Spaces:
Running
on
Zero
Running
on
Zero
File size: 19,987 Bytes
228e8c1 edd2900 97bb8f1 83b359b 228e8c1 97bb8f1 83b359b 228e8c1 81d025b 97bb8f1 edd2900 97bb8f1 edd2900 97bb8f1 81d025b 97bb8f1 228e8c1 83b359b 91afacd 83b359b 91afacd 83b359b 91afacd 83b359b 91afacd 83b359b f28502f cc2019d 83b359b 228e8c1 97bb8f1 83b359b 228e8c1 97bb8f1 228e8c1 83b359b ba9dade 83b359b 91afacd ba9dade 83b359b ba9dade 83b359b 91afacd ba9dade 83b359b ba9dade 83b359b 91afacd 83b359b 97bb8f1 83b359b 97bb8f1 83b359b 97bb8f1 228e8c1 97bb8f1 228e8c1 97bb8f1 83b359b 97bb8f1 83b359b 97bb8f1 228e8c1 97bb8f1 228e8c1 97bb8f1 228e8c1 97bb8f1 228e8c1 97bb8f1 83b359b 97bb8f1 228e8c1 97bb8f1 228e8c1 97bb8f1 228e8c1 97bb8f1 228e8c1 97bb8f1 228e8c1 83b359b 6e0ff3e 228e8c1 d5e9410 228e8c1 83b359b 228e8c1 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 |
import spaces
import gradio as gr
import os
import sys
import subprocess
import numpy as np
from paligemma2 import PaliGemma2Handler, MODELS as PALIGEMMA_MODELS
from gemma import GemmaHandler, MODELS as GEMMA_MODELS
from gemma_multiline import GemmaMultilineHandler, MODELS as GEMMA_MULTILINE_MODELS
# Initialize model handlers
paligemma_handler = PaliGemma2Handler()
gemma_handler = GemmaHandler()
gemma_multiline_handler = GemmaMultilineHandler()
@spaces.GPU
def process_image_paligemma(model_name, image, progress=gr.Progress()):
"""Process a single image with PaliGemma2"""
return paligemma_handler.process_image(model_name, image, progress)
@spaces.GPU
def process_image_gemma(model_name, image, progress=gr.Progress()):
"""Process a single image with Gemma"""
return gemma_handler.process_image(model_name, image, progress)
@spaces.GPU
def process_pdf_paligemma(pdf_path, model_name, progress=gr.Progress()):
"""Process a PDF file with PaliGemma2"""
return paligemma_handler.process_pdf(pdf_path, model_name, progress)
@spaces.GPU
def process_pdf_gemma(pdf_path, model_name, progress=gr.Progress()):
"""Process a PDF file with Gemma"""
return gemma_handler.process_pdf(pdf_path, model_name, progress)
@spaces.GPU
def process_image_multiline(model_name, image, temp, top_p, repetition_penalty, progress=gr.Progress()):
return gemma_multiline_handler.generate_text_from_image(model_name, image, temp, top_p, repetition_penalty, progress)
@spaces.GPU
def process_image_multiline_stream(model_name, image, temp, top_p, repetition_penalty, progress=gr.Progress()):
yield from gemma_multiline_handler.generate_text_stream(model_name, image, temp, top_p, repetition_penalty, progress)
@spaces.GPU
def process_pdf_multiline(model_name, pdf, temp, top_p, repetition_penalty, progress=gr.Progress()):
return gemma_multiline_handler.process_pdf(model_name, pdf, temp, top_p, repetition_penalty, progress)
@spaces.GPU
def process_pdf_multiline_stream(model_name, pdf, temp, top_p, repetition_penalty, progress=gr.Progress()):
yield from gemma_multiline_handler.process_pdf_stream(model_name, pdf, temp, top_p, repetition_penalty, progress)
# Example images for document-level OCR
document_examples = [
["ml.png", "Multi-line Dhivehi text sample"],
["ml1.png", "Multi-line Dhivehi text sample 2"],
["ml2.png", "Multi-line Dhivehi text sample 3"],
["ml3.png", "Multi-line Dhivehi text sample 4"],
]
# Example images for sentence-level OCR
sentence_examples = [
