attemp to adapt to newer version of gradio

.gitignore

@@ -9,3 +9,5 @@ pretrained/
 pretrained
 gradio_cached_examples/
 gradio_cached_examples
+.gradio/
+.gradio

README.md

@@ -4,11 +4,11 @@ emoji:
 colorFrom: blue
 colorTo: purple
 sdk: gradio
-sdk_version: 3.17.1
+sdk_version: 5.15.0
 app_file: app.py
 pinned: false
 license: mit
-python_version: 3.8.5
+python_version: 3.10.16
 ---
 
 Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -593,7 +593,7 @@ def t2i_interface(with_example=False):
             seed = gr.Number(20, label="Seed", precision=0)
             button = gr.Button("Run")
         with gr.Column():
-            img_output = gr.Gallery(label="Image Result", elem_id='customized_imbox').style(grid=n_sample_image)
+            img_output = gr.Gallery(label="Image Result", elem_id='customized_imbox')
 
     button.click(
         vd_inference.inference_t2i,
@@ -625,7 +625,7 @@ def i2i_interface(with_example=False):
             seed = gr.Number(20, label="Seed", precision=0)
             button = gr.Button("Run")
         with gr.Column():
-            img_output = gr.Gallery(label="Image Result", elem_id='customized_imbox').style(grid=n_sample_image)
+            img_output = gr.Gallery(label="Image Result", elem_id='customized_imbox')
 
         sim_flag.change(
             fn=lambda x: {
@@ -757,183 +757,183 @@ class example_visible_only_hack(object):
             if bi.value != vi:
                 fi.which_update = 'visible_oneoff'
 
