Spaces:
Running
Running
File size: 2,496 Bytes
cab1df1 ff6b5fc cab1df1 ff6b5fc cab1df1 ff6b5fc cab1df1 460bccf 3a3e2e6 460bccf cab1df1 ff6b5fc 3a3e2e6 ff6b5fc 3a3e2e6 ff6b5fc 3a3e2e6 cab1df1 460bccf cab1df1 460bccf 3a3e2e6 460bccf cab1df1 460bccf cab1df1 460bccf |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 |
import torch
from PIL import Image
from transformers import AutoProcessor, AutoModelForVision2Seq
from transformers.image_utils import load_image
import numpy as np
import gradio as gr
# Set the device (GPU or CPU)
DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
# Initialize processor and model
try:
processor = AutoProcessor.from_pretrained("HuggingFaceTB/SmolVLM-Instruct")
model = AutoModelForVision2Seq.from_pretrained(
"HuggingFaceTB/SmolVLM-Instruct",
torch_dtype=torch.bfloat16,
_attn_implementation="flash_attention_2" if DEVICE == "cuda" else "eager",
).to(DEVICE)
except Exception as e:
print(f"Error loading model or processor: {str(e)}")
exit(1)
# Define the function to answer questions
def answer_question(image, question):
# Check if the image is provided
if image is None:
return "Error: Please upload an image."
# Convert NumPy array to PIL Image if necessary
try:
if isinstance(image, np.ndarray):
image = Image.fromarray(image)
except Exception as e:
return f"Error: Unable to process the image. {str(e)}"
# Ensure question is provided
if not question.strip():
return "Error: Please provide a question."
# Create input message for the model
messages = [
{
"role": "user",
"content": [
{"type": "image", "image": image},
{"type": "text", "text": question},
],
},
]
# Apply chat template (this assumes the processor has a chat-based input format)
try:
prompt = processor.apply_chat_template(messages, add_generation_prompt=True)
inputs = processor(text=prompt, images=[image], return_tensors="pt").to(DEVICE)
except Exception as e:
return f"Error: Failed to prepare inputs. {str(e)}"
# Generate the answer
try:
outputs = model.generate(**inputs)
answer = processor.decode(outputs[0], skip_special_tokens=True)
return answer
except Exception as e:
return f"Error: Failed to generate answer. {str(e)}"
# Create Gradio interface
iface = gr.Interface(
fn=answer_question,
inputs=[
gr.Image(type="numpy"),
gr.Textbox(lines=2, placeholder="Enter your question here..."),
],
outputs="text",
title="Image Question Answering",
description="Upload an image and ask a question about it.",
)
if __name__ == "__main__":
iface.launch()
|