = 1280*28*28
processor = AutoProcessor.from_pretrained(model_path, min_pixels=min_pixels, max_pixels=max_pixels)

messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "image",
                "image": "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg",
            },
            {"type": "text", "text": "Describe this image."},
        ],
    }
]

# Preparation for inference
text = processor.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True
)
image_inputs, video_inputs = process_vision_info(messages)
inputs = processor(
    text=[text],
    images=image_inputs,
    videos=video_inputs,
    padding=True,
    return_tensors="pt",
)
inputs = inputs.to('xpu')

with torch.inference_mode():
    for i in range(3):
        st = time.time()
        # Inference: Generation of the output
        generated_ids = model.generate(**inputs, max_new_tokens=32)
        et = time.time()
        print(et - st)
        generated_ids_trimmed = [
            out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
        ]
        output_text = processor.batch_decode(
            generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
        )
        print(output_text)

3. Summary of the change

4. How to test?

• N/A
• Unit test: Please manually trigger the PR Validation here by inputting the PR number (e.g., 1234). And paste your action link here once it has been successfully finished.
• Application test
• Document test
• ...

[rnwang04]

LGTM