-def dcg_interface(with_example=False):
-    gr.HTML('<p id=myinst>&nbsp Description: ' + get_instruction("Dual-Context") + '</p>')
-    with gr.Row():
-        input_session = []
-        with gr.Column():
-            img = gr.Image(label='Image Input', type='pil', elem_id='customized_imbox')
-            fcs = gr.Slider(label="Focus (Semantic -- Style)", minimum=0, maximum=1, value=0.5, step=0.02)
-            gr.HTML('<p id=myinst>&nbsp Focus: Focus on what aspect of the image? (0-semantic, 0.5-balanced (default), 1-style).</p>')
-
-            text = gr.Textbox(lines=2, placeholder="Input prompt...", label='Text Input')
-            tstrength = gr.Slider(label="Text Domination (NoEffect -- TextOnly)", minimum=0, maximum=1, value=0, step=0.02)
-
-            seed = gr.Number(20, label="Seed", precision=0)
-            button = gr.Button("Run")
-
-        with gr.Column():
-            output_gallary = gr.Gallery(label="Image Result", elem_id='customized_imbox').style(grid=n_sample_image)
-
-    input_list = []
-    for i in input_session:
-        input_list += i
-    button.click(
-        vd_inference.inference_dcg, 
-        inputs=[img, fcs, text, tstrength, seed],
-        outputs=[output_gallary])
-
-    if with_example:
-        gr.Examples(
-            label='Examples',
-            examples=get_example('Dual-Context'),
-            fn=vd_inference.inference_dcg,
-            inputs=[img, fcs, text, tstrength, seed],
-            outputs=[output_gallary],
-            cache_examples=cache_examples)
-
-def tcg_interface(with_example=False):
-    gr.HTML('<p id=myinst>&nbsp Description: ' + get_instruction("Triple-Context") + '</p>')
-    with gr.Row():
-        input_session = []
-        with gr.Column(min_width=940):
-            with gr.Row():
-                with gr.Column():
-                    img0  = gr.Image(label='Image Input', type='pil', elem_id='customized_imbox')
-                    img0.as_example = types.MethodType(customized_as_example, img0)
-                    imgm0 = gr.Image(label='Image Input with Mask', type='pil', elem_id='customized_imbox', tool='sketch', source="upload", visible=False)
-                    imgm0.postprocess = types.MethodType(customized_postprocess, imgm0)
-                    imgm0.as_example = types.MethodType(customized_as_example, imgm0)
-                    istrength0 = gr.Slider(label="Weight", minimum=0, maximum=1, value=1, step=0.02)
-                    fcs0 = gr.Slider(label="Focus (Semantic -- Style)", minimum=0, maximum=1, value=0.5, step=0.02)
-                    msk0 = gr.Checkbox(label='Use mask?')
-                    swapf0 = image_mimage_swap(img0, imgm0)
-
-                    msk0.change(
-                        fn=swapf0,
-                        inputs=[img0, imgm0, msk0],
-                        outputs=[img0, imgm0],)
-                    input_session.append([img0, imgm0, istrength0, fcs0, msk0])
-
-                with gr.Column():
-                    img1  = gr.Image(label='Image Input', type='pil', elem_id='customized_imbox')
-                    img1.as_example = types.MethodType(customized_as_example, img1)
-                    imgm1 = gr.Image(label='Image Input with Mask', type='pil', elem_id='customized_imbox', tool='sketch', source="upload", visible=False)
-                    imgm1.postprocess = types.MethodType(customized_postprocess, imgm1)
-                    imgm1.as_example = types.MethodType(customized_as_example, imgm1)
-                    istrength1 = gr.Slider(label="Weight", minimum=0, maximum=1, value=1, step=0.02)
-                    fcs1 = gr.Slider(label="Focus (Semantic -- Style)", minimum=0, maximum=1, value=0.5, step=0.02)
-                    msk1 = gr.Checkbox(label='Use mask?')
-                    swapf1 = image_mimage_swap(img1, imgm1)
-
-                    msk1.change(
-                        fn=swapf1,
-                        inputs=[img1, imgm1, msk1],
-                        outputs=[img1, imgm1],)
-                    input_session.append([img1, imgm1, istrength1, fcs1, msk1])
-
-            gr.HTML('<p id=myinst>&nbsp Weight: The strength of the reference image. This weight is subject to <u>Text Domination</u>).</p>'+
-                    '<p id=myinst>&nbsp Focus: Focus on what aspect of the image? (0-semantic, 0.5-balanced (default), 1-style).</p>'+
-                    '<p id=myinst>&nbsp Mask: Remove regions on reference image so they will not influence the output.</p>',)
-
-            text = gr.Textbox(lines=2, placeholder="Input prompt...", label='Text Input')
-            tstrength = gr.Slider(label="Text Domination (NoEffect -- TextOnly)", minimum=0, maximum=1, value=0, step=0.02)
-
-            seed = gr.Number(20, label="Seed", precision=0)