["type_1_sl.png", "Typed Dhivehi text sample 1"],
["type_2_sl.png", "Typed Dhivehi text sample 2"],
["hw_1_sl.png", "Handwritten Dhivehi text sample 1"],
["hw_2_sl.jpg", "Handwritten Dhivehi text sample 2"],
["hw_3_sl.png", "Handwritten Dhivehi text sample 3"],
["hw_4_sl.png", "Handwritten Dhivehi text sample 4"],
["ml.png", "Multi-line Dhivehi text sample"],
]
css = """
.textbox1 textarea {
font-size: 18px !important;
font-family: 'MV_Faseyha', 'Faruma', 'A_Faruma' !important;
line-height: 1.8 !important;
}
.textbox2 textarea {
display: none;
}
"""
with gr.Blocks(title="Dhivehi Image to Text",css=css) as demo:
gr.Markdown("# Dhivehi Image to Text")
gr.Markdown("Dhivehi Image to Text experimental finetunes")
with gr.Tabs():
with gr.Tab("Gemma Document"):
with gr.Row():
model_path_dropdown = gr.Dropdown(
label="Model Checkpoint",
choices=list(GEMMA_MULTILINE_MODELS.keys()),
value=list(GEMMA_MULTILINE_MODELS.keys())[0],
interactive=True,
scale=2
)
with gr.Accordion("Advanced Options", open=False):
with gr.Row():
temperature_slider = gr.Slider(
minimum=0.1, maximum=1.9, value=0.2, step=0.1,
label="Temperature", info="Controls randomness in generation"
)
top_p_slider = gr.Slider(
minimum=0.1, maximum=1.0, value=1, step=0.1,
label="Top-p", info="Controls diversity via nucleus sampling"
)
repetition_penalty_slider = gr.Slider(
minimum=1.0, maximum=2.0, value=1.2, step=0.1,
label="Repetition Penalty", info="Penalizes repeated tokens. >1 encourages new tokens."
)
with gr.Tabs():
with gr.Tab("Image Input"):
with gr.Row():
with gr.Column():
image_input = gr.Image(type="pil", label="Upload Image")
with gr.Row():
generate_button = gr.Button("Generate Text (Non-streaming)")
stream_button = gr.Button("Generate Text (Streaming)", variant="primary")
stop_button = gr.Button("Stop", visible=False, variant="stop")
gr.Examples(
examples=[[img] for img, _ in document_examples],
inputs=[image_input],
outputs=None,
label="Example Images",
examples_per_page=7
)
with gr.Column():
text_output = gr.Textbox(
label="Extracted Dhivehi Text",
lines=20,
rtl=True,
elem_classes=["textbox1"],
show_copy_button=True,
scale=2
)
def show_stop_button_image():
return gr.update(visible=True), gr.update(interactive=False), gr.update(interactive=False)
def hide_stop_button_image():
return gr.update(visible=False), gr.update(interactive=True), gr.update(interactive=True)
generate_button.click(
fn=process_image_multiline,
inputs=[model_path_dropdown, image_input, temperature_slider, top_p_slider, repetition_penalty_slider],
outputs=text_output,
show_progress="full"
)
show_event = stream_button.click(fn=show_stop_button_image, outputs=[stop_button, stream_button, generate_button])
gen_event = show_event.then(fn=process_image_multiline_stream, inputs=[model_path_dropdown, image_input, temperature_slider, top_p_slider, repetition_penalty_slider], outputs=text_output, show_progress="full")
gen_event.then(fn=hide_stop_button_image, outputs=[stop_button, stream_button, generate_button])
stop_button.click(fn=hide_stop_button_image, outputs=[stop_button, stream_button, generate_button], cancels=[gen_event])
with gr.Tab("PDF Input"):
with gr.Row():
with gr.Column():
pdf_input = gr.File(label="Upload PDF", file_types=[".pdf"])
with gr.Row():
pdf_generate_button = gr.Button("Generate Text (Non-streaming)")
