File size: 8,719 Bytes
3166b97 07e7ef5 3166b97 de160c1 0978583 de160c1 0178eef de160c1 3166b97 e13c42d 6efb87d 595acb5 6efb87d 3166b97 6efb87d daaa9eb d82d4ca 6efb87d 4466abf 6efb87d 4466abf 6efb87d d82d4ca 6efb87d daaa9eb 6efb87d daaa9eb 3166b97 6efb87d 3166b97 76c8ebf 3166b97 6efb87d 3166b97 76c8ebf 3166b97 6efb87d 3166b97 6efb87d 3166b97 de160c1 3166b97 83fab60 595acb5 83fab60 3166b97 e13c42d 6efb87d 4860fd0 b36e19c 4860fd0 3166b97 6efb87d 3166b97 76c8ebf 3166b97 e13c42d 3166b97 6efb87d 3166b97 72a6c2f 3166b97 6efb87d 3166b97 72a6c2f 3166b97 6efb87d 3166b97 72a6c2f 3166b97 6efb87d 3166b97 72a6c2f 3166b97 e13c42d 6efb87d 72a6c2f 3166b97 83fab60 3166b97 |
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 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 |
# AUTOGENERATED! DO NOT EDIT! File to edit: app.ipynb.
# %% auto 0
__all__ = ['block', 'make_clickable_model', 'make_clickable_user', 'get_submissions']
# %% app.ipynb 0
import subprocess
import sys
def upgrade(package):
subprocess.run([sys.executable, "-m", "pip", "install", "--upgrade", package])
#upgrade("gradio==3.116")
def install_specific_version(package, version):
subprocess.run([sys.executable, "-m", "pip", "install", package+version])
install_specific_version("gradio==", "3.16.0")
import gradio as gr
import pandas as pd
from huggingface_hub import list_models
from diffusers import StableDiffusionPipeline
# %% app.ipynb 1
def get_model_list(category):
submissions_list = list_models(filter=["dreambooth-hackathon", category], full=True)
spaces_pipeline_load = [submission.id for submission in submissions_list ]
return gr.Dropdown.update(choices=spaces_pipeline_load , value=spaces_pipeline_load[4])
def get_initial_prompt(model_nm):
#Example - a photo of shbrcky dog
print(******Inside get initial prompt******)
print(f"model name from Gradio input is - {model_nm}")
user_model_nm = model_nm.split('/')[-1]
if '-' in user_model_nm:
prompt = " ".join(user_model_nm.split('-'))
else:
prompt = user_model_nm
print(f"Extracted name for constructing Prompt is - {user_model_nm}")
print(f"Extracted name for constructing Prompt is - {prompt}")
prompt = "a photo of " + prompt + " "
return gr.Textbox.update(value=prompt)
def get_pipeline(model_name): #, progress=gr.Progress(track_tqdm=True)):
#Using diffusers pipeline to generate an image for the demo
#Loading Your Dreambooth model
pipeline = StableDiffusionPipeline.from_pretrained(model_name) # Example - ("ashiqabdulkhader/shiba-dog") or ('pharma/sugar-glider')
return pipeline
def make_demo(model_name, prompt, progress=gr.Progress(track_tqdm=True)):
#Using diffusers pipeline to generate an image for the demo
progress(0, desc="Starting...")
pipeline = get_pipeline(model_name) #StableDiffusionPipeline.from_pretrained(model_name) # Example - ("ashiqabdulkhader/shiba-dog") or ('pharma/sugar-glider')
#Generating Image from your prompt
image_demo = pipeline(prompt).images[0]
return image_demo
def make_clickable_model(model_name, link=None):
if link is None:
link = "https://huggingface.co/" + model_name
# Remove user from model name
return f'<a target="_blank" href="{link}">{model_name.split("/")[-1]}</a>'
def make_clickable_user(user_id):
link = "https://huggingface.co/" + user_id
return f'<a target="_blank" href="{link}">{user_id}</a>'
# %% app.ipynb 2
def get_submissions(category, prompt):
submissions = list_models(filter=["dreambooth-hackathon", category], full=True)
leaderboard_models = []
for submission in submissions:
# user, model, likes
user_id = submission.id.split("/")[0]
model_nm = submission.id.split("/")[-1]
if '-' in model_nm:
model_nm = " ".join(model_nm.split('-'))
#button_html = get_button()
leaderboard_models.append(
(
make_clickable_user(user_id),
make_clickable_model(submission.id, prompt),
submission.likes,
#button_html #'a photo of ' + model_nm + " "
)
)
df = pd.DataFrame(data=leaderboard_models, columns=["User", "Model", "Likes", ])
df.sort_values(by=["Likes"], ascending=False, inplace=True)
df.insert(0, "Rank", list(range(1, len(df) + 1)))
return df
# %% app.ipynb 3
block = gr.Blocks()
with block:
gr.Markdown(
"""# Gradio-powered leaderboard-evaluator for the DreamBooth Hackathon
Welcome to this Gradio-powered leaderboard! Select a theme and one of the dreambooth models trained by hackathon-participants, and key in your prompt as shown (eg., a photo of Shiba dog in a jungle). Note that, the image generation might take long (around 400 seconds) as it will have to load the respective model pipeline into memory.