-            button = gr.Button("Run")
-
-        with gr.Column(min_width=470):
-            input_gallary = gr.Gallery(label="Input Display", elem_id="customized_imbox").style(grid=2)
-            output_gallary = gr.Gallery(label="Image Result", elem_id="customized_imbox").style(grid=n_sample_image)
-
-    input_list = []
-    for i in input_session:
-        input_list += i
-    input_list += [text, tstrength, seed]
-    button.click(
-        vd_inference.inference_tcg, 
-        inputs=input_list,
-        outputs=[input_gallary, output_gallary])
-
-    if with_example:
-        create_myexamples(
-            label='Examples',
-            examples=get_example('Triple-Context'),
-            fn=vd_inference.inference_tcg,
-            inputs=input_list,
-            outputs=[input_gallary, output_gallary, ],
-            cache_examples=cache_examples, )
-
-    gr.HTML('<br><p id=myinst>&nbsp How to add mask: Please see the following instructions.</p><br>'+
-            '<div id="maskinst">'+
-                '<img src="file/assets/demo/misc/mask_inst1.gif">'+
-                '<img src="file/assets/demo/misc/mask_inst2.gif">'+
-                '<img src="file/assets/demo/misc/mask_inst3.gif">'+
-            '</div>')
-
-def mcg_interface(with_example=False):
-    num_img_input = 4
-    gr.HTML('<p id=myinst>&nbsp Description: ' + get_instruction("Multi-Context") + '</p>')
-    with gr.Row():
-        input_session = []
-        with gr.Column():
-            for idx in range(num_img_input):
-                with gr.Tab('Image{}'.format(idx+1)):
-                    img = gr.Image(label='Image Input', type='pil', elem_id='customized_imbox')
-                    img.as_example = types.MethodType(customized_as_example, img)
-                    imgm = gr.Image(label='Image Input with Mask', type='pil', elem_id='customized_imbox', tool='sketch', source="upload", visible=False)
-                    imgm.postprocess = types.MethodType(customized_postprocess, imgm)
-                    imgm.as_example = types.MethodType(customized_as_example, imgm)
-
-                    with gr.Row():
-                        istrength = gr.Slider(label="Weight", minimum=0, maximum=1, value=1, step=0.02)
-                        fcs = gr.Slider(label="Focus (Semantic -- Style)", minimum=0, maximum=1, value=0.5, step=0.02)
-                    msk = gr.Checkbox(label='Use mask?')
-                    gr.HTML('<p id=myinst>&nbsp Weight: The strength of the reference image. This weight is subject to <u>Text Domination</u>).</p>'+
-                            '<p id=myinst>&nbsp Focus: Focus on what aspect of the image? (0-semantic, 0.5-balanced (default), 1-style).</p>'+
-                            '<p id=myinst>&nbsp Mask: Remove regions on reference image so they will not influence the output.</p>',)
-
-                    msk.change(
-                        fn=image_mimage_swap(img, imgm),
-                        inputs=[img, imgm, msk],
-                        outputs=[img, imgm],)
-                    input_session.append([img, imgm, istrength, fcs, msk])
-
-            text = gr.Textbox(lines=2, placeholder="Input prompt...", label='Text Input')
-            tstrength = gr.Slider(label="Text Domination (NoEffect -- TextOnly)", minimum=0, maximum=1, value=0, step=0.02)
-
-            seed = gr.Number(20, label="Seed", precision=0)
-            button = gr.Button("Run")
-
-
-        with gr.Column():
-            input_gallary = gr.Gallery(label="Input Display", elem_id='customized_imbox').style(grid=4)
-            output_gallary = gr.Gallery(label="Image Result", elem_id='customized_imbox').style(grid=n_sample_image)
-
-    input_list = []
-    for i in input_session:
-        input_list += i
-    input_list += [text, tstrength, seed]
-    button.click(
-        vd_inference.inference_mcg, 
-        inputs=input_list,
-        outputs=[input_gallary, output_gallary], )
-
-    if with_example:
-        create_myexamples(
-            label='Examples',
-            examples=get_example('Multi-Context'),
-            fn=vd_inference.inference_mcg,
-            inputs=input_list,
-            outputs=[input_gallary, output_gallary],
-            cache_examples=cache_examples, )
-
-    gr.HTML('<br><p id=myinst>&nbsp How to add mask: Please see the following instructions.</p><br>'+
-            '<div id="maskinst">'+
-                '<img src="file/assets/demo/misc/mask_inst1.gif">'+
-                '<img src="file/assets/demo/misc/mask_inst2.gif">'+
-                '<img src="file/assets/demo/misc/mask_inst3.gif">'+
-            '</div>')
+# def dcg_interface(with_example=False):
+#     gr.HTML('<p id=myinst>&nbsp Description: ' + get_instruction("Dual-Context") + '</p>')
