app.py

import spaces
import gradio as gr
from transformers import AutoModelForCausalLM, AutoProcessor, GenerationConfig
from PIL import Image
import torch
import requests

# Load the processor and model
processor = AutoProcessor.from_pretrained(
    'allenai/Molmo-7B-D-0924',
    trust_remote_code=True,
    torch_dtype='auto',
    device_map='auto'
)

model = AutoModelForCausalLM.from_pretrained(
    'allenai/Molmo-7B-D-0924',
    trust_remote_code=True,
    torch_dtype='auto',
    device_map='auto'
)

@spaces.GPU
def describe_image(image):
    # Process the image
    inputs = processor.process(images=[image], text="Describe this image.")
    
    # Move inputs to the correct device and make a batch of size 1
    inputs = {k: v.to(model.device).unsqueeze(0) for k, v in inputs.items()}
    
    # Generate output with maximum 200 new tokens
    output = model.generate_from_batch(
        inputs,
        GenerationConfig(max_new_tokens=200, stop_strings="<|endoftext|>"),
        tokenizer=processor.tokenizer
    )
    
    # Decode and return generated text
    generated_tokens = output[0, inputs['input_ids'].size(1):]
    generated_text = processor.tokenizer.decode(generated_tokens, skip_special_tokens=True)
    
    return generated_text

# Gradio interface
gr.Interface(
    fn=describe_image,
    inputs=gr.inputs.Image(type="pil"),
    outputs="text",
    title="Visual Language Model - Molmo",
    description="Upload an image, and the model will generate a detailed description of it."
).launch()