Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
@@ -163,11 +163,17 @@ app.prepare(ctx_id=0, det_size=(640, 640))
|
|
163 |
|
164 |
# download checkpoints
|
165 |
print("Downloading checkpoints")
|
166 |
-
hf_hub_download(repo_id="briaai/
|
167 |
-
hf_hub_download(repo_id="briaai/
|
168 |
-
hf_hub_download(repo_id="briaai/
|
169 |
-
|
170 |
-
hf_hub_download(repo_id="briaai/
|
|
|
|
|
|
|
|
|
|
|
|
|
171 |
|
172 |
|
173 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
@@ -176,6 +182,8 @@ device = "cuda" if torch.cuda.is_available() else "cpu"
|
|
176 |
face_adapter = f"./checkpoints/checkpoint_105000/ip-adapter.bin"
|
177 |
controlnet_path = f"./checkpoints/checkpoint_105000/controlnet"
|
178 |
base_model_path = f'briaai/BRIA-2.3'
|
|
|
|
|
179 |
resolution = 1024
|
180 |
|
181 |
# Load ControlNet models
|
@@ -206,13 +214,19 @@ pipe.load_ip_adapter_instantid(face_adapter)
|
|
206 |
|
207 |
clip_embeds=None
|
208 |
|
|
|
|
|
|
|
|
|
|
|
|
|
209 |
|
210 |
-
|
211 |
|
212 |
@spaces.GPU
|
213 |
-
|
214 |
-
def generate_image(image_path, prompt, num_steps, guidance_scale, seed, num_images, ip_adapter_scale, kps_scale, canny_scale, progress=gr.Progress(track_tqdm=True)):
|
215 |
-
|
216 |
|
217 |
if image_path is None:
|
218 |
raise gr.Error(f"Cannot find any input face image! Please upload a face image.")
|
@@ -239,9 +253,6 @@ def generate_image(image_path, prompt, num_steps, guidance_scale, seed, num_imag
|
|
239 |
files = [
|
240 |
('file', ('image_name.jpeg', image_file, 'image/jpeg')) # Specify file name, file-like object, and MIME type
|
241 |
]
|
242 |
-
# headers = {
|
243 |
-
# 'api_token': 'a10d6386dd6a11ebba800242ac130004'
|
244 |
-
# }
|
245 |
headers = {
|
246 |
'api_token': os.getenv('BRIA_RMBG_TOKEN') # Securely retrieve the token
|
247 |
}
|
@@ -269,7 +280,32 @@ def generate_image(image_path, prompt, num_steps, guidance_scale, seed, num_imag
|
|
269 |
|
270 |
generator = torch.Generator(device=device).manual_seed(seed)
|
271 |
|
272 |
-
full_prompt = prompt
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
273 |
|
274 |
print("Start inference...")
|
275 |
images = pipe(
|
@@ -341,19 +377,21 @@ with gr.Blocks(css=css) as demo:
|
|
341 |
placeholder="Enter your prompt here",
|
342 |
info="Describe what you want to generate or modify in the image."
|
343 |
)
|
344 |
-
|
|
|
|
|
345 |
submit = gr.Button("Submit", variant="primary")
|
346 |
|
347 |
with gr.Accordion(open=False, label="Advanced Options"):
|
348 |
num_steps = gr.Slider(
|
349 |
-
label="Number of
|
350 |
minimum=1,
|
351 |
maximum=100,
|
352 |
step=1,
|
353 |
value=30,
|
354 |
)
|
355 |
guidance_scale = gr.Slider(
|
356 |
-
label="
|
357 |
minimum=0.1,
|
358 |
maximum=10.0,
|
359 |
step=0.1,
|
@@ -367,27 +405,33 @@ with gr.Blocks(css=css) as demo:
|
|
367 |
value=1,
|
368 |
)
|
369 |
ip_adapter_scale = gr.Slider(
|
370 |
-
label="
|
371 |
minimum=0.0,
|
372 |
maximum=1.0,
|
373 |
step=0.01,
|
374 |
value=0.8,
|
375 |
)
|
376 |
kps_scale = gr.Slider(
|
377 |
-
label="
|
378 |
minimum=0.0,
|
379 |
maximum=1.0,
|
380 |
step=0.01,
|
381 |
value=0.6,
|
382 |
)
|
383 |
canny_scale = gr.Slider(
|
384 |
-
label="canny
