File size: 13,856 Bytes
6b5dfe6
bfdbdf6
 
 
 
 
b3d3b2f
6b5dfe6
 
 
 
 
 
bfdbdf6
b3d3b2f
bfdbdf6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b3d3b2f
 
 
 
 
 
 
 
 
 
 
89f9702
22d5d2c
bfdbdf6
 
9255bd7
bfdbdf6
d1c3953
bfdbdf6
 
 
9465fd2
bfdbdf6
36aeb40
bfdbdf6
dbfd73e
bfdbdf6
 
 
dbfd73e
bfdbdf6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2e8ed6c
 
 
 
 
 
 
dbfd73e
2e8ed6c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
36aeb40
d1c3953
 
 
 
bfdbdf6
d1c3953
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e4068c9
d1c3953
6b5dfe6
 
 
 
 
 
 
bfdbdf6
 
6b5dfe6
 
 
bfdbdf6
6b5dfe6
bfdbdf6
 
 
 
 
6b5dfe6
 
bfdbdf6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6b5dfe6
bfdbdf6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6b5dfe6
bfdbdf6
 
 
 
b3d3b2f
bfdbdf6
 
6b5dfe6
bfdbdf6
 
9255bd7
 
 
 
 
 
 
 
 
 
 
 
6b5dfe6
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
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
import gradio as gr
import os
from pathlib import Path
import argparse
import shutil
from train_dreambooth import run_training
from PIL import Image

css = '''
    .instruction{position: absolute; top: 0;right: 0;margin-top: 0px !important}
    .arrow{position: absolute;top: 0;right: -8px;margin-top: -8px !important}
    #component-4, #component-3, #component-10{min-height: 0}
'''
shutil.unpack_archive("mix.zip", "mix")
model_to_load = "multimodalart/sd-fine-tunable"
maximum_concepts = 3
def swap_values_files(*total_files):
    file_counter = 0
    for files in total_files:
        if(files):
            for file in files:
                filename = Path(file.orig_name).stem
                pt=''.join([i for i in filename if not i.isdigit()])
                pt=pt.replace("_"," ")
                pt=pt.replace("(","")
                pt=pt.replace(")","")
                instance_prompt = pt
                print(instance_prompt)
                file_counter += 1
            training_steps = (file_counter*200)
    return training_steps

def swap_text(option):
    mandatory_liability = "You must have the right to do so and you are liable for the images you use"
    if(option == "object"):
        instance_prompt_example = "cttoy"
        freeze_for = 50
        return [f"You are going to train `object`(s), upload 5-10 images of each object you are planning on training on from different angles/perspectives. {mandatory_liability}:", '''<img src="file/cat-toy.png" />''', f"You should name your concept with a unique made up word that has low chance of the model already knowing it (e.g.: `{instance_prompt_example}` here)", freeze_for]
    elif(option == "person"):
       instance_prompt_example = "julcto"
       freeze_for = 100
       return [f"You are going to train a `person`(s), upload 10-20 images of each person you are planning on training on from different angles/perspectives. {mandatory_liability}:", '''<img src="file/cat-toy.png" />''', f"You should name the files with a unique word that represent your concept (like `{instance_prompt_example}` in this example). You can train multiple concepts as well.", freeze_for]
    elif(option == "style"):
        instance_prompt_example = "mspolstyll"
        freeze_for = 10
        return [f"You are going to train a `style`, upload 10-20 images of the style you are planning on training on. Name the files with the words you would like  {mandatory_liability}:", '''<img src="file/cat-toy.png" />''', f"You should name your files with a unique word that represent your concept (as `{instance_prompt_example}` for example). You can train multiple concepts as well.", freeze_for]

def train(*inputs):
    file_counter = 0
    for i, input in enumerate(inputs):
        if(i < maximum_concepts-1):
            if(input):
                os.makedirs('instance_images',exist_ok=True)
                files = inputs[i+(maximum_concepts*2)]
                prompt = inputs[i+maximum_concepts]
                for j, file_temp in enumerate(files):
                    file = Image.open(file_temp.name)
                    width, height = file.size
                    side_length = min(width, height)
                    left = (width - side_length)/2
                    top = (height - side_length)/2
                    right = (width + side_length)/2
                    bottom = (height + side_length)/2
                    image = file.crop((left, top, right, bottom))
                    image = image.resize((512, 512))
                    extension = file_temp.name.split(".")[1]
                    image.convert('RGB')
                    image.save(f'instance_images/{prompt}_({j+1}).jpg', format="JPEG", quality = 100)
                    file_counter += 1
    
