Using loaded interface as a function in Blocks return error

[fffiloni]

Describe the bug

I want to use Stable Diffusion as a loaded interface in a Block, and use it as a function run by a button.
Calling that function leads to a AttributeError: 'Dataset' object has no attribute 'serialize'error in logs

Is there an existing issue for this?

• I have searched the existing issues

Reproduction

import gradio as gr

stable_diffusion = gr.Blocks.load(name="spaces/stabilityai/stable-diffusion")

def get_images(prompt):
    images = stable_diffusion(prompt, fn_index=0)
    return images

with gr.Blocks() as demo:
    
    with gr.Column():
        with gr.Row():
            prompt = gr.Text()
            gallery = gr.Gallery()
        submit_btn = gr.Button()
        submit_btn.click(get_images, inputs=[prompt], outputs=[gallery])
demo.launch()

Screenshot

No response

Logs

Fetching interface from: https://huggingface.co/spaces/stabilityai/stable-diffusion
Running on local URL:  http://0.0.0.0:7860

To create a public link, set `share=True` in `launch()`.
Traceback (most recent call last):
  File "/home/user/.local/lib/python3.8/site-packages/gradio/routes.py", line 273, in run_predict
    output = await app.blocks.process_api(
  File "/home/user/.local/lib/python3.8/site-packages/gradio/blocks.py", line 746, in process_api
    result = await self.call_function(fn_index, inputs, iterator)
  File "/home/user/.local/lib/python3.8/site-packages/gradio/blocks.py", line 657, in call_function
    prediction = await anyio.to_thread.run_sync(
  File "/home/user/.local/lib/python3.8/site-packages/anyio/to_thread.py", line 31, in run_sync
    return await get_asynclib().run_sync_in_worker_thread(
  File "/home/user/.local/lib/python3.8/site-packages/anyio/_backends/_asyncio.py", line 937, in run_sync_in_worker_thread
    return await future
  File "/home/user/.local/lib/python3.8/site-packages/anyio/_backends/_asyncio.py", line 867, in run
    result = context.run(func, *args)
  File "app.py", line 6, in get_images
    images = stable_diffusion(prompt, fn_index=0)
  File "/home/user/.local/lib/python3.8/site-packages/gradio/blocks.py", line 561, in __call__
    serialized_input = block.serialize(params[i])
AttributeError: 'Dataset' object has no attribute 'serialize'

### System Info

```shell
Gradio version: 3.4
Running on Huggingface

Severity

annoying

[freddyaboulton]

Thanks for filing @fffiloni !

My guess is that fn_index=0 probably corresponds to the examples component and not the actual text to image generation. I would check with fn_index=2.

[fffiloni]

Hello @freddyaboulton !

I tried with fn_index=2 and fn_index=1 but get another error:

Fetching interface from: https://huggingface.co/spaces/stabilityai/stable-diffusion
Running on local URL:  http://0.0.0.0:7860

To create a public link, set `share=True` in `launch()`.
Traceback (most recent call last):
  File "/home/user/.local/lib/python3.8/site-packages/gradio/routes.py", line 275, in run_predict
    output = await app.blocks.process_api(
  File "/home/user/.local/lib/python3.8/site-packages/gradio/blocks.py", line 764, in process_api
    predictions = self.postprocess_data(fn_index, result["prediction"], state)
  File "/home/user/.local/lib/python3.8/site-packages/gradio/blocks.py", line 731, in postprocess_data
    output_value = block.postprocess(prediction_value)
  File "/home/user/.local/lib/python3.8/site-packages/gradio/components.py", line 3297, in postprocess
    img = processing_utils.encode_url_or_file_to_base64(img)
  File "/home/user/.local/lib/python3.8/site-packages/gradio/processing_utils.py", line 37, in encode_url_or_file_to_base64
    return encode_file_to_base64(path, encryption_key=encryption_key)
  File "/home/user/.local/lib/python3.8/site-packages/gradio/processing_utils.py", line 59, in encode_file_to_base64
    with open(f, "rb") as file:
IsADirectoryError: [Errno 21] Is a directory: '/home/user/app/1422c498-85b2-487e-b5f5-cd5fe3974606'

[freddyaboulton]

Thanks for the info @fffiloni ! Will look into this

[freddyaboulton]

Hi @fffiloni ! I think I understand what's happening - the stable diffusion space is returning a path to a gallery. The images in the gallery are stored as files in that path. If you modify your function to return a list of the image files instead it works

import gradio as gr
import os

stable_diffusion = gr.Blocks.load(name="spaces/stabilityai/stable-diffusion")

def get_images(prompt):
    gallery_dir = stable_diffusion(prompt, fn_index=2)
    return [os.path.join(gallery_dir, img) for img in os.listdir(gallery_dir)]

with gr.Blocks() as demo:
    
    with gr.Column():
        with gr.Row():
            prompt = gr.Text()
            gallery = gr.Gallery()
        submit_btn = gr.Button()
        submit_btn.click(get_images, inputs=[prompt], outputs=[gallery])
demo.launch()

[fffiloni]

Ouh nice @freddyaboulton ! Yes it works !
Thank you very much for your time investigating ;)