<br>**If you like a model demo, click on the model name in the table below and UPVOTE the model on Huggingface hub**<br><br>
DreamBooth Hackathon - is an ongoing community event where participants **personalize a Stable Diffusion model** by fine-tuning it with a powerful technique called [_DreamBooth_](https://arxiv.org/abs/2208.12242). This technique allows one to implant a subject into the output domain of the model such that it can be synthesized with a _unique identifier_ (eg., shiba dog) in the prompt.
This competition comprises 5 _themes_ - Animals, Science, Food, Landscapes, and Wildcards. For details on how to participate, check out the hackathon's guide [here](https://github.com/huggingface/diffusion-models-class/blob/main/hackathon/README.md).
"""
)
with gr.Row():
with gr.Column():
theme = gr.Radio(label="Pick a Theme",choices=["animal","science", "food", "landscape", "wildcard"] )
model_list = gr.Dropdown(label="Pick a Dreamboooth model", choices = []) # choices=
with gr.Column():
prompt_in = gr.Textbox(label="Type in a Prompt in front of the given text..", value="")
button_in = gr.Button(Value = "Generate Image")
image_out = gr.Image(label="Generated image with your choice of Dreambooth model")
with gr.Tabs():
with gr.TabItem("Animal 🐨"):
with gr.Row():
animal_data = gr.components.Dataframe(
type="pandas", datatype=["number", "markdown", "markdown", "number","str"], interactive = True
)
with gr.Row():
data_run = gr.Button("Refresh")
data_run.click(
get_submissions, inputs=[gr.Variable("animal"), prompt_in], outputs=animal_data
)
with gr.TabItem("Science 🔬"):
with gr.Row():
science_data = gr.components.Dataframe(
type="pandas", datatype=["number", "markdown", "markdown", "number", "str"], interactive = True
)
with gr.Row():
data_run = gr.Button("Refresh")
data_run.click(
get_submissions, inputs=[gr.Variable("science"), prompt_in], outputs=science_data
)
with gr.TabItem("Food 🍔"):
with gr.Row():
food_data = gr.components.Dataframe(
type="pandas", datatype=["number", "markdown", "markdown", "number", "str"], interactive = True
)
with gr.Row():
data_run = gr.Button("Refresh")
data_run.click(
get_submissions, inputs=[gr.Variable("food"), prompt_in], outputs=food_data
)
with gr.TabItem("Landscape 🏔"):
with gr.Row():
landscape_data = gr.components.Dataframe(
type="pandas", datatype=["number", "markdown", "markdown", "number", "str"], interactive = True
)
with gr.Row():
data_run = gr.Button("Refresh")
data_run.click(
get_submissions,
inputs=[gr.Variable("landscape"),prompt_in],
outputs=landscape_data,
)
with gr.TabItem("Wilcard 🔥"):
with gr.Row():
wildcard_data = gr.components.Dataframe(
type="pandas", datatype=["number", "markdown", "markdown", "number", "str"], interactive = True
)
with gr.Row():
data_run = gr.Button("Refresh")
data_run.click(
get_submissions,
inputs=[gr.Variable("wildcard"),prompt_in],
outputs=wildcard_data,
)
theme.change(get_model_list, theme, model_list )
model_list.change(get_initial_prompt, model_list, prompt_in )
button_in.click(make_demo, [model_list, prompt_in], image_out)
block.load(get_submissions, inputs=[gr.Variable("animal"), prompt_in], outputs=animal_data)
block.load(get_submissions, inputs=[gr.Variable("science"), prompt_in], outputs=science_data)
block.load(get_submissions, inputs=[gr.Variable("food"), prompt_in], outputs=food_data)
block.load(get_submissions, inputs=[gr.Variable("landscape"), prompt_in], outputs=landscape_data)
block.load(get_submissions, inputs=[gr.Variable("wildcard"), prompt_in], outputs=wildcard_data)
block.queue(concurrency_count=3)
block.launch()
|