pdf_stream_button = gr.Button("Generate Text (Streaming)", variant="primary")
pdf_stop_button = gr.Button("Stop", visible=False, variant="stop")
gr.Examples(
examples=[["example.pdf", "Example PDF"]],
inputs=[pdf_input],
outputs=None,
label="Example PDFs",
examples_per_page=7
)
with gr.Column():
pdf_text_output = gr.Textbox(
label="Extracted Dhivehi Text",
lines=20,
rtl=True,
elem_classes=["textbox1"],
show_copy_button=True,
scale=2
)
def show_stop_button_pdf():
return gr.update(visible=True), gr.update(interactive=False), gr.update(interactive=False)
def hide_stop_button_pdf():
return gr.update(visible=False), gr.update(interactive=True), gr.update(interactive=True)
pdf_generate_button.click(
fn=process_pdf_multiline,
inputs=[model_path_dropdown, pdf_input, temperature_slider, top_p_slider, repetition_penalty_slider],
outputs=pdf_text_output,
show_progress="full"
)
pdf_show_event = pdf_stream_button.click(fn=show_stop_button_pdf, outputs=[pdf_stop_button, pdf_stream_button, pdf_generate_button])
pdf_gen_event = pdf_show_event.then(fn=process_pdf_multiline_stream, inputs=[model_path_dropdown, pdf_input, temperature_slider, top_p_slider, repetition_penalty_slider], outputs=pdf_text_output, show_progress="full")
pdf_gen_event.then(fn=hide_stop_button_pdf, outputs=[pdf_stop_button, pdf_stream_button, pdf_generate_button])
pdf_stop_button.click(fn=hide_stop_button_pdf, outputs=[pdf_stop_button, pdf_stream_button, pdf_generate_button], cancels=[pdf_gen_event])
# model_path_dropdown.change(fn=load_model_multiline, inputs=model_path_dropdown)
with gr.Tab("PaliGemma"):
model_dropdown_paligemma = gr.Dropdown(
choices=list(PALIGEMMA_MODELS.keys()),
value=list(PALIGEMMA_MODELS.keys())[0],
label="Select PaliGemma Model"
)
with gr.Tabs():
with gr.Tab("Image Input"):
with gr.Row():
with gr.Column(scale=2):
image_input_paligemma = gr.Image(type="pil", label="Input Image")
image_submit_btn_paligemma = gr.Button("Extract Text")
# Image examples
gr.Examples(
examples=[[img] for img, _ in sentence_examples],
inputs=[image_input_paligemma],
label="Example Images",
examples_per_page=8
)
with gr.Column(scale=3):
with gr.Tabs():
with gr.Tab("Extracted Text"):
image_text_output_paligemma = gr.Textbox(
lines=5,
label="Extracted Text",
show_copy_button=True,
rtl=True,
elem_classes="textbox1"
)
with gr.Tab("Detected Text Regions"):
image_bbox_output_paligemma = gr.Gallery(
label="Detected Text Regions",
show_label=True,
columns=2
)
with gr.Tab("PDF Input"):
with gr.Row():
with gr.Column(scale=2):
pdf_input_paligemma = gr.File(
label="Input PDF",
file_types=[".pdf"]
)
pdf_submit_btn_paligemma = gr.Button("Extract Text from PDF")
# PDF examples
gr.Examples(
examples=[
["example.pdf", "Example 1"],
],
inputs=[pdf_input_paligemma],
label="Example PDFs",
examples_per_page=8
)
with gr.Column(scale=3):
with gr.Tabs():
with gr.Tab("Extracted Text"):
pdf_text_output_paligemma = gr.Textbox(
lines=5,
label="Extracted Text",
show_copy_button=True,
rtl=True,
elem_classes="textbox1"
)
with gr.Tab("Detected Text Regions"):
pdf_bbox_output_paligemma = gr.Gallery(
label="Detected Text Regions",
show_label=True,
columns=2
)
with gr.Tab("Gemma Sentence"):
model_dropdown_gemma = gr.Dropdown(
choices=list(GEMMA_MODELS.keys()),