    os.makedirs('output_model',exist_ok=True)
    uses_custom = inputs[-1] 
    type_of_thing = inputs[-4]
    if(uses_custom):
        Training_Steps = int(inputs[-3])
        Train_text_encoder_for = int(inputs[-2])
    else:
        Training_Steps = file_counter*200
    if(type_of_thing == "person"):
            class_data_dir = "mix"
            Train_text_encoder_for=100
            args_txt_encoder = argparse.Namespace(
                image_captions_filename = True,
                train_text_encoder = True,
                pretrained_model_name_or_path=model_to_load,
                instance_data_dir="instance_images",
                class_data_dir=class_data_dir,
                output_dir="output_model",
                with_prior_preservation=True,
                prior_loss_weight=1.0,
                instance_prompt="",
                seed=42,
                resolution=512,
                mixed_precision="fp16",
                train_batch_size=1,
                gradient_accumulation_steps=1,
                gradient_checkpointing=True,
                use_8bit_adam=True,
                learning_rate=2e-6,
                lr_scheduler="polynomial",
                lr_warmup_steps=0,
                max_train_steps=Training_Steps,
                num_class_images=200
            )
            args_unet = argparse.Namespace(
                image_captions_filename = True,
                train_only_unet=True,
                Session_dir="output_model",
                save_starting_step=0,
                save_n_steps=0,
                pretrained_model_name_or_path=model_to_load,
                instance_data_dir="instance_images",
                output_dir="output_model",
                instance_prompt="",
                seed=42,
                resolution=512,
                mixed_precision="fp16",
                train_batch_size=1,
                gradient_accumulation_steps=1,
                gradient_checkpointing=False,
                use_8bit_adam=True,
                learning_rate=2e-6,
                lr_scheduler="polynomial",
                lr_warmup_steps=0,
                max_train_steps=Training_Steps
            )
            run_training(args_txt_encoder)
            run_training(args_unet)
    elif(type_of_thing == "object" or type_of_thing == "style"):
            if(type_of_thing == "object"):
                Train_text_encoder_for=30
            elif(type_of_thing == "style"):
                Train_text_encoder_for=15
            class_data_dir = None
            stptxt = int((Training_Steps*Train_text_encoder_for)/100)
            args_general = argparse.Namespace(
                image_captions_filename = True,
                train_text_encoder = True,
                stop_text_encoder_training = stptxt,
                save_n_steps = 0,
                pretrained_model_name_or_path = model_to_load,
                instance_data_dir="instance_images",
                class_data_dir=class_data_dir,
                output_dir="output_model",
                instance_prompt="",
                seed=42,
                resolution=512,
                mixed_precision="fp16",
                train_batch_size=1,
                gradient_accumulation_steps=1,
                use_8bit_adam=True,
                learning_rate=2e-6,
                lr_scheduler="polynomial",
                lr_warmup_steps = 0,
                max_train_steps=Training_Steps,     
            )
            run_training(args_general)

    shutil.rmtree('instance_images')
    shutil.make_archive("output_model", 'zip', "output_model")
    shutil.rmtree("output_model")
    return [gr.update(visible=True, value="output_model.zip"), gr.update(visible=True), gr.update(visible=True)]

with gr.Blocks(css=css) as demo:
    with gr.Box():
        # You can remove this part here for your local clone
        gr.HTML('''
            <div class="gr-prose" style="max-width: 80%">
            <h2>Attention - This Space doesn't work in this shared UI</h2>
            <p>For it to work, you have to duplicate the Space and run it on your own profile where a (paid) private GPU will be attributed to it during runtime. It will cost you < US$1 to train a model on default settings! 🤑</p> 
            <img class="instruction" src="file/duplicate.png"> 
            <img class="arrow" src="file/arrow.png" />
            </div>
        ''')
    gr.Markdown("# Dreambooth training")
    gr.Markdown("Customize Stable Diffusion by giving it with few-shot examples")
    with gr.Row():
        type_of_thing = gr.Dropdown(label="What would you like to train?", choices=["object", "person", "style"], value="object", interactive=True)
        #with gr.Column():
            #with gr.Box():
            #    gr.Textbox(label="What prompt you would like to train it on", value="The photo of a cttoy", interactive=True).style(container=False, item_container=False)
            #    gr.Markdown("You should try using words the model doesn't know. Don't use names or well known concepts.")
    with gr.Row():
        with gr.Column():
            thing_description = gr.Markdown("You are going to train an `object`, upload 5-10 images of the object you are planning on training on from different angles/perspectives. You must have the right to do so and you are liable for the images you use")
            thing_image_example = gr.HTML('''<img src="file/cat-toy.png" />''')
            things_naming = gr.Markdown("For training, you should name the files with a unique word that represent your concept (like `cctoy` in this example). You can train multiple concepts by naming multiple images at once. Images will be automatically cropped to 512x512.")
        with gr.Column():
            file_collection = []
            concept_collection = []
            buttons_collection = []
            delete_collection = []
            is_visible = []

