Qwen3-VL-8B-Instruct-SFT-pos

import torch
from transformers import Qwen3VLForConditionalGeneration, AutoProcessor
from qwen_vl_utils import process_vision_info

# Example
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "image": "https://api.memegen.link/images/ds/high_quality/small_file.jpg"},
            {"type": "text", "text": "Describe this image."},
        ],
    }
]

# Process input
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("cuda")

# Generate response
generated_ids = model.generate(**inputs, max_new_tokens=512)
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[0])
#The image is a two-panel meme illustrating a common dilemma. **Panel 1:** A hand is hovering over a control panel with two buttons. The button on the left is labeled "high quality" and the button on the right is labeled "small file". The hand is positioned near the "small file" button, suggesting an impending choice. **Panel 2:** A man is shown with a distressed expression. He has a white cloth pressed to his forehead, and sweat is dripping down his face. He looks worried and uncomfortable. The meme humorously depicts the situation where one must choose between high-quality output and a smaller file size, often resulting in a stressful or frustrating experience. The man's distressed expression in the second panel represents the negative consequences of making a suboptimal choice.