|
385 |
minimum=0.0,
|
386 |
maximum=1.0,
|
387 |
step=0.01,
|
388 |
value=0.4,
|
389 |
)
|
390 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
391 |
seed = gr.Slider(
|
392 |
label="Seed",
|
393 |
minimum=0,
|
@@ -409,7 +453,8 @@ with gr.Blocks(css=css) as demo:
|
|
409 |
api_name=False,
|
410 |
).then(
|
411 |
fn=generate_image,
|
412 |
-
inputs=[img_file, prompt, num_steps, guidance_scale, seed, num_images, ip_adapter_scale, kps_scale, canny_scale],
|
|
|
413 |
# outputs=[gallery]
|
414 |
outputs=gallery
|
415 |
)
|
|
|
163 |
|
164 |
# download checkpoints
|
165 |
print("Downloading checkpoints")
|
166 |
+
hf_hub_download(repo_id="briaai/BRIA-2.3-ID_Preservation", filename="checkpoint_105000/controlnet/config.json", local_dir="./checkpoints")
|
167 |
+
hf_hub_download(repo_id="briaai/BRIA-2.3-ID_Preservation", filename="checkpoint_105000/controlnet/diffusion_pytorch_model.safetensors", local_dir="./checkpoints")
|
168 |
+
hf_hub_download(repo_id="briaai/BRIA-2.3-ID_Preservation", filename="checkpoint_105000/ip-adapter.bin", local_dir="./checkpoints")
|
169 |
+
|
170 |
+
hf_hub_download(repo_id="briaai/BRIA-2.3-ID_Preservation", filename="image_encoder/pytorch_model.bin", local_dir="./checkpoints")
|
171 |
+
hf_hub_download(repo_id="briaai/BRIA-2.3-ID_Preservation", filename="image_encoder/config.json", local_dir="./checkpoints")
|
172 |
+
|
173 |
+
# Download Lora weights
|
174 |
+
hf_hub_download(repo_id="briaai/BRIA-2.3-ID_Preservation", filename="LoRAs/3D_avatar/pytorch_lora_weights.safetensors", local_dir=".")
|
175 |
+
hf_hub_download(repo_id="briaai/BRIA-2.3-ID_Preservation", filename="LoRAs/coloringbook/pytorch_lora_weights.safetensors", local_dir=".")
|
176 |
+
hf_hub_download(repo_id="briaai/BRIA-2.3-ID_Preservation", filename="LoRAs/One_line_portraits_Light/pytorch_lora_weights.safetensors", local_dir=".")
|
177 |
|
178 |
|
179 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
|
|
182 |
face_adapter = f"./checkpoints/checkpoint_105000/ip-adapter.bin"
|
183 |
controlnet_path = f"./checkpoints/checkpoint_105000/controlnet"
|
184 |
base_model_path = f'briaai/BRIA-2.3'
|
185 |
+
lora_base_path = f"./LoRAs"
|
186 |
+
|
187 |
resolution = 1024
|
188 |
|
189 |
# Load ControlNet models
|
|
|
214 |
|
215 |
clip_embeds=None
|
216 |
|
217 |
+
Loras_dict = {
|
218 |
+
"":"",
|
219 |
+
"One_line_portraits_Light": "An illustration of ",
|
220 |
+
"3D_avatar": "An illustration of ",
|
221 |
+
"coloringbook": "An illustration of "
|
222 |
+
}
|
223 |
|
224 |
+
lora_names = Loras_dict.keys()
|
225 |
|
226 |
@spaces.GPU
|
227 |
+
def generate_image(image_path, prompt, num_steps, guidance_scale, seed, num_images, ip_adapter_scale, kps_scale, canny_scale, lora_name, lora_scale, progress=gr.Progress(track_tqdm=True)):
|
228 |
+
# def generate_image(image_path, prompt, num_steps, guidance_scale, seed, num_images, ip_adapter_scale, kps_scale, canny_scale, progress=gr.Progress(track_tqdm=True)):
|
229 |
+
global CURRENT_LORA_NAME # Use the global variable to track LoRA
|
230 |
|
231 |
if image_path is None:
|
232 |
raise gr.Error(f"Cannot find any input face image! Please upload a face image.")
