shilinxu/NaflexVLM2_5

from PIL import Image
from transformers import AutoModelForCausalLM, AutoProcessor

import torch
model_path = 'shilinxu/NaflexVLM2_5'
model = AutoModelForCausalLM.from_pretrained(model_path, torch_dtype=torch.bfloat16 ,device_map='cuda:0', trust_remote_code=True)

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


url = 'https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg'
import requests
image = Image.open(requests.get(url, stream=True).raw)


messages = [
    {
        'role':'user', 
        'content': [
            {'type':'text', 'text': 'Describe this image in detail.'},
            {'type':'image'}
        ]
    }
]


text = processor.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True
)

inputs = processor(
    text=text,
    images=[image],
    padding=False,
    return_tensors="pt",
)
inputs = inputs.to(model.device, dtype=torch.bfloat16)

generated_ids = model.generate(**inputs, max_new_tokens=128, temperature=1.0, repetition_penalty=1.2)
generated_ids = [out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)]

output_text = processor.batch_decode(generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)
print(output_text)