Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
@@ -9,7 +9,7 @@ import logging
|
|
9 |
from diffusers import DiffusionPipeline, AutoencoderTiny, AutoencoderKL, AutoPipelineForImage2Image
|
10 |
from huggingface_hub import login
|
11 |
from diffusers.utils import load_image
|
12 |
-
|
13 |
import time
|
14 |
from datetime import datetime
|
15 |
from io import BytesIO
|
@@ -34,19 +34,13 @@ base_model = "black-forest-labs/FLUX.1-dev"
|
|
34 |
# load pipe
|
35 |
taef1 = AutoencoderTiny.from_pretrained("madebyollin/taef1", torch_dtype=dtype).to(device)
|
36 |
good_vae = AutoencoderKL.from_pretrained(base_model, subfolder="vae", torch_dtype=dtype).to(device)
|
37 |
-
pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=dtype).to(device)
|
38 |
|
39 |
-
|
40 |
-
|
41 |
-
vae=good_vae,
|
42 |
-
transformer=pipe.transformer,
|
43 |
-
text_encoder=pipe.text_encoder,
|
44 |
-
tokenizer=pipe.tokenizer,
|
45 |
-
text_encoder_2=pipe.text_encoder_2,
|
46 |
-
tokenizer_2=pipe.tokenizer_2,
|
47 |
-
torch_dtype=dtype
|
48 |
-
)
|
49 |
|
|
|
|
|
|
|
50 |
|
51 |
|
52 |
MAX_SEED = 2**32 - 1
|
@@ -76,15 +70,19 @@ def generate_image(orginal_image, prompt, adapter_names, steps, seed, image_str
|
|
76 |
|
77 |
|
78 |
gr.Info("Start to generate images ...")
|
79 |
-
with calculateDuration(f"Make a new generator:{seed}"):
|
80 |
-
|
|
|
|
|
|
|
|
|
81 |
generator = torch.Generator(device=device).manual_seed(seed)
|
82 |
|
83 |
with calculateDuration("Generating image"):
|
84 |
# Generate image
|
85 |
-
joint_attention_kwargs = {"scale": 1}
|
86 |
if orginal_image:
|
87 |
-
generated_image =
|
88 |
prompt=prompt,
|
89 |
image=orginal_image,
|
90 |
strength=image_strength,
|
@@ -96,7 +94,7 @@ def generate_image(orginal_image, prompt, adapter_names, steps, seed, image_str
|
|
96 |
joint_attention_kwargs=joint_attention_kwargs
|
97 |
).images[0]
|
98 |
else:
|
99 |
-
generated_image =
|
100 |
prompt=prompt,
|
101 |
num_inference_steps=steps,
|
102 |
guidance_scale=cfg_scale,
|
@@ -189,18 +187,18 @@ def run_lora(prompt, image_url, lora_strings_json, image_strength, cfg_scale, s
|
|
189 |
if lora_repo and weights and adapter_name:
|
190 |
try:
|
191 |
if img2img_model:
|
192 |
-
|
193 |
else:
|
194 |
-
|
195 |
except:
|
196 |
print("load lora error")
|
197 |
|
198 |
# set lora weights
|
199 |
if len(adapter_names) > 0:
|
200 |
if img2img_model:
|
201 |
-
|
202 |
else:
|
203 |
-
|
204 |
|
205 |
|
206 |
# Generate image
|
|
|
9 |
from diffusers import DiffusionPipeline, AutoencoderTiny, AutoencoderKL, AutoPipelineForImage2Image
|
10 |
from huggingface_hub import login
|
11 |
from diffusers.utils import load_image
|
12 |
+
from lora_loading_patch import load_lora_into_transformer
|
13 |
import time
|
14 |
from datetime import datetime
|
15 |
from io import BytesIO
|
|
|
34 |
# load pipe
|
35 |
taef1 = AutoencoderTiny.from_pretrained("madebyollin/taef1", torch_dtype=dtype).to(device)
|
36 |
good_vae = AutoencoderKL.from_pretrained(base_model, subfolder="vae", torch_dtype=dtype).to(device)
|
|
|
37 |
|
38 |
+
txt2img_pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=dtype).to(device)
|
39 |
+
txt2img_pipe.__class__.load_lora_into_transformer = classmethod(load_lora_into_transformer)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
|
41 |
+
# img2img model
|
42 |
+
img2img_pipe = AutoPipelineForImage2Image.from_pretrained(base_model, vae=good_vae, transformer=txt2img_pipe.transformer, text_encoder=txt2img_pipe.text_encoder, tokenizer=txt2img_pipe.tokenizer, text_encoder_2=txt2img_pipe.text_encoder_2, tokenizer_2=txt2img_pipe.tokenizer_2, torch_dtype=dtype)
|
43 |
+
img2img_pipe.__class__.load_lora_into_transformer = classmethod(load_lora_into_transformer)
|
44 |
|
45 |
|
46 |
MAX_SEED = 2**32 - 1
|
|
|
70 |
|
71 |
|
72 |
gr.Info("Start to generate images ...")
|
73 |
+
with calculateDuration(f"Make a new generator: {seed}"):
|
74 |
+
if orginal_image:
|
75 |
+
img2img_pipe.to(device)
|
76 |
+
else:
|
77 |
+
txt2img_pipe.to(device)
|
78 |
+
|
79 |
generator = torch.Generator(device=device).manual_seed(seed)
|
80 |
|
81 |
with calculateDuration("Generating image"):
|
82 |
# Generate image
|
83 |
+
joint_attention_kwargs = {"scale": 1}
|
84 |
if orginal_image:
|
85 |
+
generated_image = img2img_pipe(
|
86 |
prompt=prompt,
|
87 |
image=orginal_image,
|
88 |
strength=image_strength,
|
|
|
94 |
joint_attention_kwargs=joint_attention_kwargs
|
95 |
).images[0]
|
96 |
else:
|
97 |
+
generated_image = txt2img_pipe(
|
98 |
prompt=prompt,
|
99 |
num_inference_steps=steps,
|
100 |
guidance_scale=cfg_scale,
|
|
|
187 |
if lora_repo and weights and adapter_name:
|
188 |
try:
|
189 |
if img2img_model:
|
190 |
+
img2img_pipe.load_lora_weights(lora_repo, weight_name=weights, adapter_name=adapter_name)
|
191 |
else:
|
192 |
+
txt2img_pipe.load_lora_weights(lora_repo, weight_name=weights, adapter_name=adapter_name)
|
193 |
except:
|
194 |
print("load lora error")
|
195 |
|
196 |
# set lora weights
|
197 |
if len(adapter_names) > 0:
|
198 |
if img2img_model:
|
199 |
+
img2img_pipe.set_adapters(adapter_names, adapter_weights=adapter_weights)
|
200 |
else:
|
201 |
+
txt2img_pipe.set_adapters(adapter_names, adapter_weights=adapter_weights)
|
202 |
|
203 |
|
204 |
# Generate image
|