+#     with gr.Row():
+#         input_session = []
+#         with gr.Column():
+#             img = gr.Image(label='Image Input', type='pil', elem_id='customized_imbox')
+#             fcs = gr.Slider(label="Focus (Semantic -- Style)", minimum=0, maximum=1, value=0.5, step=0.02)
+#             gr.HTML('<p id=myinst>&nbsp Focus: Focus on what aspect of the image? (0-semantic, 0.5-balanced (default), 1-style).</p>')
+
+#             text = gr.Textbox(lines=2, placeholder="Input prompt...", label='Text Input')
+#             tstrength = gr.Slider(label="Text Domination (NoEffect -- TextOnly)", minimum=0, maximum=1, value=0, step=0.02)
+
+#             seed = gr.Number(20, label="Seed", precision=0)
+#             button = gr.Button("Run")
+
+#         with gr.Column():
+#             output_gallary = gr.Gallery(label="Image Result", elem_id='customized_imbox')
+
+#     input_list = []
+#     for i in input_session:
+#         input_list += i
+#     button.click(
+#         vd_inference.inference_dcg, 
+#         inputs=[img, fcs, text, tstrength, seed],
+#         outputs=[output_gallary])
+
+#     if with_example:
+#         gr.Examples(
+#             label='Examples',
+#             examples=get_example('Dual-Context'),
+#             fn=vd_inference.inference_dcg,
+#             inputs=[img, fcs, text, tstrength, seed],
+#             outputs=[output_gallary],
+#             cache_examples=cache_examples)
+
+# def tcg_interface(with_example=False):
+#     gr.HTML('<p id=myinst>&nbsp Description: ' + get_instruction("Triple-Context") + '</p>')
+#     with gr.Row():
+#         input_session = []
+#         with gr.Column(min_width=940):
+#             with gr.Row():
+#                 with gr.Column():
+#                     img0  = gr.Image(label='Image Input', type='pil', elem_id='customized_imbox')
+#                     img0.as_example = types.MethodType(customized_as_example, img0)
+#                     imgm0 = gr.Image(label='Image Input with Mask', type='pil', elem_id='customized_imbox', tool='sketch', source="upload", visible=False)
+#                     imgm0.postprocess = types.MethodType(customized_postprocess, imgm0)
+#                     imgm0.as_example = types.MethodType(customized_as_example, imgm0)
+#                     istrength0 = gr.Slider(label="Weight", minimum=0, maximum=1, value=1, step=0.02)
+#                     fcs0 = gr.Slider(label="Focus (Semantic -- Style)", minimum=0, maximum=1, value=0.5, step=0.02)
+#                     msk0 = gr.Checkbox(label='Use mask?')
+#                     swapf0 = image_mimage_swap(img0, imgm0)
+
+#                     msk0.change(
+#                         fn=swapf0,
+#                         inputs=[img0, imgm0, msk0],
+#                         outputs=[img0, imgm0],)
+#                     input_session.append([img0, imgm0, istrength0, fcs0, msk0])
+
+#                 with gr.Column():
+#                     img1  = gr.Image(label='Image Input', type='pil', elem_id='customized_imbox')
+#                     img1.as_example = types.MethodType(customized_as_example, img1)
+#                     imgm1 = gr.Image(label='Image Input with Mask', type='pil', elem_id='customized_imbox', tool='sketch', source="upload", visible=False)
+#                     imgm1.postprocess = types.MethodType(customized_postprocess, imgm1)
+#                     imgm1.as_example = types.MethodType(customized_as_example, imgm1)
+#                     istrength1 = gr.Slider(label="Weight", minimum=0, maximum=1, value=1, step=0.02)
+#                     fcs1 = gr.Slider(label="Focus (Semantic -- Style)", minimum=0, maximum=1, value=0.5, step=0.02)
+#                     msk1 = gr.Checkbox(label='Use mask?')
+#                     swapf1 = image_mimage_swap(img1, imgm1)
+
+#                     msk1.change(
+#                         fn=swapf1,
+#                         inputs=[img1, imgm1, msk1],
+#                         outputs=[img1, imgm1],)
+#                     input_session.append([img1, imgm1, istrength1, fcs1, msk1])
+
+#             gr.HTML('<p id=myinst>&nbsp Weight: The strength of the reference image. This weight is subject to <u>Text Domination</u>).</p>'+
+#                     '<p id=myinst>&nbsp Focus: Focus on what aspect of the image? (0-semantic, 0.5-balanced (default), 1-style).</p>'+
+#                     '<p id=myinst>&nbsp Mask: Remove regions on reference image so they will not influence the output.</p>',)
+
+#             text = gr.Textbox(lines=2, placeholder="Input prompt...", label='Text Input')
+#             tstrength = gr.Slider(label="Text Domination (NoEffect -- TextOnly)", minimum=0, maximum=1, value=0, step=0.02)
+
