| import tempfile |
| import gradio as gr |
| from main import load_artifacts_and_helpers, device |
|
|
| |
| final_pipeline, predict_fn, vocab, idx_to_answer, model, encode_fn = \ |
| load_artifacts_and_helpers(prefix="vqa_custom", map_location=device) |
|
|
| def vqa_interface(image, question): |
| |
| with tempfile.NamedTemporaryFile(suffix=".jpg") as f: |
| image.save(f.name) |
| answer = final_pipeline( |
| f.name, |
| question, |
| open_vqa_fn=None, |
| translate_fn=None |
| ) |
| return answer |
|
|
| demo = gr.Interface( |
| fn=vqa_interface, |
| inputs=[ |
| gr.Image(type="pil", label="Input Image"), |
| gr.Textbox(lines=1, label="Question", value="What is in the image?") |
| ], |
| outputs=gr.Textbox(label="Answer"), |
| title="VQA-RAD Demo", |
| description="Custom ResNet18 + LSTM VQA model (top-50 answers)." |
| ) |
|
|
| if __name__ == "__main__": |
| demo.launch() |