            row = [None] * maximum_concepts
            for x in range(maximum_concepts):
                ordinal = lambda n: "%d%s" % (n, "tsnrhtdd"[(n // 10 % 10 != 1) * (n % 10 < 4) * n % 10::4])
                if(x == 0):
                    visible = True
                    is_visible.append(gr.State(value=True))
                else:
                    visible = False
                    is_visible.append(gr.State(value=False))

                file_collection.append(gr.File(label=f"Upload the images for your {ordinal(x+1)} concept", file_count="multiple", interactive=True, visible=visible))
                with gr.Column(visible=visible) as row[x]:
                    concept_collection.append(gr.Textbox(label=f"{ordinal(x+1)} concept prompt - use a unique, made up word to avoid collisions"))  
                    with gr.Row():
                        if(x < maximum_concepts-1):
                            buttons_collection.append(gr.Button(value="Add +1 concept", visible=visible))
                        if(x > 0):
                            delete_collection.append(gr.Button(value=f"Delete {ordinal(x+1)} concept"))
            
            counter_add = 1
            for button in buttons_collection:
                if(counter_add < len(buttons_collection)):
                    button.click(lambda:
                    [gr.update(visible=True),gr.update(visible=True), gr.update(visible=False), gr.update(visible=True), True, None],
                    None, 
                    [row[counter_add], file_collection[counter_add], buttons_collection[counter_add-1], buttons_collection[counter_add], is_visible[counter_add], file_collection[counter_add]])
                else:
                    button.click(lambda:[gr.update(visible=True),gr.update(visible=True), gr.update(visible=False), True], None, [row[counter_add], file_collection[counter_add], buttons_collection[counter_add-1], is_visible[counter_add]])
                counter_add += 1
            
            counter_delete = 1
            for delete_button in delete_collection:
                if(counter_delete < len(delete_collection)+1):
                    delete_button.click(lambda:[gr.update(visible=False),gr.update(visible=False), gr.update(visible=True), False], None, [file_collection[counter_delete], row[counter_delete], buttons_collection[counter_delete-1], is_visible[counter_delete]])
                counter_delete += 1
            
            
            
    with gr.Accordion("Advanced Settings", open=False):
        swap_auto_calculated = gr.Checkbox(label="Use these advanced setting")
        gr.Markdown("If not checked, the number of steps and % of frozen encoder will be tuned automatically according to the amount of images you upload and whether you are training an `object`, `person` or `style`.")
        steps = gr.Number(label="How many steps", value=800)
        perc_txt_encoder = gr.Number(label="Percentage of the training steps the text-encoder should be trained as well", value=30)

    #for file in file_collection:
    #    file.change(fn=swap_values_files, inputs=file_collection, outputs=[steps])

    type_of_thing.change(fn=swap_text, inputs=[type_of_thing], outputs=[thing_description, thing_image_example, things_naming, perc_txt_encoder])
    train_btn = gr.Button("Start Training")
    with gr.Box(visible=False) as try_your_model:
        gr.Markdown("Try your model")
        with gr.Row():
            prompt = gr.Textbox(label="Type your prompt")
            result = gr.Image()
        generate_button = gr.Button("Generate Image")
    with gr.Box(visible=False) as push_to_hub:
        gr.Markdown("Push to Hugging Face Hub")
        model_repo_tag = gr.Textbox(label="Model name or URL", placeholder="username/model_name")
        push_button = gr.Button("Push to the Hub")
    result = gr.File(label="Download the uploaded models (zip file are diffusers weights, *.ckpt are CompVis/AUTOMATIC1111 weights)", visible=False)
    train_btn.click(fn=train, inputs=is_visible+concept_collection+file_collection+[type_of_thing]+[steps]+[perc_txt_encoder]+[swap_auto_calculated], outputs=[result, try_your_model, push_to_hub])
demo.launch()