+#             seed = gr.Number(20, label="Seed", precision=0)
+#             button = gr.Button("Run")
+
+#         with gr.Column(min_width=470):
+#             input_gallary = gr.Gallery(label="Input Display", elem_id="customized_imbox")
+#             output_gallary = gr.Gallery(label="Image Result", elem_id="customized_imbox")
+
+#     input_list = []
+#     for i in input_session:
+#         input_list += i
+#     input_list += [text, tstrength, seed]
+#     button.click(
+#         vd_inference.inference_tcg, 
+#         inputs=input_list,
+#         outputs=[input_gallary, output_gallary])
+
+#     if with_example:
+#         create_myexamples(
+#             label='Examples',
+#             examples=get_example('Triple-Context'),
+#             fn=vd_inference.inference_tcg,
+#             inputs=input_list,
+#             outputs=[input_gallary, output_gallary, ],
+#             cache_examples=cache_examples, )
+
+#     gr.HTML('<br><p id=myinst>&nbsp How to add mask: Please see the following instructions.</p><br>'+
+#             '<div id="maskinst">'+
+#                 '<img src="file/assets/demo/misc/mask_inst1.gif">'+
+#                 '<img src="file/assets/demo/misc/mask_inst2.gif">'+
+#                 '<img src="file/assets/demo/misc/mask_inst3.gif">'+
+#             '</div>')
+
+# def mcg_interface(with_example=False):
+#     num_img_input = 4
+#     gr.HTML('<p id=myinst>&nbsp Description: ' + get_instruction("Multi-Context") + '</p>')
+#     with gr.Row():
+#         input_session = []
+#         with gr.Column():
+#             for idx in range(num_img_input):
+#                 with gr.Tab('Image{}'.format(idx+1)):
+#                     img = gr.Image(label='Image Input', type='pil', elem_id='customized_imbox')
+#                     img.as_example = types.MethodType(customized_as_example, img)
+#                     imgm = gr.Image(label='Image Input with Mask', type='pil', elem_id='customized_imbox', tool='sketch', source="upload", visible=False)
+#                     imgm.postprocess = types.MethodType(customized_postprocess, imgm)
+#                     imgm.as_example = types.MethodType(customized_as_example, imgm)
+
+#                     with gr.Row():
+#                         istrength = gr.Slider(label="Weight", minimum=0, maximum=1, value=1, step=0.02)
+#                         fcs = gr.Slider(label="Focus (Semantic -- Style)", minimum=0, maximum=1, value=0.5, step=0.02)
+#                     msk = gr.Checkbox(label='Use mask?')
+#                     gr.HTML('<p id=myinst>&nbsp Weight: The strength of the reference image. This weight is subject to <u>Text Domination</u>).</p>'+
+#                             '<p id=myinst>&nbsp Focus: Focus on what aspect of the image? (0-semantic, 0.5-balanced (default), 1-style).</p>'+
+#                             '<p id=myinst>&nbsp Mask: Remove regions on reference image so they will not influence the output.</p>',)
+
+#                     msk.change(
+#                         fn=image_mimage_swap(img, imgm),
+#                         inputs=[img, imgm, msk],
+#                         outputs=[img, imgm],)
+#                     input_session.append([img, imgm, istrength, fcs, msk])
+
+#             text = gr.Textbox(lines=2, placeholder="Input prompt...", label='Text Input')
+#             tstrength = gr.Slider(label="Text Domination (NoEffect -- TextOnly)", minimum=0, maximum=1, value=0, step=0.02)
+
+#             seed = gr.Number(20, label="Seed", precision=0)
+#             button = gr.Button("Run")
+
+
+#         with gr.Column():
+#             input_gallary = gr.Gallery(label="Input Display", elem_id='customized_imbox')
+#             output_gallary = gr.Gallery(label="Image Result", elem_id='customized_imbox')
+
+#     input_list = []
+#     for i in input_session:
+#         input_list += i
+#     input_list += [text, tstrength, seed]
+#     button.click(
+#         vd_inference.inference_mcg, 
+#         inputs=input_list,
+#         outputs=[input_gallary, output_gallary], )
+
+#     if with_example:
+#         create_myexamples(
+#             label='Examples',
+#             examples=get_example('Multi-Context'),
+#             fn=vd_inference.inference_mcg,
+#             inputs=input_list,
+#             outputs=[input_gallary, output_gallary],
+#             cache_examples=cache_examples, )
+
+#     gr.HTML('<br><p id=myinst>&nbsp How to add mask: Please see the following instructions.</p><br>'+
+#             '<div id="maskinst">'+
+#                 '<img src="file/assets/demo/misc/mask_inst1.gif">'+
+#                 '<img src="file/assets/demo/misc/mask_inst2.gif">'+
+#                 '<img src="file/assets/demo/misc/mask_inst3.gif">'+
+#             '</div>')
 