|
|
|
253 |
files = [
|
254 |
('file', ('image_name.jpeg', image_file, 'image/jpeg')) # Specify file name, file-like object, and MIME type
|
255 |
]
|
|
|
|
|
|
|
256 |
headers = {
|
257 |
'api_token': os.getenv('BRIA_RMBG_TOKEN') # Securely retrieve the token
|
258 |
}
|
|
|
280 |
|
281 |
generator = torch.Generator(device=device).manual_seed(seed)
|
282 |
|
283 |
+
# full_prompt = prompt
|
284 |
+
if lora_name != CURRENT_LORA_NAME: # Check if LoRA needs to be changed
|
285 |
+
if CURRENT_LORA_NAME is not None: # If a LoRA is already loaded, unload it
|
286 |
+
pipe.disable_lora()
|
287 |
+
pipe.unfuse_lora()
|
288 |
+
pipe.unload_lora_weights()
|
289 |
+
print(f"Unloaded LoRA: {CURRENT_LORA_NAME}")
|
290 |
+
|
291 |
+
if lora_name != "": # Load the new LoRA if specified
|
292 |
+
# pipe.enable_model_cpu_offload()
|
293 |
+
lora_path = os.path.join(lora_base_path, lora_name, "pytorch_lora_weights.safetensors")
|
294 |
+
pipe.load_lora_weights(lora_path)
|
295 |
+
pipe.fuse_lora(lora_scale)
|
296 |
+
pipe.enable_lora()
|
297 |
+
|
298 |
+
# lora_prefix = Loras_dict[lora_name]
|
299 |
+
|
300 |
+
print(f"Loaded new LoRA: {lora_name}")
|
301 |
+
|
302 |
+
# Update the current LoRA name
|
303 |
+
CURRENT_LORA_NAME = lora_name
|
304 |
+
|
305 |
+
if lora_name != "":
|
306 |
+
full_prompt = f"{Loras_dict[lora_name]} + " " + {prompt}"
|
307 |
+
else:
|
308 |
+
full_prompt = prompt
|
309 |
|
310 |
print("Start inference...")
|
311 |
images = pipe(
|
|
|
377 |
placeholder="Enter your prompt here",
|
378 |
info="Describe what you want to generate or modify in the image."
|
379 |
)
|
380 |
+
|
381 |
+
lora_name = gr.Dropdown(choices=lora_names, label="LoRA", value="", info="Select a LoRA name from the list, not selecting any will disable LoRA.")
|
382 |
+
|
383 |
submit = gr.Button("Submit", variant="primary")
|
384 |
|
385 |
with gr.Accordion(open=False, label="Advanced Options"):
|
386 |
num_steps = gr.Slider(
|
387 |
+
label="Number of diffusion steps",
|
388 |
minimum=1,
|
389 |
maximum=100,
|
390 |
step=1,
|
391 |
value=30,
|
392 |
)
|
393 |
guidance_scale = gr.Slider(
|
394 |
+
label="cfg scale",
|
395 |
minimum=0.1,
|
396 |
maximum=10.0,
|
397 |
step=0.1,
|
|
|
405 |
value=1,
|
406 |
)
|
407 |
ip_adapter_scale = gr.Slider(
|
408 |
+
label="ID Adapter scale",
|
409 |
minimum=0.0,
|
410 |
maximum=1.0,
|
411 |
step=0.01,
|
412 |
value=0.8,
|
413 |
)
|
414 |
kps_scale = gr.Slider(
|
415 |
+
label="lnmks ControlNet scale",
|
416 |
minimum=0.0,
|
417 |
maximum=1.0,
|
418 |
step=0.01,
|
419 |
value=0.6,
|
420 |
)
|
421 |
canny_scale = gr.Slider(
|
422 |
+
label="canny ControlNet scale",
|
423 |
minimum=0.0,
|
424 |
maximum=1.0,
|
425 |
step=0.01,
|
426 |
value=0.4,
|
427 |
)
|
428 |
+
lora_scale = gr.Slider(
|
429 |
+
label="LoRA scale",
|
430 |
+
minimum=0.0,
|
431 |
+
maximum=1.0,
|
432 |
+
step=0.01,
|
433 |
+
value=0.7,
|
434 |
+
)
|
435 |
seed = gr.Slider(
|
436 |
label="Seed",
|
437 |
minimum=0,
|
|
|
453 |
api_name=False,
|
454 |
).then(
|
455 |
fn=generate_image,
|
456 |
+
# inputs=[img_file, prompt, num_steps, guidance_scale, seed, num_images, ip_adapter_scale, kps_scale, canny_scale],
|
457 |
+
inputs=[img_file, prompt, num_steps, guidance_scale, seed, num_images, ip_adapter_scale, kps_scale, canny_scale, lora_name, lora_scale],
|
458 |
# outputs=[gallery]
|
459 |
outputs=gallery
|
460 |
)
|