app.py

import gradio as gr
import torch
from transformers import AutoTokenizer, AutoModelForVision2Seq, AutoImageProcessor
from PIL import Image
import spaces
import os

from huggingface_hub import login
login(os.environ["HF_KEY"])

device = "cuda" if torch.cuda.is_available() else "cpu"
model = AutoModelForVision2Seq.from_pretrained("stabilityai/japanese-stable-vlm", trust_remote_code=True, device_map='auto')
processor = AutoImageProcessor.from_pretrained("stabilityai/japanese-stable-vlm", device_map='auto')
tokenizer = AutoTokenizer.from_pretrained("stabilityai/japanese-stable-vlm", device_map='auto')

# Define the helper function to build prompts
TASK2INSTRUCTION = {
    "caption": "画像を詳細に述べてください。",
    "tag": "与えられた単語を使って、画像を詳細に述べてください。",
    "vqa": "与えられた画像を下に、質問に答えてください。",
}

def build_prompt(task="caption", input=None, sep="\n\n### "):
    assert task in TASK2INSTRUCTION, f"Please choose from {list(TASK2INSTRUCTION.keys())}"
    if task in ["tag", "vqa"]:
        assert input is not None, "Please fill in `input`!"
        if task == "tag" and isinstance(input, list):
            input = "、".join(input)
    else:
        assert input is None, f"`{task}` mode doesn't support to input questions"
    sys_msg = "以下は、タスクを説明する指示と、文脈のある入力の組み合わせです。要求を適切に満たす応答を書きなさい。"
    p = sys_msg
    roles = ["指示", "応答"]
    instruction = TASK2INSTRUCTION[task]
    msgs = [": \n" + instruction, ": \n"]
    if input:
        roles.insert(1, "入力")
        msgs.insert(1, ": \n" + input)
    for role, msg in zip(roles, msgs):
        p += sep + role + msg
    return p

# Define the function to generate text from the image and prompt
@spaces.GPU(duration=120)
def generate_text(image, task, input_text=None):
    prompt = build_prompt(task=task, input=input_text)
    inputs = processor(images=image, return_tensors="pt")
    text_encoding = tokenizer(prompt, add_special_tokens=False, return_tensors="pt")
    inputs.update(text_encoding)
    outputs = model.generate(
        **inputs.to(device=device, dtype=model.dtype),
        do_sample=False,
        num_beams=5,
        max_new_tokens=128,
        min_length=1,
        repetition_penalty=1.5,
    )
    generated_text = tokenizer.batch_decode(outputs, skip_special_tokens=True)[0].strip()
    return generated_text

# Define the Gradio interface
image_input = gr.Image(label="Upload an image")
task_input = gr.Radio(choices=["caption", "tag", "vqa"], value="caption", label="Select a task")
text_input = gr.Textbox(label="Enter text (for tag or vqa tasks)")

output = gr.Textbox(label="Generated text")

interface = gr.Interface(
    fn=generate_text,
    inputs=[image_input, task_input, text_input],
    outputs=output,
    examples=[
        ["examples/example_1.jpeg", "caption", None],
        ["examples/example_2.jpg", "tag", "寿司、箸"],
        ["examples/example_3.jpg", "vqa", "この画像を説明する"],
    ],
)

interface.launch()