value=list(GEMMA_MODELS.keys())[0],
label="Select Gemma Model"
)
with gr.Tabs():
with gr.Tab("Image Input"):
with gr.Row():
with gr.Column(scale=2):
image_input_gemma = gr.Image(type="pil", label="Input Image")
image_submit_btn_gemma = gr.Button("Extract Text")
# Image examples
gr.Examples(
examples=[[img] for img, _ in sentence_examples],
inputs=[image_input_gemma],
label="Example Images",
examples_per_page=8
)
with gr.Column(scale=3):
with gr.Tabs():
with gr.Tab("Extracted Text"):
image_text_output_gemma = gr.Textbox(
lines=5,
label="Extracted Text",
show_copy_button=True,
rtl=True,
elem_classes="textbox1"
)
with gr.Tab("Detected Text Regions"):
image_bbox_output_gemma = gr.Gallery(
label="Detected Text Regions",
show_label=True,
columns=2
)
with gr.Tab("PDF Input"):
with gr.Row():
with gr.Column(scale=2):
pdf_input_gemma = gr.File(
label="Input PDF",
file_types=[".pdf"]
)
pdf_submit_btn_gemma = gr.Button("Extract Text from PDF")
# PDF examples
gr.Examples(
examples=[
["example.pdf", "Example 1"],
],
inputs=[pdf_input_gemma],
label="Example PDFs",
examples_per_page=8
)
with gr.Column(scale=3):
with gr.Tabs():
with gr.Tab("Extracted Text"):
pdf_text_output_gemma = gr.Textbox(
lines=5,
label="Extracted Text",
show_copy_button=True,
rtl=True,
elem_classes="textbox1"
)
with gr.Tab("Detected Text Regions"):
pdf_bbox_output_gemma = gr.Gallery(
label="Detected Text Regions",
show_label=True,
columns=2
)
# PaliGemma event handlers
image_submit_btn_paligemma.click(
fn=process_image_paligemma,
inputs=[model_dropdown_paligemma, image_input_paligemma],
outputs=[image_text_output_paligemma, image_bbox_output_paligemma]
)
pdf_submit_btn_paligemma.click(
fn=process_pdf_paligemma,
inputs=[pdf_input_paligemma, model_dropdown_paligemma],
outputs=[pdf_text_output_paligemma, pdf_bbox_output_paligemma]
)
# Gemma event handlers
image_submit_btn_gemma.click(
fn=process_image_gemma,
inputs=[model_dropdown_gemma, image_input_gemma],
outputs=[image_text_output_gemma, image_bbox_output_gemma]
)
pdf_submit_btn_gemma.click(
fn=process_pdf_gemma,
inputs=[pdf_input_gemma, model_dropdown_gemma],
outputs=[pdf_text_output_gemma, pdf_bbox_output_gemma]
)
# Function to install requirements
def install_requirements():
requirements_path = 'requirements.txt'
# Check if requirements.txt exists
if not os.path.exists(requirements_path):
print("Error: requirements.txt not found")
return False
try:
print("Installing requirements...")
# Using --no-cache-dir to avoid memory issues
subprocess.check_call([
sys.executable,
"-m",
"pip",
"install",
"-r",
requirements_path,
"--no-cache-dir"
])
print("Successfully installed all requirements")
return True
except subprocess.CalledProcessError as e:
print(f"Error installing requirements: {e}")
return False
except Exception as e:
print(f"Unexpected error: {e}")
return False
# Launch the app
if __name__ == "__main__":
# First install requirements
success = install_requirements()
if success:
print("All requirements installed successfully")
# Pre-load the multiline gemma model
#print("Loading default Gemma Multiline model...")
#gemma_multiline_handler.load_model(list(GEMMA_MULTILINE_MODELS.keys())[0])
#print("Default model loaded.")
demo.launch()
else:
print("Failed to install some requirements")
print("Failed to install some requirements") |