 ###########
 # Example #
@@ -1080,12 +1080,12 @@ if True:
             i2t_interface(with_example=True)
         with gr.Tab('Text-Variation'):
             t2t_interface(with_example=True)
-        with gr.Tab('Dual-Context Image-Generation'):
-            dcg_interface(with_example=True)
-        with gr.Tab('Triple-Context Image-Blender'):
-            tcg_interface(with_example=True)
-        with gr.Tab('Multi-Context Image-Blender'):
-            mcg_interface(with_example=True)
+        # with gr.Tab('Dual-Context Image-Generation'):
+        #     dcg_interface(with_example=True)
+        # with gr.Tab('Triple-Context Image-Blender'):
+        #     tcg_interface(with_example=True)
+        # with gr.Tab('Multi-Context Image-Blender'):
+        #     mcg_interface(with_example=True)
 
         gr.HTML(
             """

cusomized_gradio_blocks.py

@@ -11,8 +11,6 @@ import warnings
 from pathlib import Path
 from typing import TYPE_CHECKING, Any, Callable, Dict, Iterable, List, Tuple
 
-import matplotlib
-import matplotlib.pyplot as plt
 import numpy as np
 import PIL
 import PIL.Image
@@ -20,7 +18,7 @@ import PIL.Image
 import gradio
 from gradio import components, processing_utils, routes, utils
 from gradio.context import Context
-from gradio.documentation import document, set_documentation_group
+# from gradio.documentation import document, set_documentation_group
 from gradio.flagging import CSVLogger
 
 if TYPE_CHECKING:  # Only import for type checking (to avoid circular imports).

lib/cfg_helper.py

@@ -6,7 +6,6 @@ import time
 import pprint
 import numpy as np
 import torch
-import matplotlib
 import argparse
 import json
 import yaml

lib/model_zoo/common/utils.py

@@ -6,8 +6,6 @@ import copy
 import functools
 import itertools
 
-import matplotlib.pyplot as plt
-
 ########
 # unit #
 ########

requirements.txt

@@ -1,16 +1,16 @@
---extra-index-url https://download.pytorch.org/whl/cu113
+--extra-index-url https://download.pytorch.org/whl/cu117
+torch==2.0.0+cu117
+torchvision==0.15
 
-torch==1.12.1+cu113
-torchvision==0.13.1+cu113
+numpy==1.26.4
+gradio==5.15.0
 
-pyyaml
-easydict
-
-tqdm==4.60.0
-transformers==4.24.0
-torchmetrics==0.7.3
+transformers==4.48.2
+torchmetrics==1.6.1
 
 einops==0.3.0
 omegaconf==2.1.1
-huggingface-hub==0.11.1
-gradio==3.17.1
+
+pyyaml
+easydict
+tqdm