minicpm-v-4-tiny-random

import numpy as np
import torch
from PIL import Image
from transformers import AutoModel, AutoTokenizer

model_id = "yujiepan/minicpm-v-4-tiny-random"
model = AutoModel.from_pretrained(model_id, trust_remote_code=True,
                                  attn_implementation='sdpa', torch_dtype=torch.bfloat16)
model = model.eval().cuda()
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)

image = Image.fromarray(np.random.randint(0, 255, (224, 224, 3), dtype=np.uint8), 'RGB')
question = "What is the landform in the picture?"
msgs = [{'role': 'user', 'content': [image, question]}]
answer = model.chat(
    msgs=msgs,
    image=image,
    tokenizer=tokenizer,
    max_new_tokens=32,
)
print(answer)

# Second round chat, pass history context of multi-turn conversation
msgs.append({"role": "assistant", "content": [answer]})
msgs.append({"role": "user", "content": [
            "What should I pay attention to when traveling here?"]})
answer = model.chat(
    msgs=msgs,
    image=None,
    tokenizer=tokenizer,
    max_new_tokens=32,
)
print(answer)